Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,11 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+GraphUNets/doc/GUnet.png filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e-2.json filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e-3.json filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e-4.json filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e-8.json filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e0.json filter=lfs diff=lfs merge=lfs -text
+seeds_diam_1e3.json filter=lfs diff=lfs merge=lfs -text
+src/synthetic/Table1-Epsilon-Ablations.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,49 @@
+# Compiled class file
+*.class
+
+# Log file
+*.log
+
+# BlueJ files
+*.ctxt
+
+# Mobile Tools for Java (J2ME)
+.mtj.tmp/
+
+# Package Files #
+*.jar
+*.war
+*.nar
+*.ear
+*.zip
+*.tar.gz
+*.rar
+
+# virtual machine crash logs
+hs_err_pid*
+
+# Temporary table files
+src/tmp_table1/
+
+# macOS system files
+.DS_Store
+
+# Dataset files
+*.cites
+*.content
+
+src/synthetic/data/*
+
+src/reddit_edges.txt
+
+out/*
+
+.idea/*
+
+iclr2026_3_reorganized.tex
+
+GraphUNets/data/*
+
+temp*
+
+*.tex

.ipynb_checkpoints/requirements-checkpoint.txt

@@ -0,0 +1,4 @@
+torch
+torch-scatter
+torch-sparse
+torch-geometric

GraphUNets/README.md

@@ -0,0 +1,78 @@
+PyTorch Implementation of Graph U-Nets
+======================================
+
+Created by [Hongyang Gao](https://faculty.sites.iastate.edu/hygao/) @ Iowa State University, and
+[Shuiwang Ji](http://people.tamu.edu/~sji/) @ Texas A&M University.
+
+About
+-----
+
+PyTorch implementation of Graph U-Nets. Check http://proceedings.mlr.press/v97/gao19a/gao19a.pdf for more information.
+
+Methods
+-------
+
+### Graph Pooling Layer
+
+![gPool](./doc/GPool.png)
+
+### Graph Unpooling Layer
+
+![gPool](./doc/GUnpool.png)
+
+### Graph U-Net
+
+![gPool](./doc/GUnet.png)
+
+Installation
+------------
+
+
+Type
+
+    ./run_GNN.sh DATA FOLD GPU
+to run on dataset using fold number (1-10).
+
+You can run
+
+    ./run_GNN.sh DD 0 0
+to run on DD dataset with 10-fold cross
+validation on GPU #0.
+
+
+Code
+----
+
+The detail implementation of Graph U-Net is in src/utils/ops.py.
+
+
+Datasets
+--------
+
+Check the "data/README.md" for the format. 
+
+
+Results
+-------
+
+
+| Models   | DD              | IMDBMULTI       | PROTEINS        |
+| -------- | --------------- | --------------- | --------------- |
+| PSCN     | 76.3 ± 2.6%     | 45.2 ± 2.8%     | 75.9 ± 2.8%     |
+| DIFFPOOL | 80.6%           | -               | 76.3%           |
+| SAGPool  | 76.5%           | -               | 71.9%           |
+| GIN      | 82.0 ± 2.7%     | 52.3 ± 2.8%     | 76.2 ± 2.8%     |
+| g-U-Net  | **83.0 ± 2.2%** | **56.7 ± 2.9%** | **78.7 ± 4.2%** |
+
+Reference
+---------
+
+If you find the code useful, please cite our paper:
+
+    @inproceedings{gao2019graph,
+        title={Graph U-Nets},
+        author={Gao, Hongyang and Ji, Shuiwang},
+        booktitle={International Conference on Machine Learning},
+        pages={2083--2092},
+        year={2019}
+    }

GraphUNets/configs/COLLAB

@@ -0,0 +1,12 @@
+num_epochs=200
+batch_size=64
+learning_rate=0.001
+deg_as_tag=1
+layer_num=3
+hidden_dim=512
+layer_dim=64
+drop_network=0.3
+drop_classifier=0.2
+activation_network=ELU
+activation_classifier=ReLU
+pool_rates_layers="0.8 0.6 0.4"

GraphUNets/configs/DD

@@ -0,0 +1,12 @@
+num_epochs=200
+batch_size=32
+learning_rate=0.001
+deg_as_tag=0
+layer_num=3
+hidden_dim=512
+layer_dim=128
+drop_network=0.3
+drop_classifier=0.3
+activation_network=ELU
+activation_classifier=ELU
+pool_rates_layers="0.9 0.8 0.7"

GraphUNets/configs/IMDBMULTI

@@ -0,0 +1,12 @@
+num_epochs=200
+batch_size=64
+learning_rate=0.001
+deg_as_tag=1
+layer_num=3
+hidden_dim=512
+layer_dim=48
+drop_network=0.1
+drop_classifier=0.1
+activation_network=LeakyReLU
+activation_classifier=ELU
+pool_rates_layers="0.9 0.9 0.9"

GraphUNets/configs/PROTEINS

@@ -0,0 +1,12 @@
+num_epochs=200
+batch_size=64
+learning_rate=0.001
+deg_as_tag=0
+layer_num=3
+hidden_dim=512
+layer_dim=64
+drop_network=0.3
+drop_classifier=0.3
+activation_network=ELU
+activation_classifier=ELU
+pool_rates_layers="0.9 0.8 0.7"

GraphUNets/run_GNN.sh

@@ -0,0 +1,48 @@
+#!/bin/bash
+DATA="${1-DD}"
+fold=${2-1}  # 0 for 10-fold
+GPU=${3-0}
+
+seed=1
+
+CONFIG=configs/${DATA}
+if [ ! -f "$CONFIG" ]; then
+    echo "No config file for ${DATA} in configs folder"
+    exit 128
+fi
+source configs/${DATA}
+
+FOLDER=results
+FILE=${FOLDER}/${DATA}.txt
+if [ ! -d "$FOLDER" ]; then
+    mkdir $FOLDER
+fi
+
+run(){
+    CUDA_VISIBLE_DEVICES=${GPU} python3 src/main.py \
+        -seed $seed -data $DATA -fold $1 -num_epochs $num_epochs \
+        -batch $batch_size -lr $learning_rate -deg_as_tag $deg_as_tag \
+        -l_num $layer_num -h_dim $hidden_dim -l_dim $layer_dim \
+        -drop_n $drop_network -drop_c $drop_classifier \
+        -act_n $activation_network -act_c $activation_classifier \
+        -ks $pool_rates_layers -acc_file $FILE
+}
+
+if [ ${fold} == 0 ]; then
+    if [ -f "$FILE" ]; then
+        rm $FILE
+    fi
+    echo "Running 10-fold cross validation"
+    start=`date +%s`
+    run $fold
+    stop=`date +%s`
+    echo "End of cross-validation using $[stop - start] seconds"
+    echo "The accuracy results for ${DATA} are as follows:"
+    cat $FILE
+    echo "Mean and sstdev are:"
+    cat $FILE | datamash mean 2 sstdev 2
+else
+    run $fold
+    echo "The accuracy result for ${DATA} fold ${fold} is:"
+    tail -1 $FILE
+fi

GraphUNets/src/main.py

@@ -0,0 +1,63 @@
+import argparse
+import random
+import time
+import torch
+import numpy as np
+from network import GNet
+from trainer import Trainer
+from utils.data_loader import FileLoader
+
+
+def get_args():
+    parser = argparse.ArgumentParser(description='Args for graph predition')
+    parser.add_argument('-seed', type=int, default=1, help='seed')
+    parser.add_argument('-data', default='DD', help='data folder name')
+    parser.add_argument('-fold', type=int, default=1, help='fold (1..10)')
+    parser.add_argument('-num_epochs', type=int, default=2, help='epochs')
+    parser.add_argument('-batch', type=int, default=8, help='batch size')
+    parser.add_argument('-lr', type=float, default=0.001, help='learning rate')
+    parser.add_argument('-deg_as_tag', type=int, default=0, help='1 or degree')
+    parser.add_argument('-l_num', type=int, default=3, help='layer num')
+    parser.add_argument('-h_dim', type=int, default=512, help='hidden dim')
+    parser.add_argument('-l_dim', type=int, default=48, help='layer dim')
+    parser.add_argument('-drop_n', type=float, default=0.3, help='drop net')
+    parser.add_argument('-drop_c', type=float, default=0.2, help='drop output')
+    parser.add_argument('-act_n', type=str, default='ELU', help='network act')
+    parser.add_argument('-act_c', type=str, default='ELU', help='output act')
+    parser.add_argument('-ks', nargs='+', type=float, default='0.9 0.8 0.7')
+    parser.add_argument('-acc_file', type=str, default='re', help='acc file')
+    args, _ = parser.parse_known_args()
+    return args
+
+
+def set_random(seed):
+    random.seed(seed)
+    np.random.seed(seed)
+    torch.manual_seed(seed)
+
+
+def app_run(args, G_data, fold_idx):
+    G_data.use_fold_data(fold_idx)
+    net = GNet(G_data.feat_dim, G_data.num_class, args)
+    trainer = Trainer(args, net, G_data)
+    trainer.train()
+
+
+def main():
+    args = get_args()
+    print(args)
+    set_random(args.seed)
+    start = time.time()
+    G_data = FileLoader(args).load_data()
+    print('load data using ------>', time.time()-start)
+    if args.fold == 0:
+        for fold_idx in range(10):
+            print('start training ------> fold', fold_idx+1)
+            app_run(args, G_data, fold_idx)
+    else:
+        print('start training ------> fold', args.fold)
+        app_run(args, G_data, args.fold-1)
+
+
+if __name__ == "__main__":
+    main()

GraphUNets/src/network.py

@@ -0,0 +1,60 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from utils.ops import GCN, GraphUnet, Initializer, norm_g
+
+
+class GNet(nn.Module):
+    def __init__(self, in_dim, n_classes, args):
+        super(GNet, self).__init__()
+        self.n_act = getattr(nn, args.act_n)()
+        self.c_act = getattr(nn, args.act_c)()
+        self.s_gcn = GCN(in_dim, args.l_dim, self.n_act, args.drop_n)
+        self.g_unet = GraphUnet(
+            args.ks, args.l_dim, args.l_dim, args.l_dim, self.n_act,
+            args.drop_n)
+        self.out_l_1 = nn.Linear(3*args.l_dim*(args.l_num+1), args.h_dim)
+        self.out_l_2 = nn.Linear(args.h_dim, n_classes)
+        self.out_drop = nn.Dropout(p=args.drop_c)
+        Initializer.weights_init(self)
+
+    def forward(self, gs, hs, labels):
+        hs = self.embed(gs, hs)
+        logits = self.classify(hs)
+        return self.metric(logits, labels)
+
+    def embed(self, gs, hs):
+        o_hs = []
+        for g, h in zip(gs, hs):
+            h = self.embed_one(g, h)
+            o_hs.append(h)
+        hs = torch.stack(o_hs, 0)
+        return hs
+
+    def embed_one(self, g, h):
+        g = norm_g(g)
+        h = self.s_gcn(g, h)
+        hs = self.g_unet(g, h)
+        h = self.readout(hs)
+        return h
+
+    def readout(self, hs):
+        h_max = [torch.max(h, 0)[0] for h in hs]
+        h_sum = [torch.sum(h, 0) for h in hs]
+        h_mean = [torch.mean(h, 0) for h in hs]
+        h = torch.cat(h_max + h_sum + h_mean)
+        return h
+
+    def classify(self, h):
+        h = self.out_drop(h)
+        h = self.out_l_1(h)
+        h = self.c_act(h)
+        h = self.out_drop(h)
+        h = self.out_l_2(h)
+        return F.log_softmax(h, dim=1)
+
+    def metric(self, logits, labels):
+        loss = F.nll_loss(logits, labels)
+        _, preds = torch.max(logits, 1)
+        acc = torch.mean((preds == labels).float())
+        return loss, acc

GraphUNets/src/trainer.py

@@ -0,0 +1,69 @@
+import torch
+from tqdm import tqdm
+import torch.optim as optim
+from utils.dataset import GraphData
+
+
+class Trainer:
+    def __init__(self, args, net, G_data):
+        self.args = args
+        self.net = net
+        self.feat_dim = G_data.feat_dim
+        self.fold_idx = G_data.fold_idx
+        self.init(args, G_data.train_gs, G_data.test_gs)
+        if torch.cuda.is_available():
+            self.net.cuda()
+
+    def init(self, args, train_gs, test_gs):
+        print('#train: %d, #test: %d' % (len(train_gs), len(test_gs)))
+        train_data = GraphData(train_gs, self.feat_dim)
+        test_data = GraphData(test_gs, self.feat_dim)
+        self.train_d = train_data.loader(self.args.batch, True)
+        self.test_d = test_data.loader(self.args.batch, False)
+        self.optimizer = optim.Adam(
+            self.net.parameters(), lr=self.args.lr, amsgrad=True,
+            weight_decay=0.0008)
+
+    def to_cuda(self, gs):
+        if torch.cuda.is_available():
+            if type(gs) == list:
+                return [g.cuda() for g in gs]
+            return gs.cuda()
+        return gs
+
+    def run_epoch(self, epoch, data, model, optimizer):
+        losses, accs, n_samples = [], [], 0
+        for batch in tqdm(data, desc=str(epoch), unit='b'):
+            cur_len, gs, hs, ys = batch
+            gs, hs, ys = map(self.to_cuda, [gs, hs, ys])
+            loss, acc = model(gs, hs, ys)
+            losses.append(loss*cur_len)
+            accs.append(acc*cur_len)
+            n_samples += cur_len
+            if optimizer is not None:
+                optimizer.zero_grad()
+                loss.backward()
+                optimizer.step()
+
+        avg_loss, avg_acc = sum(losses) / n_samples, sum(accs) / n_samples
+        return avg_loss.item(), avg_acc.item()
+
+    def train(self):
+        max_acc = 0.0
+        train_str = 'Train epoch %d: loss %.5f acc %.5f'
+        test_str = 'Test epoch %d: loss %.5f acc %.5f max %.5f'
+        line_str = '%d:\t%.5f\n'
+        for e_id in range(self.args.num_epochs):
+            self.net.train()
+            loss, acc = self.run_epoch(
+                e_id, self.train_d, self.net, self.optimizer)
+            print(train_str % (e_id, loss, acc))
+
+            with torch.no_grad():
+                self.net.eval()
+                loss, acc = self.run_epoch(e_id, self.test_d, self.net, None)
+            max_acc = max(max_acc, acc)
+            print(test_str % (e_id, loss, acc, max_acc))
+
+        with open(self.args.acc_file, 'a+') as f:
+            f.write(line_str % (self.fold_idx, max_acc))

GraphUNets/src/utils/data_loader.py

@@ -0,0 +1,101 @@
+import torch
+from tqdm import tqdm
+import networkx as nx
+import numpy as np
+import torch.nn.functional as F
+from sklearn.model_selection import StratifiedKFold
+from functools import partial
+
+
+class G_data(object):
+    def __init__(self, num_class, feat_dim, g_list):
+        self.num_class = num_class
+        self.feat_dim = feat_dim
+        self.g_list = g_list
+        self.sep_data()
+
+    def sep_data(self, seed=0):
+        skf = StratifiedKFold(n_splits=10, shuffle=True, random_state=seed)
+        labels = [g.label for g in self.g_list]
+        self.idx_list = list(skf.split(np.zeros(len(labels)), labels))
+
+    def use_fold_data(self, fold_idx):
+        self.fold_idx = fold_idx+1
+        train_idx, test_idx = self.idx_list[fold_idx]
+        self.train_gs = [self.g_list[i] for i in train_idx]
+        self.test_gs = [self.g_list[i] for i in test_idx]
+
+
+class FileLoader(object):
+    def __init__(self, args):
+        self.args = args
+
+    def line_genor(self, lines):
+        for line in lines:
+            yield line
+
+    def gen_graph(self, f, i, label_dict, feat_dict, deg_as_tag):
+        row = next(f).strip().split()
+        n, label = [int(w) for w in row]
+        if label not in label_dict:
+            label_dict[label] = len(label_dict)
+        g = nx.Graph()
+        g.add_nodes_from(list(range(n)))
+        node_tags = []
+        for j in range(n):
+            row = next(f).strip().split()
+            tmp = int(row[1]) + 2
+            row = [int(w) for w in row[:tmp]]
+            if row[0] not in feat_dict:
+                feat_dict[row[0]] = len(feat_dict)
+            for k in range(2, len(row)):
+                if j != row[k]:
+                    g.add_edge(j, row[k])
+            if len(row) > 2:
+                node_tags.append(feat_dict[row[0]])
+        g.label = label
+        g.remove_nodes_from(list(nx.isolates(g)))
+        if deg_as_tag:
+            g.node_tags = list(dict(g.degree).values())
+        else:
+            g.node_tags = node_tags
+        return g
+
+    def process_g(self, label_dict, tag2index, tagset, g):
+        g.label = label_dict[g.label]
+        g.feas = torch.tensor([tag2index[tag] for tag in g.node_tags])
+        g.feas = F.one_hot(g.feas, len(tagset))
+        A = torch.FloatTensor(nx.to_numpy_matrix(g))
+        g.A = A + torch.eye(g.number_of_nodes())
+        return g
+
+    def load_data(self):
+        args = self.args
+        print('loading data ...')
+        g_list = []
+        label_dict = {}
+        feat_dict = {}
+
+        with open('data/%s/%s.txt' % (args.data, args.data), 'r') as f:
+            lines = f.readlines()
+        f = self.line_genor(lines)
+        n_g = int(next(f).strip())
+        for i in tqdm(range(n_g), desc="Create graph", unit='graphs'):
+            g = self.gen_graph(f, i, label_dict, feat_dict, args.deg_as_tag)
+            g_list.append(g)
+
+        tagset = set([])
+        for g in g_list:
+            tagset = tagset.union(set(g.node_tags))
+        tagset = list(tagset)
+        tag2index = {tagset[i]: i for i in range(len(tagset))}
+
+        f_n = partial(self.process_g, label_dict, tag2index, tagset)
+        new_g_list = []
+        for g in tqdm(g_list, desc="Process graph", unit='graphs'):
+            new_g_list.append(f_n(g))
+        num_class = len(label_dict)
+        feat_dim = len(tagset)
+
+        print('# classes: %d' % num_class, '# maximum node tag: %d' % feat_dim)
+        return G_data(num_class, feat_dim, new_g_list)

GraphUNets/src/utils/dataset.py

@@ -0,0 +1,45 @@
+import random
+import torch
+
+
+class GraphData(object):
+
+    def __init__(self, data, feat_dim):
+        super(GraphData, self).__init__()
+        self.data = data
+        self.feat_dim = feat_dim
+        self.idx = list(range(len(data)))
+        self.pos = 0
+
+    def __reset__(self):
+        self.pos = 0
+        if self.shuffle:
+            random.shuffle(self.idx)
+
+    def __len__(self):
+        return len(self.data) // self.batch + 1
+
+    def __getitem__(self, idx):
+        g = self.data[idx]
+        return g.A, g.feas.float(), g.label
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        if self.pos >= len(self.data):
+            self.__reset__()
+            raise StopIteration
+
+        cur_idx = self.idx[self.pos: self.pos+self.batch]
+        data = [self.__getitem__(idx) for idx in cur_idx]
+        self.pos += len(cur_idx)
+        gs, hs, labels = map(list, zip(*data))
+        return len(gs), gs, hs, torch.LongTensor(labels)
+
+    def loader(self, batch, shuffle, *args):
+        self.batch = batch
+        self.shuffle = shuffle
+        if shuffle:
+            random.shuffle(self.idx)
+        return self

GraphUNets/src/utils/ops.py

@@ -0,0 +1,145 @@
+import torch
+import torch.nn as nn
+import numpy as np
+
+
+class GraphUnet(nn.Module):
+
+    def __init__(self, ks, in_dim, out_dim, dim, act, drop_p):
+        super(GraphUnet, self).__init__()
+        self.ks = ks
+        self.bottom_gcn = GCN(dim, dim, act, drop_p)
+        self.down_gcns = nn.ModuleList()
+        self.up_gcns = nn.ModuleList()
+        self.pools = nn.ModuleList()
+        self.unpools = nn.ModuleList()
+        self.l_n = len(ks)
+        for i in range(self.l_n):
+            self.down_gcns.append(GCN(dim, dim, act, drop_p))
+            self.up_gcns.append(GCN(dim, dim, act, drop_p))
+            self.pools.append(Pool(ks[i], dim, drop_p))
+            self.unpools.append(Unpool(dim, dim, drop_p))
+
+    def forward(self, g, h):
+        adj_ms = []
+        indices_list = []
+        down_outs = []
+        hs = []
+        org_h = h
+        for i in range(self.l_n):
+            h = self.down_gcns[i](g, h)
+            adj_ms.append(g)
+            down_outs.append(h)
+            g, h, idx = self.pools[i](g, h)
+            indices_list.append(idx)
+        h = self.bottom_gcn(g, h)
+        for i in range(self.l_n):
+            up_idx = self.l_n - i - 1
+            g, idx = adj_ms[up_idx], indices_list[up_idx]
+            g, h = self.unpools[i](g, h, down_outs[up_idx], idx)
+            h = self.up_gcns[i](g, h)
+            h = h.add(down_outs[up_idx])
+            hs.append(h)
+        h = h.add(org_h)
+        hs.append(h)
+        return hs
+
+
+class GCN(nn.Module):
+
+    def __init__(self, in_dim, out_dim, act, p):
+        super(GCN, self).__init__()
+        self.proj = nn.Linear(in_dim, out_dim)
+        self.act = act
+        self.drop = nn.Dropout(p=p) if p > 0.0 else nn.Identity()
+
+    def forward(self, g, h):
+        h = self.drop(h)
+        h = torch.matmul(g, h)
+        h = self.proj(h)
+        h = self.act(h)
+        return h
+
+
+class Pool(nn.Module):
+
+    def __init__(self, k, in_dim, p):
+        super(Pool, self).__init__()
+        self.k = k
+        self.sigmoid = nn.Sigmoid()
+        self.proj = nn.Linear(in_dim, 1)
+        self.drop = nn.Dropout(p=p) if p > 0 else nn.Identity()
+
+    def forward(self, g, h):
+        Z = self.drop(h)
+        weights = self.proj(Z).squeeze()
+        scores = self.sigmoid(weights)
+        return top_k_graph(scores, g, h, self.k)
+
+
+class Unpool(nn.Module):
+
+    def __init__(self, *args):
+        super(Unpool, self).__init__()
+
+    def forward(self, g, h, pre_h, idx):
+        new_h = h.new_zeros([g.shape[0], h.shape[1]])
+        new_h[idx] = h
+        return g, new_h
+
+
+def top_k_graph(scores, g, h, k):
+    num_nodes = g.shape[0]
+    values, idx = torch.topk(scores, max(2, int(k*num_nodes)))
+    new_h = h[idx, :]
+    values = torch.unsqueeze(values, -1)
+    new_h = torch.mul(new_h, values)
+    un_g = g.bool().float()
+    un_g = torch.matmul(un_g, un_g).bool().float()
+    un_g = un_g[idx, :]
+    un_g = un_g[:, idx]
+    g = norm_g(un_g)
+    return g, new_h, idx
+
+
+def norm_g(g):
+    degrees = torch.sum(g, 1)
+    g = g / degrees
+    return g
+
+
+class Initializer(object):
+
+    @classmethod
+    def _glorot_uniform(cls, w):
+        if len(w.size()) == 2:
+            fan_in, fan_out = w.size()
+        elif len(w.size()) == 3:
+            fan_in = w.size()[1] * w.size()[2]
+            fan_out = w.size()[0] * w.size()[2]
+        else:
+            fan_in = np.prod(w.size())
+            fan_out = np.prod(w.size())
+        limit = np.sqrt(6.0 / (fan_in + fan_out))
+        w.uniform_(-limit, limit)
+
+    @classmethod
+    def _param_init(cls, m):
+        if isinstance(m, nn.parameter.Parameter):
+            cls._glorot_uniform(m.data)
+        elif isinstance(m, nn.Linear):
+            m.bias.data.zero_()
+            cls._glorot_uniform(m.weight.data)
+
+    @classmethod
+    def weights_init(cls, m):
+        for p in m.modules():
+            if isinstance(p, nn.ParameterList):
+                for pp in p:
+                    cls._param_init(pp)
+            else:
+                cls._param_init(p)
+
+        for name, p in m.named_parameters():
+            if '.' not in name:
+                cls._param_init(p)

cora/README

@@ -0,0 +1,29 @@
+This directory contains the a selection of the Cora dataset (www.research.whizbang.com/data).
+
+The Cora dataset consists of Machine Learning papers. These papers are classified into one of the following seven classes:
+		Case_Based
+		Genetic_Algorithms
+		Neural_Networks
+		Probabilistic_Methods
+		Reinforcement_Learning
+		Rule_Learning
+		Theory
+
+The papers were selected in a way such that in the final corpus every paper cites or is cited by atleast one other paper. There are 2708 papers in the whole corpus. 
+
+After stemming and removing stopwords we were left with a vocabulary of size 1433 unique words. All words with document frequency less than 10 were removed.
+
+
+THE DIRECTORY CONTAINS TWO FILES:
+
+The .content file contains descriptions of the papers in the following format:
+
+		<paper_id> <word_attributes>+ <class_label>
+
+The first entry in each line contains the unique string ID of the paper followed by binary values indicating whether each word in the vocabulary is present (indicated by 1) or absent (indicated by 0) in the paper. Finally, the last entry in the line contains the class label of the paper.
+
+The .cites file contains the citation graph of the corpus. Each line describes a link in the following format:
+
+		<ID of cited paper> <ID of citing paper>
+
+Each line contains two paper IDs. The first entry is the ID of the paper being cited and the second ID stands for the paper which contains the citation. The direction of the link is from right to left. If a line is represented by "paper1 paper2" then the link is "paper2->paper1". 

input

@@ -0,0 +1,2 @@
+5 3 4
+0 2 3

requirements.txt

@@ -0,0 +1,4 @@
+torch
+torch-scatter
+torch-sparse
+torch-geometric

results.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,51.152,99112.473,38.589
+dens_x0.5,10000,761,2,50.359,99112.473,38.589
+dens_x0.5,10000,761,3,50.185,99112.473,38.589
+dens_x0.5,12000,1077,1,61.595,122827.840,47.823
+dens_x0.5,12000,1077,2,61.517,122827.840,47.823
+dens_x0.5,12000,1077,3,60.927,122827.840,47.823
+dens_x0.5,14000,1455,1,69.076,147545.989,57.447
+dens_x0.5,14000,1455,2,80.602,147545.989,57.447
+dens_x0.5,14000,1455,3,69.987,147545.989,57.447
+dens_x0.5,16000,1937,1,80.775,173636.330,67.605
+dens_x0.5,16000,1937,2,81.025,173636.330,67.605
+dens_x0.5,16000,1937,3,87.149,173636.330,67.605
+dens_x0.5,19000,2715,1,97.314,213940.398,83.297
+dens_x0.5,19000,2715,2,96.800,213940.398,83.297
+dens_x0.5,19000,2715,3,99.001,213940.398,83.297
+dens_x0.5,22000,3611,1,114.149,256079.209,99.704
+dens_x0.5,22000,3611,2,107.349,256079.209,99.704
+dens_x0.5,22000,3611,3,128.944,256079.209,99.704
+dens_x0.5,26000,5194,1,169.081,317113.582,123.468
+dens_x0.5,26000,5194,2,124.848,317113.582,123.468
+dens_x0.5,26000,5194,3,126.180,317113.582,123.468
+dens_x0.5,30000,6763,1,137.270,378988.027,147.558
+dens_x0.5,30000,6763,2,174.605,378988.027,147.558
+dens_x0.5,30000,6763,3,347.805,378988.027,147.558
+dens_x0.5,36000,9660,1,170.769,479031.581,186.510
+dens_x0.5,36000,9660,2,167.815,479031.581,186.510
+dens_x0.5,36000,9660,3,157.285,479031.581,186.510
+dens_x0.5,42000,13346,1,208.810,589181.686,229.397
+dens_x0.5,42000,13346,2,191.181,589181.686,229.397
+dens_x0.5,42000,13346,3,188.304,589181.686,229.397
+dens_x0.5,49000,17900,1,274.029,722491.606,281.301
+dens_x0.5,49000,17900,2,262.442,722491.606,281.301
+dens_x0.5,49000,17900,3,268.845,722491.606,281.301
+dens_x0.5,57000,24363,1,312.798,890990.473,346.906
+dens_x0.5,57000,24363,2,301.761,890990.473,346.906
+dens_x0.5,57000,24363,3,327.824,890990.473,346.906
+dens_x0.5,67000,33829,1,424.048,1120457.009,436.248
+dens_x0.5,67000,33829,2,375.659,1120457.009,436.248
+dens_x0.5,67000,33829,3,375.023,1120457.009,436.248
+dens_x0.5,79000,46917,1,447.988,1419991.587,552.871
+dens_x0.5,79000,46917,2,543.833,1419991.587,552.871
+dens_x0.5,79000,46917,3,490.562,1419991.587,552.871
+dens_x0.5,92000,63864,1,605.730,1781454.424,693.606
+dens_x0.5,92000,63864,2,597.292,1781454.424,693.606
+dens_x0.5,92000,63864,3,650.577,1781454.424,693.606
+dens_x0.5,108000,87350,1,608.491,2264084.329,881.518
+dens_x0.5,108000,87350,2,568.816,2264084.329,881.518
+dens_x0.5,108000,87350,3,605.038,2264084.329,881.518
+dens_x0.5,127000,120912,1,701.876,2913447.536,1134.346
+dens_x0.5,127000,120912,2,719.486,2913447.536,1134.346
+dens_x0.5,127000,120912,3,717.672,2913447.536,1134.346
+dens_x0.5,149000,166365,1,875.466,3756533.770,1462.600
+dens_x0.5,149000,166365,2,841.735,3756533.770,1462.600
+dens_x0.5,149000,166365,3,986.523,3756533.770,1462.600
+dens_x0.5,174000,226551,1,1247.354,4833373.044,1881.866
+dens_x0.5,174000,226551,2,1151.140,4833373.044,1881.866
+dens_x0.5,174000,226551,3,1234.685,4833373.044,1881.866
+dens_x0.5,204000,312325,1,1410.429,6312525.050,2457.772
+dens_x0.5,204000,312325,2,1327.941,6312525.050,2457.772
+dens_x0.5,204000,312325,3,1383.628,6312525.050,2457.772
+dens_x0.5,240000,431142,1,1747.568,8314371.662,3237.187
+dens_x0.5,240000,431142,2,1674.865,8314371.662,3237.187
+dens_x0.5,240000,431142,3,1688.159,8314371.662,3237.187
+dens_x0.5,281000,589828,1,2166.406,10925503.834,4253.828
+dens_x0.5,281000,589828,2,3446.245,10925503.834,4253.828
+dens_x0.5,281000,589828,3,2455.903,10925503.834,4253.828
+dens_x0.5,329000,811844,1,3235.985,14493068.872,5642.853
+dens_x0.5,329000,811844,2,3914.621,14493068.872,5642.853
+dens_x0.5,329000,811844,3,3149.491,14493068.872,5642.853
+dens_x0.5,386000,1115568,1,4208.325,19315559.086,7520.482
+dens_x0.5,386000,1115568,2,4153.722,19315559.086,7520.482
+dens_x0.5,386000,1115568,3,4030.253,19315559.086,7520.482
+dens_x0.5,452000,1531464,1,5754.485,25827552.428,10055.916
+dens_x0.5,452000,1531464,2,5686.995,25827552.428,10055.916
+dens_x0.5,452000,1531464,3,5806.431,25827552.428,10055.916
+dens_x0.5,530000,2103874,1,8092.970,34716124.680,13516.668
+dens_x0.5,530000,2103874,2,8378.748,34716124.680,13516.668
+dens_x0.5,530000,2103874,3,8476.747,34716124.680,13516.668
+dens_x0.5,621000,2887681,1,12003.507,46802598.872,18222.517
+dens_x0.5,621000,2887681,2,11475.289,46802598.872,18222.517
+dens_x0.5,621000,2887681,3,11643.507,46802598.872,18222.517
+dens_x0.5,728000,3971764,1,15740.247,63437679.196,24699.359
+dens_x0.5,728000,3971764,2,15861.904,63437679.196,24699.359
+dens_x0.5,728000,3971764,3,16104.897,63437679.196,24699.359
+dens_x0.5,853000,5448500,1,23924.937,86056528.179,33505.972
+dens_x0.5,853000,5448500,2,23993.406,86056528.179,33505.972
+dens_x0.5,853000,5448500,3,25817.687,86056528.179,33505.972
+dens_x0.5,1000000,7489974,1,33333.025,117293325.434,45667.970
+dens_x0.5,1000000,7489974,2,32787.127,117293325.434,45667.970
+dens_x0.5,1000000,7489974,3,35885.585,117293325.434,45667.970
+dens_x1.0,10000,1550,1,58.206,106379.431,41.419
+dens_x1.0,10000,1550,2,58.889,106379.431,41.419
+dens_x1.0,10000,1550,3,59.592,106379.431,41.419
+dens_x1.0,12000,2135,1,68.803,132765.276,51.692
+dens_x1.0,12000,2135,2,69.910,132765.276,51.692
+dens_x1.0,12000,2135,3,69.644,132765.276,51.692
+dens_x1.0,14000,2963,1,94.906,161942.582,63.052
+dens_x1.0,14000,2963,2,80.102,161942.582,63.052
+dens_x1.0,14000,2963,3,84.479,161942.582,63.052
+dens_x1.0,16000,3912,1,94.907,192755.010,75.049
+dens_x1.0,16000,3912,2,97.008,192755.010,75.049
+dens_x1.0,16000,3912,3,94.134,192755.010,75.049
+dens_x1.0,19000,5413,1,108.859,240521.618,93.647
+dens_x1.0,19000,5413,2,108.454,240521.618,93.647
+dens_x1.0,19000,5413,3,108.726,240521.618,93.647
+dens_x1.0,22000,7290,1,116.383,292864.786,114.026
+dens_x1.0,22000,7290,2,119.034,292864.786,114.026
+dens_x1.0,22000,7290,3,124.547,292864.786,114.026
+dens_x1.0,26000,10078,1,134.102,366763.602,142.799
+dens_x1.0,26000,10078,2,138.847,366763.602,142.799
+dens_x1.0,26000,10078,3,145.234,366763.602,142.799
+dens_x1.0,30000,13336,1,205.487,446748.773,173.941
+dens_x1.0,30000,13336,2,201.550,446748.773,173.941
+dens_x1.0,30000,13336,3,185.775,446748.773,173.941
+dens_x1.0,36000,19373,1,195.098,580933.327,226.185
+dens_x1.0,36000,19373,2,221.406,580933.327,226.185
+dens_x1.0,36000,19373,3,229.693,580933.327,226.185
+dens_x1.0,42000,26057,1,268.682,724495.682,282.081
+dens_x1.0,42000,26057,2,235.697,724495.682,282.081
+dens_x1.0,42000,26057,3,260.327,724495.682,282.081
+dens_x1.0,49000,35903,1,319.756,916916.365,357.000
+dens_x1.0,49000,35903,2,247.543,916916.365,357.000
+dens_x1.0,49000,35903,3,242.526,916916.365,357.000
+dens_x1.0,57000,48648,1,397.064,1156930.810,450.449
+dens_x1.0,57000,48648,2,380.093,1156930.810,450.449
+dens_x1.0,57000,48648,3,346.247,1156930.810,450.449
+dens_x1.0,67000,67188,1,359.800,1491157.159,580.580
+dens_x1.0,67000,67188,2,361.464,1491157.159,580.580
+dens_x1.0,67000,67188,3,369.751,1491157.159,580.580
+dens_x1.0,79000,93552,1,469.624,1945903.955,757.635
+dens_x1.0,79000,93552,2,523.395,1945903.955,757.635
+dens_x1.0,79000,93552,3,451.623,1945903.955,757.635
+dens_x1.0,92000,126776,1,557.575,2500509.887,973.570
+dens_x1.0,92000,126776,2,555.152,2500509.887,973.570
+dens_x1.0,92000,126776,3,559.637,2500509.887,973.570
+dens_x1.0,108000,175020,1,693.661,3280169.679,1277.129
+dens_x1.0,108000,175020,2,694.646,3280169.679,1277.129
+dens_x1.0,108000,175020,3,688.850,3280169.679,1277.129
+dens_x1.0,127000,241838,1,887.291,4334562.918,1687.655
+dens_x1.0,127000,241838,2,917.991,4334562.918,1687.655
+dens_x1.0,127000,241838,3,887.282,4334562.918,1687.655
+dens_x1.0,149000,333094,1,1218.530,5742559.864,2235.857
+dens_x1.0,149000,333094,2,1262.687,5742559.864,2235.857
+dens_x1.0,149000,333094,3,1221.463,5742559.864,2235.857
+dens_x1.0,174000,454003,1,1653.458,7577993.244,2950.480
+dens_x1.0,174000,454003,2,1589.535,7577993.244,2950.480
+dens_x1.0,174000,454003,3,1597.942,7577993.244,2950.480
+dens_x1.0,204000,623499,1,2081.263,10116899.562,3938.999
+dens_x1.0,204000,623499,2,2064.348,10116899.562,3938.999
+dens_x1.0,204000,623499,3,2011.870,10116899.562,3938.999
+dens_x1.0,240000,862431,1,2886.728,13657349.809,5317.467
+dens_x1.0,240000,862431,2,2914.305,13657349.809,5317.467
+dens_x1.0,240000,862431,3,2874.904,13657349.809,5317.467
+dens_x1.0,281000,1182168,1,4453.264,18357066.601,7147.295
+dens_x1.0,281000,1182168,2,4139.544,18357066.601,7147.295
+dens_x1.0,281000,1182168,3,4573.999,18357066.601,7147.295
+dens_x1.0,329000,1623777,1,5959.977,24807713.897,9658.844
+dens_x1.0,329000,1623777,2,6951.901,24807713.897,9658.844
+dens_x1.0,329000,1623777,3,5727.519,24807713.897,9658.844
+dens_x1.0,386000,2232671,1,8057.699,33685517.024,13115.402
+dens_x1.0,386000,2232671,2,8162.875,33685517.024,13115.402
+dens_x1.0,386000,2232671,3,8187.878,33685517.024,13115.402
+dens_x1.0,452000,3055874,1,11484.033,45677561.904,17784.486
+dens_x1.0,452000,3055874,2,11583.497,45677561.904,17784.486
+dens_x1.0,452000,3055874,3,11713.283,45677561.904,17784.486
+dens_x1.0,530000,4208065,1,16968.262,62450692.510,24315.078
+dens_x1.0,530000,4208065,2,15686.033,62450692.510,24315.078
+dens_x1.0,530000,4208065,3,16661.082,62450692.510,24315.078
+dens_x1.0,621000,5774507,1,24345.921,85310220.195,33215.398
+dens_x1.0,621000,5774507,2,24451.492,85310220.195,33215.398
+dens_x1.0,621000,5774507,3,23900.906,85310220.195,33215.398
+dens_x1.0,728000,7935888,1,36299.762,116945648.236,45532.602
+dens_x1.0,728000,7935888,2,33059.393,116945648.236,45532.602
+dens_x1.0,728000,7935888,3,37098.302,116945648.236,45532.602
+dens_x1.0,853000,10900416,1,53083.830,160510699.866,62494.585
+dens_x1.0,853000,10900416,2,53282.730,160510699.866,62494.585
+dens_x1.0,853000,10900416,3,52312.206,160510699.866,62494.585
+dens_x1.0,1000000,14986435,1,77987.779,220860761.527,85991.786
+dens_x1.0,1000000,14986435,2,78617.841,220860761.527,85991.786
+dens_x1.0,1000000,14986435,3,79625.105,220860761.527,85991.786
+dens_x2.0,10000,3007,1,72.111,119798.897,46.644
+dens_x2.0,10000,3007,2,70.572,119798.897,46.644
+dens_x2.0,10000,3007,3,69.460,119798.897,46.644
+dens_x2.0,12000,4365,1,88.829,153710.912,59.847
+dens_x2.0,12000,4365,2,87.598,153710.912,59.847
+dens_x2.0,12000,4365,3,92.566,153710.912,59.847
+dens_x2.0,14000,5830,1,98.751,189313.294,73.709
+dens_x2.0,14000,5830,2,99.427,189313.294,73.709
+dens_x2.0,14000,5830,3,97.017,189313.294,73.709
+dens_x2.0,16000,7624,1,113.441,228688.447,89.039
+dens_x2.0,16000,7624,2,114.051,228688.447,89.039
+dens_x2.0,16000,7624,3,111.923,228688.447,89.039
+dens_x2.0,19000,10837,1,140.996,293959.920,114.453
+dens_x2.0,19000,10837,2,131.449,293959.920,114.453
+dens_x2.0,19000,10837,3,137.347,293959.920,114.453
+dens_x2.0,22000,14556,1,165.281,365516.050,142.313
+dens_x2.0,22000,14556,2,146.564,365516.050,142.313
+dens_x2.0,22000,14556,3,144.894,365516.050,142.313
+dens_x2.0,26000,20395,1,181.262,471644.695,183.634
+dens_x2.0,26000,20395,2,198.833,471644.695,183.634
+dens_x2.0,26000,20395,3,207.951,471644.695,183.634
+dens_x2.0,30000,26790,1,225.380,585445.422,227.942
+dens_x2.0,30000,26790,2,212.692,585445.422,227.942
+dens_x2.0,30000,26790,3,220.381,585445.422,227.942
+dens_x2.0,36000,38800,1,291.967,784747.311,305.540
+dens_x2.0,36000,38800,2,287.036,784747.311,305.540
+dens_x2.0,36000,38800,3,285.022,784747.311,305.540
+dens_x2.0,42000,53391,1,306.692,1015477.726,395.375
+dens_x2.0,42000,53391,2,317.915,1015477.726,395.375
+dens_x2.0,42000,53391,3,324.443,1015477.726,395.375
+dens_x2.0,49000,71945,1,431.617,1306154.668,508.549
+dens_x2.0,49000,71945,2,411.852,1306154.668,508.549
+dens_x2.0,49000,71945,3,442.893,1306154.668,508.549
+dens_x2.0,57000,97182,1,444.690,1688417.255,657.383
+dens_x2.0,57000,97182,2,414.667,1688417.255,657.383
+dens_x2.0,57000,97182,3,413.077,1688417.255,657.383
+dens_x2.0,67000,134082,1,528.155,2234513.248,870.004
+dens_x2.0,67000,134082,2,532.767,2234513.248,870.004
+dens_x2.0,67000,134082,3,528.469,2234513.248,870.004
+dens_x2.0,79000,186876,1,676.211,2998337.660,1167.398
+dens_x2.0,79000,186876,2,664.701,2998337.660,1167.398
+dens_x2.0,79000,186876,3,697.816,2998337.660,1167.398
+dens_x2.0,92000,253371,1,849.705,3947432.991,1536.927
+dens_x2.0,92000,253371,2,904.967,3947432.991,1536.927
+dens_x2.0,92000,253371,3,870.314,3947432.991,1536.927
+dens_x2.0,108000,349108,1,1181.033,5297829.841,2062.702
+dens_x2.0,108000,349108,2,1179.125,5297829.841,2062.702
+dens_x2.0,108000,349108,3,1169.062,5297829.841,2062.702
+dens_x2.0,127000,484181,1,1568.306,7182563.887,2796.520
+dens_x2.0,127000,484181,2,1527.148,7182563.887,2796.520
+dens_x2.0,127000,484181,3,1527.026,7182563.887,2796.520
+dens_x2.0,149000,666491,1,2130.278,9713885.437,3782.086
+dens_x2.0,149000,666491,2,2095.352,9713885.437,3782.086
+dens_x2.0,149000,666491,3,2116.762,9713885.437,3782.086
+dens_x2.0,174000,907603,1,2955.051,13051498.522,5081.580
+dens_x2.0,174000,907603,2,2968.651,13051498.522,5081.580
+dens_x2.0,174000,907603,3,2901.389,13051498.522,5081.580
+dens_x2.0,204000,1247723,1,4011.758,17748584.329,6910.383
+dens_x2.0,204000,1247723,2,3967.668,17748584.329,6910.383
+dens_x2.0,204000,1247723,3,4002.986,17748584.329,6910.383
+dens_x2.0,240000,1725541,1,5827.568,24349896.729,9480.594
+dens_x2.0,240000,1725541,2,5824.487,24349896.729,9480.594
+dens_x2.0,240000,1725541,3,5785.284,24349896.729,9480.594
+dens_x2.0,281000,2365586,1,8196.369,33204358.944,12928.064
+dens_x2.0,281000,2365586,2,8396.722,33204358.944,12928.064
+dens_x2.0,281000,2365586,3,8292.723,33204358.944,12928.064
+dens_x2.0,329000,3242756,1,12064.337,45374920.412,17666.653
+dens_x2.0,329000,3242756,2,12045.907,45374920.412,17666.653
+dens_x2.0,329000,3242756,3,12026.978,45374920.412,17666.653
+dens_x2.0,386000,4461354,1,17662.636,62354387.279,24277.581
+dens_x2.0,386000,4461354,2,17484.608,62354387.279,24277.581
+dens_x2.0,386000,4461354,3,17401.132,62354387.279,24277.581
+dens_x2.0,452000,6119886,1,31461.157,85575402.535,33318.647
+dens_x2.0,452000,6119886,2,27648.193,85575402.535,33318.647
+dens_x2.0,452000,6119886,3,25986.776,85575402.535,33318.647
+dens_x2.0,530000,8413791,1,39027.499,117884820.410,45898.267
+dens_x2.0,530000,8413791,2,37862.747,117884820.410,45898.267
+dens_x2.0,530000,8413791,3,37738.933,117884820.410,45898.267
+dens_x2.0,621000,11552079,1,56434.022,162377752.059,63221.519
+dens_x2.0,621000,11552079,2,55647.184,162377752.059,63221.519
+dens_x2.0,621000,11552079,3,55382.089,162377752.059,63221.519
+dens_x2.0,728000,15870417,1,84297.240,224046367.608,87232.097
+dens_x2.0,728000,15870417,2,84495.719,224046367.608,87232.097
+dens_x2.0,728000,15870417,3,84830.940,224046367.608,87232.097
+dens_x2.0,853000,21800415,1,127846.865,309366697.818,120451.431
+dens_x2.0,853000,21800415,2,128578.916,309366697.818,120451.431
+dens_x2.0,853000,21800415,3,127293.075,309366697.818,120451.431
+dens_x2.0,1000000,29958277,1,193809.065,427704402.750,166526.028
+dens_x2.0,1000000,29958277,2,192667.212,427704402.750,166526.028
+dens_x2.0,1000000,29958277,3,194000.905,427704402.750,166526.028

results_clique.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,38.276,99112.473,18.938
+dens_x0.5,10000,761,2,44.437,99112.473,18.938
+dens_x0.5,10000,761,3,47.050,99112.473,18.938
+dens_x0.5,12000,1077,1,61.835,122827.840,23.469
+dens_x0.5,12000,1077,2,56.548,122827.840,23.469
+dens_x0.5,12000,1077,3,57.742,122827.840,23.469
+dens_x0.5,14000,1455,1,56.974,147545.989,28.192
+dens_x0.5,14000,1455,2,72.837,147545.989,28.192
+dens_x0.5,14000,1455,3,53.347,147545.989,28.192
+dens_x0.5,16000,1937,1,78.537,173636.330,33.178
+dens_x0.5,16000,1937,2,85.969,173636.330,33.178
+dens_x0.5,16000,1937,3,86.131,173636.330,33.178
+dens_x0.5,19000,2715,1,98.627,213940.398,40.879
+dens_x0.5,19000,2715,2,116.253,213940.398,40.879
+dens_x0.5,19000,2715,3,108.861,213940.398,40.879
+dens_x0.5,22000,3611,1,130.744,256079.209,48.930
+dens_x0.5,22000,3611,2,89.643,256079.209,48.930
+dens_x0.5,22000,3611,3,131.201,256079.209,48.930
+dens_x0.5,26000,5194,1,120.353,317113.582,60.593
+dens_x0.5,26000,5194,2,145.893,317113.582,60.593
+dens_x0.5,26000,5194,3,144.641,317113.582,60.593
+dens_x0.5,30000,6763,1,193.243,378988.027,72.415
+dens_x0.5,30000,6763,2,138.053,378988.027,72.415
+dens_x0.5,30000,6763,3,159.792,378988.027,72.415
+dens_x0.5,36000,9660,1,188.103,479031.581,91.531
+dens_x0.5,36000,9660,2,181.788,479031.581,91.531
+dens_x0.5,36000,9660,3,183.152,479031.581,91.531
+dens_x0.5,42000,13346,1,151.805,589181.686,112.578
+dens_x0.5,42000,13346,2,237.285,589181.686,112.578
+dens_x0.5,42000,13346,3,241.080,589181.686,112.578
+dens_x0.5,49000,17900,1,225.210,722491.606,138.051
+dens_x0.5,49000,17900,2,256.325,722491.606,138.051
+dens_x0.5,49000,17900,3,254.573,722491.606,138.051
+dens_x0.5,57000,24363,1,205.892,890990.473,170.247
+dens_x0.5,57000,24363,2,218.569,890990.473,170.247
+dens_x0.5,57000,24363,3,193.055,890990.473,170.247
+dens_x0.5,67000,33829,1,312.891,1120457.009,214.092
+dens_x0.5,67000,33829,2,332.210,1120457.009,214.092
+dens_x0.5,67000,33829,3,246.322,1120457.009,214.092
+dens_x0.5,79000,46917,1,302.178,1419991.587,271.326
+dens_x0.5,79000,46917,2,362.654,1419991.587,271.326
+dens_x0.5,79000,46917,3,388.967,1419991.587,271.326
+dens_x0.5,92000,63864,1,405.253,1781454.424,340.392
+dens_x0.5,92000,63864,2,428.211,1781454.424,340.392
+dens_x0.5,92000,63864,3,424.595,1781454.424,340.392
+dens_x0.5,108000,87350,1,531.480,2264084.329,432.611
+dens_x0.5,108000,87350,2,456.242,2264084.329,432.611
+dens_x0.5,108000,87350,3,480.178,2264084.329,432.611
+dens_x0.5,127000,120912,1,584.835,2913447.536,556.689
+dens_x0.5,127000,120912,2,574.881,2913447.536,556.689
+dens_x0.5,127000,120912,3,591.637,2913447.536,556.689
+dens_x0.5,149000,166365,1,770.336,3756533.770,717.782
+dens_x0.5,149000,166365,2,749.424,3756533.770,717.782
+dens_x0.5,149000,166365,3,758.684,3756533.770,717.782
+dens_x0.5,174000,226551,1,977.870,4833373.044,923.540
+dens_x0.5,174000,226551,2,911.794,4833373.044,923.540
+dens_x0.5,174000,226551,3,1973.549,4833373.044,923.540
+dens_x0.5,204000,312325,1,1322.555,6312525.050,1206.169
+dens_x0.5,204000,312325,2,1448.643,6312525.050,1206.169
+dens_x0.5,204000,312325,3,1015.363,6312525.050,1206.169
+dens_x0.5,240000,431142,1,1260.775,8314371.662,1588.673
+dens_x0.5,240000,431142,2,1571.054,8314371.662,1588.673
+dens_x0.5,240000,431142,3,2325.376,8314371.662,1588.673
+dens_x0.5,281000,589828,1,1716.171,10925503.834,2087.597
+dens_x0.5,281000,589828,2,1772.932,10925503.834,2087.597
+dens_x0.5,281000,589828,3,1770.317,10925503.834,2087.597
+dens_x0.5,329000,811844,1,2497.336,14493068.872,2769.272
+dens_x0.5,329000,811844,2,3011.600,14493068.872,2769.272
+dens_x0.5,329000,811844,3,2262.160,14493068.872,2769.272
+dens_x0.5,386000,1115568,1,3122.086,19315559.086,3690.732
+dens_x0.5,386000,1115568,2,3034.498,19315559.086,3690.732
+dens_x0.5,386000,1115568,3,3167.076,19315559.086,3690.732
+dens_x0.5,452000,1531464,1,4496.003,25827552.428,4935.015
+dens_x0.5,452000,1531464,2,4372.080,25827552.428,4935.015
+dens_x0.5,452000,1531464,3,4231.171,25827552.428,4935.015
+dens_x0.5,530000,2103874,1,6131.828,34716124.680,6633.404
+dens_x0.5,530000,2103874,2,6007.724,34716124.680,6633.404
+dens_x0.5,530000,2103874,3,5781.477,34716124.680,6633.404
+dens_x0.5,621000,2887681,1,7942.139,46802598.872,8942.834
+dens_x0.5,621000,2887681,2,7923.217,46802598.872,8942.834
+dens_x0.5,621000,2887681,3,7941.885,46802598.872,8942.834
+dens_x0.5,728000,3971764,1,11260.380,63437679.196,12121.392
+dens_x0.5,728000,3971764,2,11239.027,63437679.196,12121.392
+dens_x0.5,728000,3971764,3,10927.841,63437679.196,12121.392
+dens_x0.5,853000,5448500,1,15269.468,86056528.179,16443.301
+dens_x0.5,853000,5448500,2,15545.416,86056528.179,16443.301
+dens_x0.5,853000,5448500,3,15453.498,86056528.179,16443.301
+dens_x0.5,1000000,7489974,1,21313.952,117293325.434,22411.891
+dens_x0.5,1000000,7489974,2,21260.455,117293325.434,22411.891
+dens_x0.5,1000000,7489974,3,21363.975,117293325.434,22411.891
+dens_x1.0,10000,1550,1,61.324,106379.431,20.327
+dens_x1.0,10000,1550,2,42.309,106379.431,20.327
+dens_x1.0,10000,1550,3,43.722,106379.431,20.327
+dens_x1.0,12000,2135,1,59.932,132765.276,25.368
+dens_x1.0,12000,2135,2,58.811,132765.276,25.368
+dens_x1.0,12000,2135,3,59.965,132765.276,25.368
+dens_x1.0,14000,2963,1,67.045,161942.582,30.943
+dens_x1.0,14000,2963,2,64.095,161942.582,30.943
+dens_x1.0,14000,2963,3,98.423,161942.582,30.943
+dens_x1.0,16000,3912,1,67.059,192755.010,36.831
+dens_x1.0,16000,3912,2,73.193,192755.010,36.831
+dens_x1.0,16000,3912,3,78.115,192755.010,36.831
+dens_x1.0,19000,5413,1,86.648,240521.618,45.958
+dens_x1.0,19000,5413,2,129.568,240521.618,45.958
+dens_x1.0,19000,5413,3,78.989,240521.618,45.958
+dens_x1.0,22000,7290,1,111.022,292864.786,55.959
+dens_x1.0,22000,7290,2,133.328,292864.786,55.959
+dens_x1.0,22000,7290,3,99.040,292864.786,55.959
+dens_x1.0,26000,10078,1,152.963,366763.602,70.080
+dens_x1.0,26000,10078,2,122.924,366763.602,70.080
+dens_x1.0,26000,10078,3,138.414,366763.602,70.080
+dens_x1.0,30000,13336,1,129.110,446748.773,85.363
+dens_x1.0,30000,13336,2,186.901,446748.773,85.363
+dens_x1.0,30000,13336,3,145.582,446748.773,85.363
+dens_x1.0,36000,19373,1,157.848,580933.327,111.002
+dens_x1.0,36000,19373,2,214.319,580933.327,111.002
+dens_x1.0,36000,19373,3,144.951,580933.327,111.002
+dens_x1.0,42000,26057,1,274.180,724495.682,138.433
+dens_x1.0,42000,26057,2,190.711,724495.682,138.433
+dens_x1.0,42000,26057,3,282.941,724495.682,138.433
+dens_x1.0,49000,35903,1,223.180,916916.365,175.200
+dens_x1.0,49000,35903,2,323.990,916916.365,175.200
+dens_x1.0,49000,35903,3,227.125,916916.365,175.200
+dens_x1.0,57000,48648,1,317.633,1156930.810,221.061
+dens_x1.0,57000,48648,2,269.861,1156930.810,221.061
+dens_x1.0,57000,48648,3,301.058,1156930.810,221.061
+dens_x1.0,67000,67188,1,372.238,1491157.159,284.924
+dens_x1.0,67000,67188,2,310.594,1491157.159,284.924
+dens_x1.0,67000,67188,3,327.349,1491157.159,284.924
+dens_x1.0,79000,93552,1,399.597,1945903.955,371.815
+dens_x1.0,79000,93552,2,399.330,1945903.955,371.815
+dens_x1.0,79000,93552,3,378.416,1945903.955,371.815
+dens_x1.0,92000,126776,1,436.287,2500509.887,477.786
+dens_x1.0,92000,126776,2,458.982,2500509.887,477.786
+dens_x1.0,92000,126776,3,455.902,2500509.887,477.786
+dens_x1.0,108000,175020,1,552.531,3280169.679,626.760
+dens_x1.0,108000,175020,2,541.982,3280169.679,626.760
+dens_x1.0,108000,175020,3,541.280,3280169.679,626.760
+dens_x1.0,127000,241838,1,737.481,4334562.918,828.229
+dens_x1.0,127000,241838,2,728.767,4334562.918,828.229
+dens_x1.0,127000,241838,3,720.321,4334562.918,828.229
+dens_x1.0,149000,333094,1,934.288,5742559.864,1097.263
+dens_x1.0,149000,333094,2,1065.436,5742559.864,1097.263
+dens_x1.0,149000,333094,3,950.058,5742559.864,1097.263
+dens_x1.0,174000,454003,1,1169.608,7577993.244,1447.969
+dens_x1.0,174000,454003,2,1263.648,7577993.244,1447.969
+dens_x1.0,174000,454003,3,1170.809,7577993.244,1447.969
+dens_x1.0,204000,623499,1,1608.676,10116899.562,1933.092
+dens_x1.0,204000,623499,2,1533.414,10116899.562,1933.092
+dens_x1.0,204000,623499,3,1524.210,10116899.562,1933.092
+dens_x1.0,240000,862431,1,2177.558,13657349.809,2609.586
+dens_x1.0,240000,862431,2,2162.622,13657349.809,2609.586
+dens_x1.0,240000,862431,3,2250.789,13657349.809,2609.586
+dens_x1.0,281000,1182168,1,2836.469,18357066.601,3507.587
+dens_x1.0,281000,1182168,2,2819.405,18357066.601,3507.587
+dens_x1.0,281000,1182168,3,2877.142,18357066.601,3507.587
+dens_x1.0,329000,1623777,1,3866.669,24807713.897,4740.148
+dens_x1.0,329000,1623777,2,3998.567,24807713.897,4740.148
+dens_x1.0,329000,1623777,3,4059.382,24807713.897,4740.148
+dens_x1.0,386000,2232671,1,5485.621,33685517.024,6436.480
+dens_x1.0,386000,2232671,2,5580.961,33685517.024,6436.480
+dens_x1.0,386000,2232671,3,5513.234,33685517.024,6436.480
+dens_x1.0,452000,3055874,1,7599.997,45677561.904,8727.867
+dens_x1.0,452000,3055874,2,7705.885,45677561.904,8727.867
+dens_x1.0,452000,3055874,3,7632.451,45677561.904,8727.867
+dens_x1.0,530000,4208065,1,10645.239,62450692.510,11932.802
+dens_x1.0,530000,4208065,2,10608.786,62450692.510,11932.802
+dens_x1.0,530000,4208065,3,10661.889,62450692.510,11932.802
+dens_x1.0,621000,5774507,1,14916.794,85310220.195,16300.700
+dens_x1.0,621000,5774507,2,14760.646,85310220.195,16300.700
+dens_x1.0,621000,5774507,3,15000.037,85310220.195,16300.700
+dens_x1.0,728000,7935888,1,21350.019,116945648.236,22345.458
+dens_x1.0,728000,7935888,2,20599.824,116945648.236,22345.458
+dens_x1.0,728000,7935888,3,20715.817,116945648.236,22345.458
+dens_x1.0,853000,10900416,1,28939.165,160510699.866,30669.676
+dens_x1.0,853000,10900416,2,28468.021,160510699.866,30669.676
+dens_x1.0,853000,10900416,3,31349.436,160510699.866,30669.676
+dens_x1.0,1000000,14986435,1,66847.745,220860761.527,42201.099
+dens_x1.0,1000000,14986435,2,41387.027,220860761.527,42201.099
+dens_x1.0,1000000,14986435,3,42951.041,220860761.527,42201.099
+dens_x2.0,10000,3007,1,87.427,119798.897,22.891
+dens_x2.0,10000,3007,2,63.753,119798.897,22.891
+dens_x2.0,10000,3007,3,52.585,119798.897,22.891
+dens_x2.0,12000,4365,1,68.433,153710.912,29.370
+dens_x2.0,12000,4365,2,82.225,153710.912,29.370
+dens_x2.0,12000,4365,3,62.498,153710.912,29.370
+dens_x2.0,14000,5830,1,146.832,189313.294,36.173
+dens_x2.0,14000,5830,2,92.943,189313.294,36.173
+dens_x2.0,14000,5830,3,77.130,189313.294,36.173
+dens_x2.0,16000,7624,1,151.648,228688.447,43.697
+dens_x2.0,16000,7624,2,86.382,228688.447,43.697
+dens_x2.0,16000,7624,3,118.420,228688.447,43.697
+dens_x2.0,19000,10837,1,115.938,293959.920,56.169
+dens_x2.0,19000,10837,2,138.467,293959.920,56.169
+dens_x2.0,19000,10837,3,110.107,293959.920,56.169
+dens_x2.0,22000,14556,1,150.381,365516.050,69.841
+dens_x2.0,22000,14556,2,126.595,365516.050,69.841
+dens_x2.0,22000,14556,3,164.303,365516.050,69.841
+dens_x2.0,26000,20395,1,143.082,471644.695,90.120
+dens_x2.0,26000,20395,2,188.214,471644.695,90.120
+dens_x2.0,26000,20395,3,150.807,471644.695,90.120
+dens_x2.0,30000,26790,1,156.010,585445.422,111.864
+dens_x2.0,30000,26790,2,245.802,585445.422,111.864
+dens_x2.0,30000,26790,3,158.062,585445.422,111.864
+dens_x2.0,36000,38800,1,193.929,784747.311,149.946
+dens_x2.0,36000,38800,2,193.447,784747.311,149.946
+dens_x2.0,36000,38800,3,320.627,784747.311,149.946
+dens_x2.0,42000,53391,1,245.425,1015477.726,194.033
+dens_x2.0,42000,53391,2,259.902,1015477.726,194.033
+dens_x2.0,42000,53391,3,259.556,1015477.726,194.033
+dens_x2.0,49000,71945,1,307.752,1306154.668,249.574
+dens_x2.0,49000,71945,2,310.579,1306154.668,249.574
+dens_x2.0,49000,71945,3,332.788,1306154.668,249.574
+dens_x2.0,57000,97182,1,349.160,1688417.255,322.615
+dens_x2.0,57000,97182,2,344.003,1688417.255,322.615
+dens_x2.0,57000,97182,3,309.397,1688417.255,322.615
+dens_x2.0,67000,134082,1,415.061,2234513.248,426.961
+dens_x2.0,67000,134082,2,418.921,2234513.248,426.961
+dens_x2.0,67000,134082,3,542.886,2234513.248,426.961
+dens_x2.0,79000,186876,1,506.513,2998337.660,572.909
+dens_x2.0,79000,186876,2,509.927,2998337.660,572.909
+dens_x2.0,79000,186876,3,511.928,2998337.660,572.909
+dens_x2.0,92000,253371,1,636.724,3947432.991,754.258
+dens_x2.0,92000,253371,2,641.103,3947432.991,754.258
+dens_x2.0,92000,253371,3,630.308,3947432.991,754.258
+dens_x2.0,108000,349108,1,1719.294,5297829.841,1012.286
+dens_x2.0,108000,349108,2,1146.188,5297829.841,1012.286
+dens_x2.0,108000,349108,3,926.098,5297829.841,1012.286
+dens_x2.0,127000,484181,1,1180.387,7182563.887,1372.413
+dens_x2.0,127000,484181,2,1185.449,7182563.887,1372.413
+dens_x2.0,127000,484181,3,1133.979,7182563.887,1372.413
+dens_x2.0,149000,666491,1,1659.692,9713885.437,1856.086
+dens_x2.0,149000,666491,2,1612.972,9713885.437,1856.086
+dens_x2.0,149000,666491,3,1517.271,9713885.437,1856.086
+dens_x2.0,174000,907603,1,2171.669,13051498.522,2493.823
+dens_x2.0,174000,907603,2,2170.243,13051498.522,2493.823
+dens_x2.0,174000,907603,3,2138.745,13051498.522,2493.823
+dens_x2.0,204000,1247723,1,3040.113,17748584.329,3391.321
+dens_x2.0,204000,1247723,2,2865.235,17748584.329,3391.321
+dens_x2.0,204000,1247723,3,2818.529,17748584.329,3391.321
+dens_x2.0,240000,1725541,1,3911.937,24349896.729,4652.671
+dens_x2.0,240000,1725541,2,3864.313,24349896.729,4652.671
+dens_x2.0,240000,1725541,3,4019.525,24349896.729,4652.671
+dens_x2.0,281000,2365586,1,5440.080,33204358.944,6344.542
+dens_x2.0,281000,2365586,2,5280.197,33204358.944,6344.542
+dens_x2.0,281000,2365586,3,5331.377,33204358.944,6344.542
+dens_x2.0,329000,3242756,1,7509.897,45374920.412,8670.039
+dens_x2.0,329000,3242756,2,7297.788,45374920.412,8670.039
+dens_x2.0,329000,3242756,3,7308.891,45374920.412,8670.039
+dens_x2.0,386000,4461354,1,10341.541,62354387.279,11914.401
+dens_x2.0,386000,4461354,2,11255.962,62354387.279,11914.401
+dens_x2.0,386000,4461354,3,10464.571,62354387.279,11914.401
+dens_x2.0,452000,6119886,1,14859.560,85575402.535,16351.370
+dens_x2.0,452000,6119886,2,14641.356,85575402.535,16351.370
+dens_x2.0,452000,6119886,3,14368.936,85575402.535,16351.370
+dens_x2.0,530000,8413791,1,20623.174,117884820.410,22524.911
+dens_x2.0,530000,8413791,2,20776.519,117884820.410,22524.911
+dens_x2.0,530000,8413791,3,35375.719,117884820.410,22524.911
+dens_x2.0,621000,11552079,1,35126.303,162377752.059,31026.424
+dens_x2.0,621000,11552079,2,28352.579,162377752.059,31026.424
+dens_x2.0,621000,11552079,3,28463.199,162377752.059,31026.424
+dens_x2.0,728000,15870417,1,40801.199,224046367.608,42809.790
+dens_x2.0,728000,15870417,2,39676.070,224046367.608,42809.790
+dens_x2.0,728000,15870417,3,40929.421,224046367.608,42809.790
+dens_x2.0,853000,21800415,1,57857.381,309366697.818,59112.422
+dens_x2.0,853000,21800415,2,57089.829,309366697.818,59112.422
+dens_x2.0,853000,21800415,3,56924.223,309366697.818,59112.422
+dens_x2.0,1000000,29958277,1,81612.419,427704402.750,81723.868
+dens_x2.0,1000000,29958277,2,81139.967,427704402.750,81723.868
+dens_x2.0,1000000,29958277,3,85039.919,427704402.750,81723.868

results_lct.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,35.961,99112.473,62.983
+dens_x0.5,10000,761,2,34.394,99112.473,62.983
+dens_x0.5,10000,761,3,50.859,99112.473,62.983
+dens_x0.5,12000,1077,1,45.321,122827.840,78.053
+dens_x0.5,12000,1077,2,62.352,122827.840,78.053
+dens_x0.5,12000,1077,3,47.778,122827.840,78.053
+dens_x0.5,14000,1455,1,73.521,147545.989,93.761
+dens_x0.5,14000,1455,2,62.999,147545.989,93.761
+dens_x0.5,14000,1455,3,60.800,147545.989,93.761
+dens_x0.5,16000,1937,1,87.146,173636.330,110.340
+dens_x0.5,16000,1937,2,67.794,173636.330,110.340
+dens_x0.5,16000,1937,3,164.459,173636.330,110.340
+dens_x0.5,19000,2715,1,145.995,213940.398,135.952
+dens_x0.5,19000,2715,2,175.168,213940.398,135.952
+dens_x0.5,19000,2715,3,116.191,213940.398,135.952
+dens_x0.5,22000,3611,1,99.170,256079.209,162.730
+dens_x0.5,22000,3611,2,92.240,256079.209,162.730
+dens_x0.5,22000,3611,3,92.266,256079.209,162.730
+dens_x0.5,26000,5194,1,111.996,317113.582,201.515
+dens_x0.5,26000,5194,2,99.461,317113.582,201.515
+dens_x0.5,26000,5194,3,110.166,317113.582,201.515
+dens_x0.5,30000,6763,1,183.794,378988.027,240.834
+dens_x0.5,30000,6763,2,181.493,378988.027,240.834
+dens_x0.5,30000,6763,3,186.844,378988.027,240.834
+dens_x0.5,36000,9660,1,182.254,479031.581,304.409
+dens_x0.5,36000,9660,2,200.999,479031.581,304.409
+dens_x0.5,36000,9660,3,143.599,479031.581,304.409
+dens_x0.5,42000,13346,1,179.495,589181.686,374.405
+dens_x0.5,42000,13346,2,181.815,589181.686,374.405
+dens_x0.5,42000,13346,3,165.094,589181.686,374.405
+dens_x0.5,49000,17900,1,284.586,722491.606,459.120
+dens_x0.5,49000,17900,2,233.091,722491.606,459.120
+dens_x0.5,49000,17900,3,208.590,722491.606,459.120
+dens_x0.5,57000,24363,1,309.408,890990.473,566.195
+dens_x0.5,57000,24363,2,261.349,890990.473,566.195
+dens_x0.5,57000,24363,3,265.088,890990.473,566.195
+dens_x0.5,67000,33829,1,335.052,1120457.009,712.013
+dens_x0.5,67000,33829,2,449.145,1120457.009,712.013
+dens_x0.5,67000,33829,3,410.054,1120457.009,712.013
+dens_x0.5,79000,46917,1,511.215,1419991.587,902.358
+dens_x0.5,79000,46917,2,531.563,1419991.587,902.358
+dens_x0.5,79000,46917,3,437.912,1419991.587,902.358
+dens_x0.5,92000,63864,1,600.704,1781454.424,1132.055
+dens_x0.5,92000,63864,2,666.763,1781454.424,1132.055
+dens_x0.5,92000,63864,3,602.090,1781454.424,1132.055
+dens_x0.5,108000,87350,1,767.170,2264084.329,1438.751
+dens_x0.5,108000,87350,2,736.665,2264084.329,1438.751
+dens_x0.5,108000,87350,3,674.970,2264084.329,1438.751
+dens_x0.5,127000,120912,1,913.971,2913447.536,1851.400
+dens_x0.5,127000,120912,2,814.923,2913447.536,1851.400
+dens_x0.5,127000,120912,3,740.522,2913447.536,1851.400
+dens_x0.5,149000,166365,1,967.782,3756533.770,2387.153
+dens_x0.5,149000,166365,2,1230.397,3756533.770,2387.153
+dens_x0.5,149000,166365,3,1109.607,3756533.770,2387.153
+dens_x0.5,174000,226551,1,1557.154,4833373.044,3071.449
+dens_x0.5,174000,226551,2,1316.292,4833373.044,3071.449
+dens_x0.5,174000,226551,3,1317.413,4833373.044,3071.449
+dens_x0.5,204000,312325,1,1605.837,6312525.050,4011.401
+dens_x0.5,204000,312325,2,1914.829,6312525.050,4011.401
+dens_x0.5,204000,312325,3,1488.417,6312525.050,4011.401
+dens_x0.5,240000,431142,1,1820.898,8314371.662,5283.508
+dens_x0.5,240000,431142,2,1964.801,8314371.662,5283.508
+dens_x0.5,240000,431142,3,2009.085,8314371.662,5283.508
+dens_x0.5,281000,589828,1,2600.835,10925503.834,6942.796
+dens_x0.5,281000,589828,2,2540.880,10925503.834,6942.796
+dens_x0.5,281000,589828,3,2590.190,10925503.834,6942.796
+dens_x0.5,329000,811844,1,3473.126,14493068.872,9209.866
+dens_x0.5,329000,811844,2,3470.392,14493068.872,9209.866
+dens_x0.5,329000,811844,3,3314.117,14493068.872,9209.866
+dens_x0.5,386000,1115568,1,5107.043,19315559.086,12274.399
+dens_x0.5,386000,1115568,2,4930.561,19315559.086,12274.399
+dens_x0.5,386000,1115568,3,6380.473,19315559.086,12274.399
+dens_x0.5,452000,1531464,1,10735.902,25827552.428,16412.555
+dens_x0.5,452000,1531464,2,6940.655,25827552.428,16412.555
+dens_x0.5,452000,1531464,3,11508.761,25827552.428,16412.555
+dens_x0.5,530000,2103874,1,11136.995,34716124.680,22060.949
+dens_x0.5,530000,2103874,2,9777.029,34716124.680,22060.949
+dens_x0.5,530000,2103874,3,9783.617,34716124.680,22060.949
+dens_x0.5,621000,2887681,1,14149.726,46802598.872,29741.503
+dens_x0.5,621000,2887681,2,14757.810,46802598.872,29741.503
+dens_x0.5,621000,2887681,3,14496.587,46802598.872,29741.503
+dens_x0.5,728000,3971764,1,19466.477,63437679.196,40312.547
+dens_x0.5,728000,3971764,2,20275.542,63437679.196,40312.547
+dens_x0.5,728000,3971764,3,21114.354,63437679.196,40312.547
+dens_x0.5,853000,5448500,1,30840.224,86056528.179,54686.077
+dens_x0.5,853000,5448500,2,30346.139,86056528.179,54686.077
+dens_x0.5,853000,5448500,3,31170.105,86056528.179,54686.077
+dens_x0.5,1000000,7489974,1,47582.788,117293325.434,74536.028
+dens_x0.5,1000000,7489974,2,46482.519,117293325.434,74536.028
+dens_x0.5,1000000,7489974,3,44026.110,117293325.434,74536.028
+dens_x1.0,10000,1550,1,44.766,106379.431,67.601
+dens_x1.0,10000,1550,2,80.772,106379.431,67.601
+dens_x1.0,10000,1550,3,64.825,106379.431,67.601
+dens_x1.0,12000,2135,1,117.380,132765.276,84.368
+dens_x1.0,12000,2135,2,83.140,132765.276,84.368
+dens_x1.0,12000,2135,3,124.942,132765.276,84.368
+dens_x1.0,14000,2963,1,153.056,161942.582,102.909
+dens_x1.0,14000,2963,2,107.005,161942.582,102.909
+dens_x1.0,14000,2963,3,176.352,161942.582,102.909
+dens_x1.0,16000,3912,1,125.583,192755.010,122.489
+dens_x1.0,16000,3912,2,134.302,192755.010,122.489
+dens_x1.0,16000,3912,3,185.688,192755.010,122.489
+dens_x1.0,19000,5413,1,123.266,240521.618,152.844
+dens_x1.0,19000,5413,2,203.327,240521.618,152.844
+dens_x1.0,19000,5413,3,122.239,240521.618,152.844
+dens_x1.0,22000,7290,1,181.987,292864.786,186.106
+dens_x1.0,22000,7290,2,164.418,292864.786,186.106
+dens_x1.0,22000,7290,3,156.525,292864.786,186.106
+dens_x1.0,26000,10078,1,259.958,366763.602,233.066
+dens_x1.0,26000,10078,2,206.015,366763.602,233.066
+dens_x1.0,26000,10078,3,197.615,366763.602,233.066
+dens_x1.0,30000,13336,1,247.134,446748.773,283.894
+dens_x1.0,30000,13336,2,249.997,446748.773,283.894
+dens_x1.0,30000,13336,3,248.417,446748.773,283.894
+dens_x1.0,36000,19373,1,342.043,580933.327,369.164
+dens_x1.0,36000,19373,2,286.624,580933.327,369.164
+dens_x1.0,36000,19373,3,352.649,580933.327,369.164
+dens_x1.0,42000,26057,1,351.357,724495.682,460.393
+dens_x1.0,42000,26057,2,353.756,724495.682,460.393
+dens_x1.0,42000,26057,3,405.657,724495.682,460.393
+dens_x1.0,49000,35903,1,419.918,916916.365,582.670
+dens_x1.0,49000,35903,2,482.741,916916.365,582.670
+dens_x1.0,49000,35903,3,429.988,916916.365,582.670
+dens_x1.0,57000,48648,1,520.940,1156930.810,735.191
+dens_x1.0,57000,48648,2,440.392,1156930.810,735.191
+dens_x1.0,57000,48648,3,427.737,1156930.810,735.191
+dens_x1.0,67000,67188,1,570.235,1491157.159,947.581
+dens_x1.0,67000,67188,2,553.552,1491157.159,947.581
+dens_x1.0,67000,67188,3,583.040,1491157.159,947.581
+dens_x1.0,79000,93552,1,683.788,1945903.955,1236.558
+dens_x1.0,79000,93552,2,669.714,1945903.955,1236.558
+dens_x1.0,79000,93552,3,732.646,1945903.955,1236.558
+dens_x1.0,92000,126776,1,769.173,2500509.887,1588.991
+dens_x1.0,92000,126776,2,765.284,2500509.887,1588.991
+dens_x1.0,92000,126776,3,765.463,2500509.887,1588.991
+dens_x1.0,108000,175020,1,929.263,3280169.679,2084.439
+dens_x1.0,108000,175020,2,956.523,3280169.679,2084.439
+dens_x1.0,108000,175020,3,950.606,3280169.679,2084.439
+dens_x1.0,127000,241838,1,1181.597,4334562.918,2754.471
+dens_x1.0,127000,241838,2,1092.162,4334562.918,2754.471
+dens_x1.0,127000,241838,3,1079.477,4334562.918,2754.471
+dens_x1.0,149000,333094,1,1442.657,5742559.864,3649.207
+dens_x1.0,149000,333094,2,1365.485,5742559.864,3649.207
+dens_x1.0,149000,333094,3,1393.380,5742559.864,3649.207
+dens_x1.0,174000,454003,1,1856.201,7577993.244,4815.564
+dens_x1.0,174000,454003,2,1991.080,7577993.244,4815.564
+dens_x1.0,174000,454003,3,1853.065,7577993.244,4815.564
+dens_x1.0,204000,623499,1,2612.491,10116899.562,6428.955
+dens_x1.0,204000,623499,2,2464.648,10116899.562,6428.955
+dens_x1.0,204000,623499,3,2486.181,10116899.562,6428.955
+dens_x1.0,240000,862431,1,3883.620,13657349.809,8678.794
+dens_x1.0,240000,862431,2,3654.052,13657349.809,8678.794
+dens_x1.0,240000,862431,3,3628.481,13657349.809,8678.794
+dens_x1.0,281000,1182168,1,4952.134,18357066.601,11665.309
+dens_x1.0,281000,1182168,2,5243.182,18357066.601,11665.309
+dens_x1.0,281000,1182168,3,5014.568,18357066.601,11665.309
+dens_x1.0,329000,1623777,1,6905.065,24807713.897,15764.482
+dens_x1.0,329000,1623777,2,7251.999,24807713.897,15764.482
+dens_x1.0,329000,1623777,3,6730.380,24807713.897,15764.482
+dens_x1.0,386000,2232671,1,10210.314,33685517.024,21406.032
+dens_x1.0,386000,2232671,2,10712.272,33685517.024,21406.032
+dens_x1.0,386000,2232671,3,10134.142,33685517.024,21406.032
+dens_x1.0,452000,3055874,1,15197.842,45677561.904,29026.580
+dens_x1.0,452000,3055874,2,14589.433,45677561.904,29026.580
+dens_x1.0,452000,3055874,3,13996.179,45677561.904,29026.580
+dens_x1.0,530000,4208065,1,21873.394,62450692.510,39685.349
+dens_x1.0,530000,4208065,2,20504.892,62450692.510,39685.349
+dens_x1.0,530000,4208065,3,22332.396,62450692.510,39685.349
+dens_x1.0,621000,5774507,1,33111.195,85310220.195,54211.823
+dens_x1.0,621000,5774507,2,33271.836,85310220.195,54211.823
+dens_x1.0,621000,5774507,3,33340.294,85310220.195,54211.823
+dens_x1.0,728000,7935888,1,47822.973,116945648.236,74315.091
+dens_x1.0,728000,7935888,2,48739.415,116945648.236,74315.091
+dens_x1.0,728000,7935888,3,48949.025,116945648.236,74315.091
+dens_x1.0,853000,10900416,1,84950.950,160510699.866,101999.240
+dens_x1.0,853000,10900416,2,79421.655,160510699.866,101999.240
+dens_x1.0,853000,10900416,3,87799.883,160510699.866,101999.240
+dens_x1.0,1000000,14986435,1,113252.341,220860761.527,140349.708
+dens_x1.0,1000000,14986435,2,123807.781,220860761.527,140349.708
+dens_x1.0,1000000,14986435,3,127188.415,220860761.527,140349.708
+dens_x2.0,10000,3007,1,123.275,119798.897,76.128
+dens_x2.0,10000,3007,2,79.558,119798.897,76.128
+dens_x2.0,10000,3007,3,128.487,119798.897,76.128
+dens_x2.0,12000,4365,1,82.814,153710.912,97.678
+dens_x2.0,12000,4365,2,120.726,153710.912,97.678
+dens_x2.0,12000,4365,3,73.037,153710.912,97.678
+dens_x2.0,14000,5830,1,156.142,189313.294,120.302
+dens_x2.0,14000,5830,2,98.668,189313.294,120.302
+dens_x2.0,14000,5830,3,111.172,189313.294,120.302
+dens_x2.0,16000,7624,1,114.385,228688.447,145.324
+dens_x2.0,16000,7624,2,163.465,228688.447,145.324
+dens_x2.0,16000,7624,3,180.540,228688.447,145.324
+dens_x2.0,19000,10837,1,212.731,293959.920,186.802
+dens_x2.0,19000,10837,2,311.261,293959.920,186.802
+dens_x2.0,19000,10837,3,287.109,293959.920,186.802
+dens_x2.0,22000,14556,1,240.075,365516.050,232.273
+dens_x2.0,22000,14556,2,150.187,365516.050,232.273
+dens_x2.0,22000,14556,3,295.141,365516.050,232.273
+dens_x2.0,26000,20395,1,223.860,471644.695,299.715
+dens_x2.0,26000,20395,2,237.952,471644.695,299.715
+dens_x2.0,26000,20395,3,202.743,471644.695,299.715
+dens_x2.0,30000,26790,1,393.274,585445.422,372.031
+dens_x2.0,30000,26790,2,301.000,585445.422,372.031
+dens_x2.0,30000,26790,3,349.833,585445.422,372.031
+dens_x2.0,36000,38800,1,370.541,784747.311,498.681
+dens_x2.0,36000,38800,2,364.912,784747.311,498.681
+dens_x2.0,36000,38800,3,376.424,784747.311,498.681
+dens_x2.0,42000,53391,1,448.852,1015477.726,645.303
+dens_x2.0,42000,53391,2,441.742,1015477.726,645.303
+dens_x2.0,42000,53391,3,451.396,1015477.726,645.303
+dens_x2.0,49000,71945,1,442.243,1306154.668,830.018
+dens_x2.0,49000,71945,2,472.422,1306154.668,830.018
+dens_x2.0,49000,71945,3,523.281,1306154.668,830.018
+dens_x2.0,57000,97182,1,578.557,1688417.255,1072.933
+dens_x2.0,57000,97182,2,489.410,1688417.255,1072.933
+dens_x2.0,57000,97182,3,518.427,1688417.255,1072.933
+dens_x2.0,67000,134082,1,780.094,2234513.248,1419.959
+dens_x2.0,67000,134082,2,755.497,2234513.248,1419.959
+dens_x2.0,67000,134082,3,674.201,2234513.248,1419.959
+dens_x2.0,79000,186876,1,833.358,2998337.660,1905.344
+dens_x2.0,79000,186876,2,961.736,2998337.660,1905.344
+dens_x2.0,79000,186876,3,816.439,2998337.660,1905.344
+dens_x2.0,92000,253371,1,1203.082,3947432.991,2508.463
+dens_x2.0,92000,253371,2,1186.119,3947432.991,2508.463
+dens_x2.0,92000,253371,3,1192.455,3947432.991,2508.463
+dens_x2.0,108000,349108,1,1618.963,5297829.841,3366.596
+dens_x2.0,108000,349108,2,1345.165,5297829.841,3366.596
+dens_x2.0,108000,349108,3,1338.014,5297829.841,3366.596
+dens_x2.0,127000,484181,1,2170.970,7182563.887,4564.282
+dens_x2.0,127000,484181,2,2114.696,7182563.887,4564.282
+dens_x2.0,127000,484181,3,2058.957,7182563.887,4564.282
+dens_x2.0,149000,666491,1,2903.698,9713885.437,6172.853
+dens_x2.0,149000,666491,2,2768.252,9713885.437,6172.853
+dens_x2.0,149000,666491,3,2683.809,9713885.437,6172.853
+dens_x2.0,174000,907603,1,3750.869,13051498.522,8293.796
+dens_x2.0,174000,907603,2,3788.667,13051498.522,8293.796
+dens_x2.0,174000,907603,3,3595.097,13051498.522,8293.796
+dens_x2.0,204000,1247723,1,5598.056,17748584.329,11278.638
+dens_x2.0,204000,1247723,2,5033.484,17748584.329,11278.638
+dens_x2.0,204000,1247723,3,5031.559,17748584.329,11278.638
+dens_x2.0,240000,1725541,1,8026.526,24349896.729,15473.554
+dens_x2.0,240000,1725541,2,7715.847,24349896.729,15473.554
+dens_x2.0,240000,1725541,3,7493.727,24349896.729,15473.554
+dens_x2.0,281000,2365586,1,12315.262,33204358.944,21100.272
+dens_x2.0,281000,2365586,2,10750.730,33204358.944,21100.272
+dens_x2.0,281000,2365586,3,10910.945,33204358.944,21100.272
+dens_x2.0,329000,3242756,1,17218.988,45374920.412,28834.261
+dens_x2.0,329000,3242756,2,15704.277,45374920.412,28834.261
+dens_x2.0,329000,3242756,3,16909.888,45374920.412,28834.261
+dens_x2.0,386000,4461354,1,22930.235,62354387.279,39624.150
+dens_x2.0,386000,4461354,2,23229.405,62354387.279,39624.150
+dens_x2.0,386000,4461354,3,23342.470,62354387.279,39624.150
+dens_x2.0,452000,6119886,1,35110.122,85575402.535,54380.337
+dens_x2.0,452000,6119886,2,35385.954,85575402.535,54380.337
+dens_x2.0,452000,6119886,3,35167.751,85575402.535,54380.337
+dens_x2.0,530000,8413791,1,55796.898,117884820.410,74911.904
+dens_x2.0,530000,8413791,2,59173.450,117884820.410,74911.904
+dens_x2.0,530000,8413791,3,56385.714,117884820.410,74911.904
+dens_x2.0,621000,11552079,1,87065.836,162377752.059,103185.690
+dens_x2.0,621000,11552079,2,88510.360,162377752.059,103185.690
+dens_x2.0,621000,11552079,3,87154.836,162377752.059,103185.690
+dens_x2.0,728000,15870417,1,131315.993,224046367.608,142374.055
+dens_x2.0,728000,15870417,2,129778.579,224046367.608,142374.055
+dens_x2.0,728000,15870417,3,129852.250,224046367.608,142374.055
+dens_x2.0,853000,21800415,1,216168.770,309366697.818,196592.303
+dens_x2.0,853000,21800415,2,208242.155,309366697.818,196592.303
+dens_x2.0,853000,21800415,3,198993.350,309366697.818,196592.303
+dens_x2.0,1000000,29958277,1,365605.129,427704402.750,271792.000
+dens_x2.0,1000000,29958277,2,338676.140,427704402.750,271792.000
+dens_x2.0,1000000,29958277,3,326162.678,427704402.750,271792.000

results_mk.csv

@@ -0,0 +1,307 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,15.688,99112.473,21.422
+dens_x0.5,10000,761,2,15.767,99112.473,21.422
+dens_x0.5,10000,761,3,15.154,99112.473,21.422
+dens_x0.5,12000,1077,1,20.751,122827.840,26.548
+dens_x0.5,12000,1077,2,20.008,122827.840,26.548
+dens_x0.5,12000,1077,3,18.613,122827.840,26.548
+dens_x0.5,14000,1455,1,22.026,147545.989,31.890
+dens_x0.5,14000,1455,2,21.376,147545.989,31.890
+dens_x0.5,14000,1455,3,22.028,147545.989,31.890
+dens_x0.5,16000,1937,1,26.161,173636.330,37.529
+dens_x0.5,16000,1937,2,27.869,173636.330,37.529
+dens_x0.5,16000,1937,3,28.148,173636.330,37.529
+dens_x0.5,19000,2715,1,39.460,213940.398,46.241
+dens_x0.5,19000,2715,2,39.874,213940.398,46.241
+dens_x0.5,19000,2715,3,37.501,213940.398,46.241
+dens_x0.5,22000,3611,1,45.224,256079.209,55.348
+dens_x0.5,22000,3611,2,44.134,256079.209,55.348
+dens_x0.5,22000,3611,3,44.888,256079.209,55.348
+dens_x0.5,26000,5194,1,56.877,317113.582,68.540
+dens_x0.5,26000,5194,2,56.417,317113.582,68.540
+dens_x0.5,26000,5194,3,50.854,317113.582,68.540
+dens_x0.5,30000,6763,1,55.551,378988.027,81.914
+dens_x0.5,30000,6763,2,56.281,378988.027,81.914
+dens_x0.5,30000,6763,3,61.705,378988.027,81.914
+dens_x0.5,36000,9660,1,58.902,479031.581,103.537
+dens_x0.5,36000,9660,2,59.343,479031.581,103.537
+dens_x0.5,36000,9660,3,56.766,479031.581,103.537
+dens_x0.5,42000,13346,1,71.784,589181.686,127.344
+dens_x0.5,42000,13346,2,70.883,589181.686,127.344
+dens_x0.5,42000,13346,3,67.382,589181.686,127.344
+dens_x0.5,49000,17900,1,77.784,722491.606,156.158
+dens_x0.5,49000,17900,2,98.588,722491.606,156.158
+dens_x0.5,49000,17900,3,78.477,722491.606,156.158
+dens_x0.5,57000,24363,1,100.125,890990.473,192.577
+dens_x0.5,57000,24363,2,98.099,890990.473,192.577
+dens_x0.5,57000,24363,3,95.559,890990.473,192.577
+dens_x0.5,67000,33829,1,120.753,1120457.009,242.173
+dens_x0.5,67000,33829,2,157.183,1120457.009,242.173
+dens_x0.5,67000,33829,3,119.986,1120457.009,242.173
+dens_x0.5,79000,46917,1,158.279,1419991.587,306.914
+dens_x0.5,79000,46917,2,160.867,1419991.587,306.914
+dens_x0.5,79000,46917,3,188.158,1419991.587,306.914
+dens_x0.5,92000,63864,1,217.457,1781454.424,385.039
+dens_x0.5,92000,63864,2,201.234,1781454.424,385.039
+dens_x0.5,92000,63864,3,202.297,1781454.424,385.039
+dens_x0.5,108000,87350,1,280.966,2264084.329,489.354
+dens_x0.5,108000,87350,2,249.816,2264084.329,489.354
+dens_x0.5,108000,87350,3,300.323,2264084.329,489.354
+dens_x0.5,127000,120912,1,303.758,2913447.536,629.706
+dens_x0.5,127000,120912,2,353.401,2913447.536,629.706
+dens_x0.5,127000,120912,3,611.729,2913447.536,629.706
+dens_x0.5,149000,166365,1,520.963,3756533.770,811.928
+dens_x0.5,149000,166365,2,567.283,3756533.770,811.928
+dens_x0.5,149000,166365,3,464.261,3756533.770,811.928
+dens_x0.5,174000,226551,1,579.208,4833373.044,1044.674
+dens_x0.5,174000,226551,2,645.586,4833373.044,1044.674
+dens_x0.5,174000,226551,3,566.923,4833373.044,1044.674
+dens_x0.5,204000,312325,1,908.258,6312525.050,1364.374
+dens_x0.5,204000,312325,2,597.708,6312525.050,1364.374
+dens_x0.5,204000,312325,3,614.792,6312525.050,1364.374
+dens_x0.5,240000,431142,1,830.654,8314371.662,1797.049
+dens_x0.5,240000,431142,2,747.721,8314371.662,1797.049
+dens_x0.5,240000,431142,3,793.694,8314371.662,1797.049
+dens_x0.5,281000,589828,1,1009.589,10925503.834,2361.413
+dens_x0.5,281000,589828,2,999.220,10925503.834,2361.413
+dens_x0.5,281000,589828,3,968.745,10925503.834,2361.413
+dens_x0.5,329000,811844,1,1346.701,14493068.872,3132.498
+dens_x0.5,329000,811844,2,1300.301,14493068.872,3132.498
+dens_x0.5,329000,811844,3,1336.633,14493068.872,3132.498
+dens_x0.5,386000,1115568,1,1930.593,19315559.086,4174.819
+dens_x0.5,386000,1115568,2,1943.367,19315559.086,4174.819
+dens_x0.5,386000,1115568,3,1826.516,19315559.086,4174.819
+dens_x0.5,452000,1531464,1,2626.389,25827552.428,5582.306
+dens_x0.5,452000,1531464,2,2527.856,25827552.428,5582.306
+dens_x0.5,452000,1531464,3,2552.325,25827552.428,5582.306
+dens_x0.5,530000,2103874,1,3927.716,34716124.680,7503.461
+dens_x0.5,530000,2103874,2,3968.094,34716124.680,7503.461
+dens_x0.5,530000,2103874,3,3782.988,34716124.680,7503.461
+dens_x0.5,621000,2887681,1,5852.172,46802598.872,10115.803
+dens_x0.5,621000,2887681,2,5655.723,46802598.872,10115.803
+dens_x0.5,621000,2887681,3,5788.834,46802598.872,10115.803
+dens_x0.5,728000,3971764,1,8425.377,63437679.196,13711.270
+dens_x0.5,728000,3971764,2,8226.609,63437679.196,13711.270
+dens_x0.5,728000,3971764,3,8583.745,63437679.196,13711.270
+dens_x0.5,853000,5448500,1,12175.891,86056528.179,18600.055
+dens_x0.5,853000,5448500,2,15218.073,86056528.179,18600.055
+dens_x0.5,853000,5448500,3,11273.111,86056528.179,18600.055
+dens_x0.5,1000000,7489974,1,22260.390,117293325.434,25351.503
+dens_x0.5,1000000,7489974,2,21735.102,117293325.434,25351.503
+dens_x0.5,1000000,7489974,3,16119.248,117293325.434,25351.503
+dens_x0.5,1300000,12655006,1,36062.526,196456827.614,73316.712
+dens_x0.5,1300000,12655006,2,37491.996,196456827.614,73316.712
+dens_x0.5,1300000,12655006,3,40901.895,196456827.614,73316.712
+dens_x0.5,1500000,16856907,1,55060.072,261053127.751,97423.730
+dens_x0.5,1500000,16856907,2,51489.209,261053127.751,97423.730
+dens_x0.5,1500000,16856907,3,51762.161,261053127.751,97423.730
+dens_x1.0,10000,1550,1,20.015,106379.431,22.993
+dens_x1.0,10000,1550,2,16.967,106379.431,22.993
+dens_x1.0,10000,1550,3,33.054,106379.431,22.993
+dens_x1.0,12000,2135,1,40.238,132765.276,28.696
+dens_x1.0,12000,2135,2,22.697,132765.276,28.696
+dens_x1.0,12000,2135,3,23.725,132765.276,28.696
+dens_x1.0,14000,2963,1,35.596,161942.582,35.002
+dens_x1.0,14000,2963,2,28.078,161942.582,35.002
+dens_x1.0,14000,2963,3,33.954,161942.582,35.002
+dens_x1.0,16000,3912,1,44.754,192755.010,41.662
+dens_x1.0,16000,3912,2,46.279,192755.010,41.662
+dens_x1.0,16000,3912,3,43.305,192755.010,41.662
+dens_x1.0,19000,5413,1,59.872,240521.618,51.986
+dens_x1.0,19000,5413,2,54.141,240521.618,51.986
+dens_x1.0,19000,5413,3,59.136,240521.618,51.986
+dens_x1.0,22000,7290,1,57.037,292864.786,63.299
+dens_x1.0,22000,7290,2,80.374,292864.786,63.299
+dens_x1.0,22000,7290,3,51.274,292864.786,63.299
+dens_x1.0,26000,10078,1,100.147,366763.602,79.271
+dens_x1.0,26000,10078,2,64.036,366763.602,79.271
+dens_x1.0,26000,10078,3,118.154,366763.602,79.271
+dens_x1.0,30000,13336,1,63.956,446748.773,96.559
+dens_x1.0,30000,13336,2,90.179,446748.773,96.559
+dens_x1.0,30000,13336,3,63.027,446748.773,96.559
+dens_x1.0,36000,19373,1,133.343,580933.327,125.562
+dens_x1.0,36000,19373,2,82.262,580933.327,125.562
+dens_x1.0,36000,19373,3,126.932,580933.327,125.562
+dens_x1.0,42000,26057,1,310.700,724495.682,156.591
+dens_x1.0,42000,26057,2,182.416,724495.682,156.591
+dens_x1.0,42000,26057,3,146.894,724495.682,156.591
+dens_x1.0,49000,35903,1,152.136,916916.365,198.180
+dens_x1.0,49000,35903,2,134.394,916916.365,198.180
+dens_x1.0,49000,35903,3,140.925,916916.365,198.180
+dens_x1.0,57000,48648,1,283.893,1156930.810,250.056
+dens_x1.0,57000,48648,2,234.125,1156930.810,250.056
+dens_x1.0,57000,48648,3,190.222,1156930.810,250.056
+dens_x1.0,67000,67188,1,289.951,1491157.159,322.295
+dens_x1.0,67000,67188,2,237.208,1491157.159,322.295
+dens_x1.0,67000,67188,3,287.380,1491157.159,322.295
+dens_x1.0,79000,93552,1,275.404,1945903.955,420.583
+dens_x1.0,79000,93552,2,307.363,1945903.955,420.583
+dens_x1.0,79000,93552,3,253.094,1945903.955,420.583
+dens_x1.0,92000,126776,1,342.068,2500509.887,540.454
+dens_x1.0,92000,126776,2,276.337,2500509.887,540.454
+dens_x1.0,92000,126776,3,288.147,2500509.887,540.454
+dens_x1.0,108000,175020,1,316.023,3280169.679,708.968
+dens_x1.0,108000,175020,2,353.284,3280169.679,708.968
+dens_x1.0,108000,175020,3,339.276,3280169.679,708.968
+dens_x1.0,127000,241838,1,391.496,4334562.918,936.862
+dens_x1.0,127000,241838,2,384.289,4334562.918,936.862
+dens_x1.0,127000,241838,3,400.746,4334562.918,936.862
+dens_x1.0,149000,333094,1,601.271,5742559.864,1241.183
+dens_x1.0,149000,333094,2,542.543,5742559.864,1241.183
+dens_x1.0,149000,333094,3,507.314,5742559.864,1241.183
+dens_x1.0,174000,454003,1,705.319,7577993.244,1637.890
+dens_x1.0,174000,454003,2,708.632,7577993.244,1637.890
+dens_x1.0,174000,454003,3,662.705,7577993.244,1637.890
+dens_x1.0,204000,623499,1,944.243,10116899.562,2186.643
+dens_x1.0,204000,623499,2,929.347,10116899.562,2186.643
+dens_x1.0,204000,623499,3,942.194,10116899.562,2186.643
+dens_x1.0,240000,862431,1,1338.912,13657349.809,2951.867
+dens_x1.0,240000,862431,2,1197.529,13657349.809,2951.867
+dens_x1.0,240000,862431,3,1288.236,13657349.809,2951.867
+dens_x1.0,281000,1182168,1,1812.009,18357066.601,3967.653
+dens_x1.0,281000,1182168,2,1831.549,18357066.601,3967.653
+dens_x1.0,281000,1182168,3,1800.877,18357066.601,3967.653
+dens_x1.0,329000,1623777,1,2607.569,24807713.897,5361.881
+dens_x1.0,329000,1623777,2,2721.496,24807713.897,5361.881
+dens_x1.0,329000,1623777,3,2692.989,24807713.897,5361.881
+dens_x1.0,386000,2232671,1,3896.705,33685517.024,7280.708
+dens_x1.0,386000,2232671,2,4093.978,33685517.024,7280.708
+dens_x1.0,386000,2232671,3,4157.255,33685517.024,7280.708
+dens_x1.0,452000,3055874,1,6800.304,45677561.904,9872.640
+dens_x1.0,452000,3055874,2,6063.371,45677561.904,9872.640
+dens_x1.0,452000,3055874,3,5619.696,45677561.904,9872.640
+dens_x1.0,530000,4208065,1,7359.967,62450692.510,13497.945
+dens_x1.0,530000,4208065,2,8442.994,62450692.510,13497.945
+dens_x1.0,530000,4208065,3,9818.880,62450692.510,13497.945
+dens_x1.0,621000,5774507,1,13500.980,85310220.195,18438.750
+dens_x1.0,621000,5774507,2,15204.199,85310220.195,18438.750
+dens_x1.0,621000,5774507,3,12344.941,85310220.195,18438.750
+dens_x1.0,728000,7935888,1,20263.595,116945648.236,25276.356
+dens_x1.0,728000,7935888,2,21043.008,116945648.236,25276.356
+dens_x1.0,728000,7935888,3,20002.944,116945648.236,25276.356
+dens_x1.0,853000,10900416,1,36276.239,160510699.866,34692.404
+dens_x1.0,853000,10900416,2,28334.668,160510699.866,34692.404
+dens_x1.0,853000,10900416,3,35198.553,160510699.866,34692.404
+dens_x1.0,1000000,14986435,1,45843.089,220860761.527,47736.324
+dens_x1.0,1000000,14986435,2,47166.816,220860761.527,47736.324
+dens_x1.0,1000000,14986435,3,49091.110,220860761.527,47736.324
+dens_x1.0,1300000,25315094,1,98721.978,374683961.719,139830.193
+dens_x1.0,1300000,25315094,2,78676.425,374683961.719,139830.193
+dens_x1.0,1300000,25315094,3,103389.019,374683961.719,139830.193
+dens_x1.0,1500000,33713952,1,132205.466,500776754.498,186887.396
+dens_x1.0,1500000,33713952,2,152518.556,500776754.498,186887.396
+dens_x1.0,1500000,33713952,3,112566.137,500776754.498,186887.396
+dens_x2.0,10000,3007,1,26.720,119798.897,25.893
+dens_x2.0,10000,3007,2,43.318,119798.897,25.893
+dens_x2.0,10000,3007,3,31.487,119798.897,25.893
+dens_x2.0,12000,4365,1,48.927,153710.912,33.223
+dens_x2.0,12000,4365,2,34.995,153710.912,33.223
+dens_x2.0,12000,4365,3,46.545,153710.912,33.223
+dens_x2.0,14000,5830,1,49.223,189313.294,40.918
+dens_x2.0,14000,5830,2,37.392,189313.294,40.918
+dens_x2.0,14000,5830,3,56.664,189313.294,40.918
+dens_x2.0,16000,7624,1,46.943,228688.447,49.428
+dens_x2.0,16000,7624,2,57.763,228688.447,49.428
+dens_x2.0,16000,7624,3,47.823,228688.447,49.428
+dens_x2.0,19000,10837,1,98.560,293959.920,63.536
+dens_x2.0,19000,10837,2,62.016,293959.920,63.536
+dens_x2.0,19000,10837,3,58.108,293959.920,63.536
+dens_x2.0,22000,14556,1,101.815,365516.050,79.002
+dens_x2.0,22000,14556,2,63.192,365516.050,79.002
+dens_x2.0,22000,14556,3,93.428,365516.050,79.002
+dens_x2.0,26000,20395,1,109.113,471644.695,101.940
+dens_x2.0,26000,20395,2,82.642,471644.695,101.940
+dens_x2.0,26000,20395,3,117.889,471644.695,101.940
+dens_x2.0,30000,26790,1,94.074,585445.422,126.537
+dens_x2.0,30000,26790,2,143.902,585445.422,126.537
+dens_x2.0,30000,26790,3,124.759,585445.422,126.537
+dens_x2.0,36000,38800,1,162.750,784747.311,169.613
+dens_x2.0,36000,38800,2,202.980,784747.311,169.613
+dens_x2.0,36000,38800,3,142.628,784747.311,169.613
+dens_x2.0,42000,53391,1,206.343,1015477.726,219.483
+dens_x2.0,42000,53391,2,142.426,1015477.726,219.483
+dens_x2.0,42000,53391,3,289.445,1015477.726,219.483
+dens_x2.0,49000,71945,1,264.377,1306154.668,282.309
+dens_x2.0,49000,71945,2,199.305,1306154.668,282.309
+dens_x2.0,49000,71945,3,262.257,1306154.668,282.309
+dens_x2.0,57000,97182,1,307.139,1688417.255,364.931
+dens_x2.0,57000,97182,2,326.031,1688417.255,364.931
+dens_x2.0,57000,97182,3,350.852,1688417.255,364.931
+dens_x2.0,67000,134082,1,247.385,2234513.248,482.962
+dens_x2.0,67000,134082,2,359.392,2234513.248,482.962
+dens_x2.0,67000,134082,3,399.988,2234513.248,482.962
+dens_x2.0,79000,186876,1,342.322,2998337.660,648.054
+dens_x2.0,79000,186876,2,332.860,2998337.660,648.054
+dens_x2.0,79000,186876,3,344.360,2998337.660,648.054
+dens_x2.0,92000,253371,1,379.277,3947432.991,853.189
+dens_x2.0,92000,253371,2,425.224,3947432.991,853.189
+dens_x2.0,92000,253371,3,363.135,3947432.991,853.189
+dens_x2.0,108000,349108,1,491.590,5297829.841,1145.060
+dens_x2.0,108000,349108,2,527.905,5297829.841,1145.060
+dens_x2.0,108000,349108,3,514.033,5297829.841,1145.060
+dens_x2.0,127000,484181,1,822.984,7182563.887,1552.422
+dens_x2.0,127000,484181,2,713.202,7182563.887,1552.422
+dens_x2.0,127000,484181,3,730.441,7182563.887,1552.422
+dens_x2.0,149000,666491,1,902.169,9713885.437,2099.536
+dens_x2.0,149000,666491,2,921.216,9713885.437,2099.536
+dens_x2.0,149000,666491,3,1008.476,9713885.437,2099.536
+dens_x2.0,174000,907603,1,1210.904,13051498.522,2820.920
+dens_x2.0,174000,907603,2,1263.896,13051498.522,2820.920
+dens_x2.0,174000,907603,3,1232.927,13051498.522,2820.920
+dens_x2.0,204000,1247723,1,1851.569,17748584.329,3836.137
+dens_x2.0,204000,1247723,2,1769.742,17748584.329,3836.137
+dens_x2.0,204000,1247723,3,1841.024,17748584.329,3836.137
+dens_x2.0,240000,1725541,1,2987.372,24349896.729,5262.929
+dens_x2.0,240000,1725541,2,2659.560,24349896.729,5262.929
+dens_x2.0,240000,1725541,3,2649.774,24349896.729,5262.929
+dens_x2.0,281000,2365586,1,3782.265,33204358.944,7176.712
+dens_x2.0,281000,2365586,2,4613.097,33204358.944,7176.712
+dens_x2.0,281000,2365586,3,4571.901,33204358.944,7176.712
+dens_x2.0,329000,3242756,1,7011.528,45374920.412,9807.228
+dens_x2.0,329000,3242756,2,6958.774,45374920.412,9807.228
+dens_x2.0,329000,3242756,3,6887.107,45374920.412,9807.228
+dens_x2.0,386000,4461354,1,9100.421,62354387.279,13477.130
+dens_x2.0,386000,4461354,2,11851.489,62354387.279,13477.130
+dens_x2.0,386000,4461354,3,10179.649,62354387.279,13477.130
+dens_x2.0,452000,6119886,1,13421.960,85575402.535,18496.066
+dens_x2.0,452000,6119886,2,18355.367,85575402.535,18496.066
+dens_x2.0,452000,6119886,3,15388.386,85575402.535,18496.066
+dens_x2.0,530000,8413791,1,25985.246,117884820.410,25479.347
+dens_x2.0,530000,8413791,2,18929.372,117884820.410,25479.347
+dens_x2.0,530000,8413791,3,26384.270,117884820.410,25479.347
+dens_x2.0,621000,11552079,1,30976.711,162377752.059,35095.944
+dens_x2.0,621000,11552079,2,31445.815,162377752.059,35095.944
+dens_x2.0,621000,11552079,3,30916.426,162377752.059,35095.944
+dens_x2.0,728000,15870417,1,45213.376,224046367.608,48424.853
+dens_x2.0,728000,15870417,2,46801.158,224046367.608,48424.853
+dens_x2.0,728000,15870417,3,46000.598,224046367.608,48424.853
+dens_x2.0,853000,21800415,1,80293.217,309366697.818,66865.788
+dens_x2.0,853000,21800415,2,72885.247,309366697.818,66865.788
+dens_x2.0,853000,21800415,3,73704.347,309366697.818,66865.788
+dens_x2.0,1000000,29958277,1,93513.717,427704402.750,92443.020
+dens_x2.0,1000000,29958277,2,93710.242,427704402.750,92443.020
+dens_x2.0,1000000,29958277,3,114397.999,427704402.750,92443.020
+dens_x0.5,1300000,12655006,1,36062.526,196456827.614,73316.712
+dens_x0.5,1300000,12655006,2,37491.996,196456827.614,73316.712
+dens_x0.5,1300000,12655006,3,40901.895,196456827.614,73316.712
+dens_x0.5,1500000,16856907,1,55060.072,261053127.751,97423.730
+dens_x0.5,1500000,16856907,2,51489.209,261053127.751,97423.730
+dens_x0.5,1500000,16856907,3,51762.161,261053127.751,97423.730
+dens_x1.0,1300000,25315094,1,98721.978,374683961.719,139830.193
+dens_x1.0,1300000,25315094,2,78676.425,374683961.719,139830.193
+dens_x1.0,1300000,25315094,3,103389.019,374683961.719,139830.193
+dens_x1.0,1500000,33713952,1,132205.466,500776754.498,186887.396
+dens_x1.0,1500000,33713952,2,152518.556,500776754.498,186887.396
+dens_x1.0,1500000,33713952,3,112566.137,500776754.498,186887.396
+dens_x2.0,1300000,50620887,1,269269.465,730935747.856,272781.589
+dens_x2.0,1300000,50620887,2,256085.939,730935747.856,272781.589
+dens_x2.0,1300000,50620887,3,281477.940,730935747.856,272781.589
+dens_x2.0,1500000,67408553,1,490996.350,979946855.397,365711.297
+dens_x2.0,1500000,67408553,2,398152.570,979946855.397,365711.297
+dens_x2.0,1500000,67408553,3,409926.988,979946855.397,365711.297
+dens_x2.0,1300000,50620887,1,269269.465,730935747.856,272781.589
+dens_x2.0,1300000,50620887,2,256085.939,730935747.856,272781.589
+dens_x2.0,1300000,50620887,3,281477.940,730935747.856,272781.589
+dens_x2.0,1500000,67408553,1,490996.350,979946855.397,365711.297
+dens_x2.0,1500000,67408553,2,398152.570,979946855.397,365711.297
+dens_x2.0,1500000,67408553,3,409926.988,979946855.397,365711.297

results_mk_paper.csv

@@ -0,0 +1,91 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms,p_intra,p_inter
+p0.010,10000,6855,1,32.883,168950.287,5.622,0.010000,0.000100000
+p0.010,10000,6855,2,35.347,168950.287,5.622,0.010000,0.000100000
+p0.010,10000,6855,3,35.753,168950.287,5.622,0.010000,0.000100000
+p0.010,12000,9996,1,44.591,226592.992,7.540,0.010000,0.000100000
+p0.010,12000,9996,2,49.257,226592.992,7.540,0.010000,0.000100000
+p0.010,12000,9996,3,46.698,226592.992,7.540,0.010000,0.000100000
+p0.010,14000,13651,1,55.784,291280.915,9.693,0.010000,0.000100000
+p0.010,14000,13651,2,53.707,291280.915,9.693,0.010000,0.000100000
+p0.010,14000,13651,3,55.903,291280.915,9.693,0.010000,0.000100000
+p0.010,16000,17847,1,57.833,381191.603,12.685,0.010000,0.000100000
+p0.010,16000,17847,2,57.146,381191.603,12.685,0.010000,0.000100000
+p0.010,16000,17847,3,54.132,381191.603,12.685,0.010000,0.000100000
+p0.010,19000,25186,1,64.764,510887.055,17.000,0.010000,0.000100000
+p0.010,19000,25186,2,73.329,510887.055,17.000,0.010000,0.000100000
+p0.010,19000,25186,3,76.659,510887.055,17.000,0.010000,0.000100000
+p0.010,22000,34048,1,89.388,662556.615,22.047,0.010000,0.000100000
+p0.010,22000,34048,2,90.601,662556.615,22.047,0.010000,0.000100000
+p0.010,22000,34048,3,90.021,662556.615,22.047,0.010000,0.000100000
+p0.010,26000,47355,1,142.680,935136.060,31.118,0.010000,0.000100000
+p0.010,26000,47355,2,117.901,935136.060,31.118,0.010000,0.000100000
+p0.010,26000,47355,3,134.182,935136.060,31.118,0.010000,0.000100000
+p0.010,31000,66944,1,190.615,1280687.626,42.616,0.010000,0.000100000
+p0.010,31000,66944,2,168.829,1280687.626,42.616,0.010000,0.000100000
+p0.010,31000,66944,3,170.883,1280687.626,42.616,0.010000,0.000100000
+p0.010,37000,95444,1,215.912,1870355.152,62.238,0.010000,0.000100000
+p0.010,37000,95444,2,268.993,1870355.152,62.238,0.010000,0.000100000
+p0.010,37000,95444,3,242.674,1870355.152,62.238,0.010000,0.000100000
+p0.010,43000,129467,1,265.009,2616844.730,87.079,0.010000,0.000100000
+p0.010,43000,129467,2,283.592,2616844.730,87.079,0.010000,0.000100000
+p0.010,43000,129467,3,254.891,2616844.730,87.079,0.010000,0.000100000
+p0.010,51000,181666,1,333.548,3793663.932,126.239,0.010000,0.000100000
+p0.010,51000,181666,2,349.489,3793663.932,126.239,0.010000,0.000100000
+p0.010,51000,181666,3,410.205,3793663.932,126.239,0.010000,0.000100000
+p0.010,59000,242699,1,503.184,5255843.829,174.895,0.010000,0.000100000
+p0.010,59000,242699,2,381.119,5255843.829,174.895,0.010000,0.000100000
+p0.010,59000,242699,3,478.618,5255843.829,174.895,0.010000,0.000100000
+p0.010,70000,340953,1,518.918,7994224.620,266.018,0.010000,0.000100000
+p0.010,70000,340953,2,594.535,7994224.620,266.018,0.010000,0.000100000
+p0.010,70000,340953,3,581.086,7994224.620,266.018,0.010000,0.000100000
+p0.010,82000,468721,1,739.602,11855770.726,394.516,0.010000,0.000100000
+p0.010,82000,468721,2,671.713,11855770.726,394.516,0.010000,0.000100000
+p0.010,82000,468721,3,737.615,11855770.726,394.516,0.010000,0.000100000
+p0.010,97000,657203,1,926.296,18518155.499,616.215,0.010000,0.000100000
+p0.010,97000,657203,2,954.613,18518155.499,616.215,0.010000,0.000100000
+p0.010,97000,657203,3,1049.569,18518155.499,616.215,0.010000,0.000100000
+p0.010,114000,905961,1,1377.284,27277715.688,907.700,0.010000,0.000100000
+p0.010,114000,905961,2,1347.628,27277715.688,907.700,0.010000,0.000100000
+p0.010,114000,905961,3,1253.559,27277715.688,907.700,0.010000,0.000100000
+p0.010,134000,1254092,1,1900.711,42723184.084,1421.667,0.010000,0.000100000
+p0.010,134000,1254092,2,1952.478,42723184.084,1421.667,0.010000,0.000100000
+p0.010,134000,1254092,3,1959.497,42723184.084,1421.667,0.010000,0.000100000
+p0.010,157000,1721008,1,2679.850,65493689.748,2179.384,0.010000,0.000100000
+p0.010,157000,1721008,2,2582.330,65493689.748,2179.384,0.010000,0.000100000
+p0.010,157000,1721008,3,2514.934,65493689.748,2179.384,0.010000,0.000100000
+p0.010,185000,2387648,1,4411.791,102830800.776,3421.823,0.010000,0.000100000
+p0.010,185000,2387648,2,4158.479,102830800.776,3421.823,0.010000,0.000100000
+p0.010,185000,2387648,3,4728.596,102830800.776,3421.823,0.010000,0.000100000
+p0.010,218000,3318758,1,7240.039,162950002.734,5422.364,0.010000,0.000100000
+p0.010,218000,3318758,2,7185.162,162950002.734,5422.364,0.010000,0.000100000
+p0.010,218000,3318758,3,7265.966,162950002.734,5422.364,0.010000,0.000100000
+p0.010,256000,4576568,1,10987.979,252395500.186,8398.775,0.010000,0.000100000
+p0.010,256000,4576568,2,10888.895,252395500.186,8398.775,0.010000,0.000100000
+p0.010,256000,4576568,3,10992.033,252395500.186,8398.775,0.010000,0.000100000
+p0.010,301000,6321401,1,14079.826,405932149.192,13507.898,0.010000,0.000100000
+p0.010,301000,6321401,2,16659.217,405932149.192,13507.898,0.010000,0.000100000
+p0.010,301000,6321401,3,16325.723,405932149.192,13507.898,0.010000,0.000100000
+p0.010,354000,8746503,1,25978.646,649814284.806,21623.380,0.010000,0.000100000
+p0.010,354000,8746503,2,26495.244,649814284.806,21623.380,0.010000,0.000100000
+p0.010,354000,8746503,3,25287.206,649814284.806,21623.380,0.010000,0.000100000
+p0.010,416000,12073015,1,32732.342,1030845457.971,34302.668,0.010000,0.000100000
+p0.010,416000,12073015,2,28855.088,1030845457.971,34302.668,0.010000,0.000100000
+p0.010,416000,12073015,3,29107.222,1030845457.971,34302.668,0.010000,0.000100000
+p0.010,489000,16686318,1,46434.755,1660022036.511,55239.303,0.010000,0.000100000
+p0.010,489000,16686318,2,50506.399,1660022036.511,55239.303,0.010000,0.000100000
+p0.010,489000,16686318,3,45908.078,1660022036.511,55239.303,0.010000,0.000100000
+p0.010,575000,23083981,1,67075.956,2691418413.960,89560.303,0.010000,0.000100000
+p0.010,575000,23083981,2,72917.385,2691418413.960,89560.303,0.010000,0.000100000
+p0.010,575000,23083981,3,77338.323,2691418413.960,89560.303,0.010000,0.000100000
+p0.010,676000,31898369,1,130134.315,4328877665.154,144048.800,0.010000,0.000100000
+p0.010,676000,31898369,2,148812.790,4328877665.154,144048.800,0.010000,0.000100000
+p0.010,676000,31898369,3,136450.553,4328877665.154,144048.800,0.010000,0.000100000
+p0.010,795000,44104536,1,232507.154,7005167189.025,233105.670,0.010000,0.000100000
+p0.010,795000,44104536,2,173564.025,7005167189.025,233105.670,0.010000,0.000100000
+p0.010,795000,44104536,3,202268.941,7005167189.025,233105.670,0.010000,0.000100000
+p0.010,935000,61021179,1,389899.582,11408456214.779,379630.602,0.010000,0.000100000
+p0.010,935000,61021179,2,392860.546,11408456214.779,379630.602,0.010000,0.000100000
+p0.010,935000,61021179,3,349062.020,11408456214.779,379630.602,0.010000,0.000100000
+p0.010,1100000,84452506,1,576853.282,18502869304.633,615706.040,0.010000,0.000100000
+p0.010,1100000,84452506,2,674198.550,18502869304.633,615706.040,0.010000,0.000100000
+p0.010,1100000,84452506,3,648210.917,18502869304.633,615706.040,0.010000,0.000100000

results_mk_paper2.csv

@@ -0,0 +1,7 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms,p_intra,p_inter
+p0.010,894000,55784686,1,434722.717,9981167356.202,415219.440,0.010000,0.000100000
+p0.010,894000,55784686,2,443367.313,9981167356.202,415219.440,0.010000,0.000100000
+p0.010,894000,55784686,3,405549.165,9981167356.202,415219.440,0.010000,0.000100000
+p0.010,1050000,76942931,1,769922.022,16085188994.585,669148.500,0.010000,0.000100000
+p0.010,1050000,76942931,2,633155.668,16085188994.585,669148.500,0.010000,0.000100000
+p0.010,1050000,76942931,3,580799.949,16085188994.585,669148.500,0.010000,0.000100000

results_mk_peeling.csv

@@ -0,0 +1,91 @@
+series,n,m,trial,ms_lrmc,ms_charikar,ratio_char_over_lrmc,p_intra,p_inter
+p0.010,10000,6855,1,31.992,13.314,0.4162,0.010000,0.000100000
+p0.010,10000,6855,2,31.391,5.962,0.1899,0.010000,0.000100000
+p0.010,10000,6855,3,32.466,3.270,0.1007,0.010000,0.000100000
+p0.010,12000,9996,1,43.876,4.399,0.1003,0.010000,0.000100000
+p0.010,12000,9996,2,45.240,3.439,0.0760,0.010000,0.000100000
+p0.010,12000,9996,3,47.128,3.028,0.0643,0.010000,0.000100000
+p0.010,14000,13651,1,56.129,8.664,0.1544,0.010000,0.000100000
+p0.010,14000,13651,2,52.710,4.570,0.0867,0.010000,0.000100000
+p0.010,14000,13651,3,53.127,4.179,0.0787,0.010000,0.000100000
+p0.010,16000,17847,1,55.469,5.666,0.1022,0.010000,0.000100000
+p0.010,16000,17847,2,59.261,5.577,0.0941,0.010000,0.000100000
+p0.010,16000,17847,3,52.821,5.330,0.1009,0.010000,0.000100000
+p0.010,19000,25186,1,63.495,7.508,0.1182,0.010000,0.000100000
+p0.010,19000,25186,2,70.504,7.240,0.1027,0.010000,0.000100000
+p0.010,19000,25186,3,62.448,6.297,0.1008,0.010000,0.000100000
+p0.010,22000,34048,1,90.847,8.440,0.0929,0.010000,0.000100000
+p0.010,22000,34048,2,76.658,10.948,0.1428,0.010000,0.000100000
+p0.010,22000,34048,3,80.600,9.722,0.1206,0.010000,0.000100000
+p0.010,26000,47355,1,141.174,12.880,0.0912,0.010000,0.000100000
+p0.010,26000,47355,2,119.388,13.316,0.1115,0.010000,0.000100000
+p0.010,26000,47355,3,118.538,13.578,0.1145,0.010000,0.000100000
+p0.010,31000,66944,1,186.135,17.147,0.0921,0.010000,0.000100000
+p0.010,31000,66944,2,170.393,16.096,0.0945,0.010000,0.000100000
+p0.010,31000,66944,3,179.323,15.829,0.0883,0.010000,0.000100000
+p0.010,36000,90395,1,202.701,24.991,0.1233,0.010000,0.000100000
+p0.010,36000,90395,2,251.185,24.376,0.0970,0.010000,0.000100000
+p0.010,36000,90395,3,203.004,24.936,0.1228,0.010000,0.000100000
+p0.010,42000,123740,1,261.486,32.913,0.1259,0.010000,0.000100000
+p0.010,42000,123740,2,236.691,35.356,0.1494,0.010000,0.000100000
+p0.010,42000,123740,3,364.157,32.413,0.0890,0.010000,0.000100000
+p0.010,50000,174903,1,430.400,44.907,0.1043,0.010000,0.000100000
+p0.010,50000,174903,2,458.256,47.685,0.1041,0.010000,0.000100000
+p0.010,50000,174903,3,332.602,42.574,0.1280,0.010000,0.000100000
+p0.010,58000,234422,1,606.597,65.650,0.1082,0.010000,0.000100000
+p0.010,58000,234422,2,502.739,58.246,0.1159,0.010000,0.000100000
+p0.010,58000,234422,3,477.328,59.802,0.1253,0.010000,0.000100000
+p0.010,69000,331231,1,559.928,81.968,0.1464,0.010000,0.000100000
+p0.010,69000,331231,2,514.083,86.528,0.1683,0.010000,0.000100000
+p0.010,69000,331231,3,601.672,95.934,0.1594,0.010000,0.000100000
+p0.010,81000,457271,1,675.088,119.821,0.1775,0.010000,0.000100000
+p0.010,81000,457271,2,641.671,114.564,0.1785,0.010000,0.000100000
+p0.010,81000,457271,3,651.985,119.556,0.1834,0.010000,0.000100000
+p0.010,95000,629660,1,928.485,162.729,0.1753,0.010000,0.000100000
+p0.010,95000,629660,2,869.368,161.968,0.1863,0.010000,0.000100000
+p0.010,95000,629660,3,876.931,165.358,0.1886,0.010000,0.000100000
+p0.010,111000,858910,1,1290.964,232.245,0.1799,0.010000,0.000100000
+p0.010,111000,858910,2,1210.956,223.658,0.1847,0.010000,0.000100000
+p0.010,111000,858910,3,1233.032,235.343,0.1909,0.010000,0.000100000
+p0.010,130000,1180808,1,1835.596,302.229,0.1646,0.010000,0.000100000
+p0.010,130000,1180808,2,1728.640,301.218,0.1743,0.010000,0.000100000
+p0.010,130000,1180808,3,1894.693,323.953,0.1710,0.010000,0.000100000
+p0.010,153000,1635104,1,2434.417,481.471,0.1978,0.010000,0.000100000
+p0.010,153000,1635104,2,2835.456,565.099,0.1993,0.010000,0.000100000
+p0.010,153000,1635104,3,2509.212,462.239,0.1842,0.010000,0.000100000
+p0.010,180000,2259463,1,3470.979,753.372,0.2170,0.010000,0.000100000
+p0.010,180000,2259463,2,3840.293,888.440,0.2313,0.010000,0.000100000
+p0.010,180000,2259463,3,4039.016,774.241,0.1917,0.010000,0.000100000
+p0.010,211000,3107369,1,5367.980,1196.096,0.2228,0.010000,0.000100000
+p0.010,211000,3107369,2,5169.248,1260.174,0.2438,0.010000,0.000100000
+p0.010,211000,3107369,3,5381.383,1125.088,0.2091,0.010000,0.000100000
+p0.010,248000,4294862,1,8809.494,1714.655,0.1946,0.010000,0.000100000
+p0.010,248000,4294862,2,10826.117,1550.755,0.1432,0.010000,0.000100000
+p0.010,248000,4294862,3,9121.021,1334.802,0.1463,0.010000,0.000100000
+p0.010,291000,5909593,1,11920.301,1879.288,0.1577,0.010000,0.000100000
+p0.010,291000,5909593,2,12681.672,1886.153,0.1487,0.010000,0.000100000
+p0.010,291000,5909593,3,13008.623,1904.175,0.1464,0.010000,0.000100000
+p0.010,341000,8116027,1,20886.926,2635.949,0.1262,0.010000,0.000100000
+p0.010,341000,8116027,2,23657.120,2641.528,0.1117,0.010000,0.000100000
+p0.010,341000,8116027,3,24323.864,2660.392,0.1094,0.010000,0.000100000
+p0.010,401000,11224438,1,26442.424,3698.956,0.1399,0.010000,0.000100000
+p0.010,401000,11224438,2,27944.718,3785.824,0.1355,0.010000,0.000100000
+p0.010,401000,11224438,3,35573.425,4131.695,0.1161,0.010000,0.000100000
+p0.010,471000,15478323,1,46818.609,5235.147,0.1118,0.010000,0.000100000
+p0.010,471000,15478323,2,43437.508,5320.091,0.1225,0.010000,0.000100000
+p0.010,471000,15478323,3,44888.857,5216.460,0.1162,0.010000,0.000100000
+p0.010,553000,21347547,1,70903.340,7237.961,0.1021,0.010000,0.000100000
+p0.010,553000,21347547,2,86748.593,7565.931,0.0872,0.010000,0.000100000
+p0.010,553000,21347547,3,73878.376,7251.312,0.0982,0.010000,0.000100000
+p0.010,649000,29406812,1,110734.007,10133.955,0.0915,0.010000,0.000100000
+p0.010,649000,29406812,2,98126.391,10132.573,0.1033,0.010000,0.000100000
+p0.010,649000,29406812,3,94134.658,10104.309,0.1073,0.010000,0.000100000
+p0.010,762000,40512199,1,238696.389,14388.104,0.0603,0.010000,0.000100000
+p0.010,762000,40512199,2,197853.014,14915.448,0.0754,0.010000,0.000100000
+p0.010,762000,40512199,3,185054.047,14096.983,0.0762,0.010000,0.000100000
+p0.010,894000,55788477,1,342670.024,19924.599,0.0581,0.010000,0.000100000
+p0.010,894000,55788477,2,514606.529,19798.917,0.0385,0.010000,0.000100000
+p0.010,894000,55788477,3,448368.373,20717.981,0.0462,0.010000,0.000100000
+p0.010,1050000,76952553,1,868027.280,37071.813,0.0427,0.010000,0.000100000
+p0.010,1050000,76952553,2,519475.395,38502.183,0.0741,0.010000,0.000100000
+p0.010,1050000,76952553,3,511793.094,35704.567,0.0698,0.010000,0.000100000

results_nosquared.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,13.928,99112.473,32.641
+dens_x0.5,10000,761,2,14.427,99112.473,32.641
+dens_x0.5,10000,761,3,14.068,99112.473,32.641
+dens_x0.5,12000,1077,1,24.819,122827.840,40.451
+dens_x0.5,12000,1077,2,18.116,122827.840,40.451
+dens_x0.5,12000,1077,3,17.827,122827.840,40.451
+dens_x0.5,14000,1455,1,22.211,147545.989,48.592
+dens_x0.5,14000,1455,2,22.745,147545.989,48.592
+dens_x0.5,14000,1455,3,21.998,147545.989,48.592
+dens_x0.5,16000,1937,1,25.174,173636.330,57.184
+dens_x0.5,16000,1937,2,24.886,173636.330,57.184
+dens_x0.5,16000,1937,3,25.206,173636.330,57.184
+dens_x0.5,19000,2715,1,30.350,213940.398,70.458
+dens_x0.5,19000,2715,2,27.632,213940.398,70.458
+dens_x0.5,19000,2715,3,30.365,213940.398,70.458
+dens_x0.5,22000,3611,1,37.591,256079.209,84.335
+dens_x0.5,22000,3611,2,39.308,256079.209,84.335
+dens_x0.5,22000,3611,3,37.304,256079.209,84.335
+dens_x0.5,26000,5194,1,49.215,317113.582,104.436
+dens_x0.5,26000,5194,2,45.546,317113.582,104.436
+dens_x0.5,26000,5194,3,46.041,317113.582,104.436
+dens_x0.5,30000,6763,1,43.645,378988.027,124.813
+dens_x0.5,30000,6763,2,45.099,378988.027,124.813
+dens_x0.5,30000,6763,3,45.535,378988.027,124.813
+dens_x0.5,36000,9660,1,47.761,479031.581,157.761
+dens_x0.5,36000,9660,2,67.519,479031.581,157.761
+dens_x0.5,36000,9660,3,48.563,479031.581,157.761
+dens_x0.5,42000,13346,1,50.651,589181.686,194.037
+dens_x0.5,42000,13346,2,67.829,589181.686,194.037
+dens_x0.5,42000,13346,3,58.132,589181.686,194.037
+dens_x0.5,49000,17900,1,71.334,722491.606,237.940
+dens_x0.5,49000,17900,2,81.959,722491.606,237.940
+dens_x0.5,49000,17900,3,175.205,722491.606,237.940
+dens_x0.5,57000,24363,1,100.968,890990.473,293.433
+dens_x0.5,57000,24363,2,102.184,890990.473,293.433
+dens_x0.5,57000,24363,3,101.227,890990.473,293.433
+dens_x0.5,67000,33829,1,119.900,1120457.009,369.004
+dens_x0.5,67000,33829,2,131.959,1120457.009,369.004
+dens_x0.5,67000,33829,3,133.951,1120457.009,369.004
+dens_x0.5,79000,46917,1,165.619,1419991.587,467.650
+dens_x0.5,79000,46917,2,165.403,1419991.587,467.650
+dens_x0.5,79000,46917,3,161.530,1419991.587,467.650
+dens_x0.5,92000,63864,1,200.791,1781454.424,586.692
+dens_x0.5,92000,63864,2,255.603,1781454.424,586.692
+dens_x0.5,92000,63864,3,233.085,1781454.424,586.692
+dens_x0.5,108000,87350,1,258.690,2264084.329,745.638
+dens_x0.5,108000,87350,2,279.017,2264084.329,745.638
+dens_x0.5,108000,87350,3,209.424,2264084.329,745.638
+dens_x0.5,127000,120912,1,296.722,2913447.536,959.495
+dens_x0.5,127000,120912,2,297.731,2913447.536,959.495
+dens_x0.5,127000,120912,3,323.214,2913447.536,959.495
+dens_x0.5,149000,166365,1,389.742,3756533.770,1237.151
+dens_x0.5,149000,166365,2,315.302,3756533.770,1237.151
+dens_x0.5,149000,166365,3,369.940,3756533.770,1237.151
+dens_x0.5,174000,226551,1,384.883,4833373.044,1591.789
+dens_x0.5,174000,226551,2,474.197,4833373.044,1591.789
+dens_x0.5,174000,226551,3,385.850,4833373.044,1591.789
+dens_x0.5,204000,312325,1,513.323,6312525.050,2078.923
+dens_x0.5,204000,312325,2,548.002,6312525.050,2078.923
+dens_x0.5,204000,312325,3,561.628,6312525.050,2078.923
+dens_x0.5,240000,431142,1,707.811,8314371.662,2738.197
+dens_x0.5,240000,431142,2,649.676,8314371.662,2738.197
+dens_x0.5,240000,431142,3,701.845,8314371.662,2738.197
+dens_x0.5,281000,589828,1,888.249,10925503.834,3598.130
+dens_x0.5,281000,589828,2,891.876,10925503.834,3598.130
+dens_x0.5,281000,589828,3,908.785,10925503.834,3598.130
+dens_x0.5,329000,811844,1,1222.536,14493068.872,4773.047
+dens_x0.5,329000,811844,2,1304.879,14493068.872,4773.047
+dens_x0.5,329000,811844,3,1212.275,14493068.872,4773.047
+dens_x0.5,386000,1115568,1,2245.253,19315559.086,6361.252
+dens_x0.5,386000,1115568,2,2000.723,19315559.086,6361.252
+dens_x0.5,386000,1115568,3,2589.654,19315559.086,6361.252
+dens_x0.5,452000,1531464,1,3012.595,25827552.428,8505.867
+dens_x0.5,452000,1531464,2,2750.113,25827552.428,8505.867
+dens_x0.5,452000,1531464,3,2729.488,25827552.428,8505.867
+dens_x0.5,530000,2103874,1,3809.212,34716124.680,11433.167
+dens_x0.5,530000,2103874,2,4309.380,34716124.680,11433.167
+dens_x0.5,530000,2103874,3,4481.890,34716124.680,11433.167
+dens_x0.5,621000,2887681,1,6422.418,46802598.872,15413.643
+dens_x0.5,621000,2887681,2,6585.869,46802598.872,15413.643
+dens_x0.5,621000,2887681,3,6765.516,46802598.872,15413.643
+dens_x0.5,728000,3971764,1,11958.752,63437679.196,20892.124
+dens_x0.5,728000,3971764,2,10308.737,63437679.196,20892.124
+dens_x0.5,728000,3971764,3,11056.578,63437679.196,20892.124
+dens_x0.5,853000,5448500,1,14109.654,86056528.179,28341.259
+dens_x0.5,853000,5448500,2,14769.758,86056528.179,28341.259
+dens_x0.5,853000,5448500,3,14404.127,86056528.179,28341.259
+dens_x0.5,1000000,7489974,1,24320.614,117293325.434,38628.569
+dens_x0.5,1000000,7489974,2,24047.511,117293325.434,38628.569
+dens_x0.5,1000000,7489974,3,22553.170,117293325.434,38628.569
+dens_x1.0,10000,1550,1,15.795,106379.431,35.034
+dens_x1.0,10000,1550,2,17.105,106379.431,35.034
+dens_x1.0,10000,1550,3,17.056,106379.431,35.034
+dens_x1.0,12000,2135,1,20.849,132765.276,43.724
+dens_x1.0,12000,2135,2,21.907,132765.276,43.724
+dens_x1.0,12000,2135,3,22.254,132765.276,43.724
+dens_x1.0,14000,2963,1,28.417,161942.582,53.333
+dens_x1.0,14000,2963,2,27.935,161942.582,53.333
+dens_x1.0,14000,2963,3,27.837,161942.582,53.333
+dens_x1.0,16000,3912,1,32.832,192755.010,63.481
+dens_x1.0,16000,3912,2,32.494,192755.010,63.481
+dens_x1.0,16000,3912,3,31.028,192755.010,63.481
+dens_x1.0,19000,5413,1,40.669,240521.618,79.212
+dens_x1.0,19000,5413,2,39.997,240521.618,79.212
+dens_x1.0,19000,5413,3,43.415,240521.618,79.212
+dens_x1.0,22000,7290,1,40.898,292864.786,96.450
+dens_x1.0,22000,7290,2,68.058,292864.786,96.450
+dens_x1.0,22000,7290,3,37.365,292864.786,96.450
+dens_x1.0,26000,10078,1,49.070,366763.602,120.787
+dens_x1.0,26000,10078,2,51.834,366763.602,120.787
+dens_x1.0,26000,10078,3,46.146,366763.602,120.787
+dens_x1.0,30000,13336,1,80.316,446748.773,147.129
+dens_x1.0,30000,13336,2,74.729,446748.773,147.129
+dens_x1.0,30000,13336,3,57.861,446748.773,147.129
+dens_x1.0,36000,19373,1,83.950,580933.327,191.321
+dens_x1.0,36000,19373,2,90.415,580933.327,191.321
+dens_x1.0,36000,19373,3,77.043,580933.327,191.321
+dens_x1.0,42000,26057,1,131.538,724495.682,238.600
+dens_x1.0,42000,26057,2,94.107,724495.682,238.600
+dens_x1.0,42000,26057,3,157.959,724495.682,238.600
+dens_x1.0,49000,35903,1,121.913,916916.365,301.971
+dens_x1.0,49000,35903,2,142.426,916916.365,301.971
+dens_x1.0,49000,35903,3,142.150,916916.365,301.971
+dens_x1.0,57000,48648,1,169.368,1156930.810,381.016
+dens_x1.0,57000,48648,2,170.218,1156930.810,381.016
+dens_x1.0,57000,48648,3,171.756,1156930.810,381.016
+dens_x1.0,67000,67188,1,214.470,1491157.159,491.087
+dens_x1.0,67000,67188,2,208.270,1491157.159,491.087
+dens_x1.0,67000,67188,3,220.916,1491157.159,491.087
+dens_x1.0,79000,93552,1,237.551,1945903.955,640.850
+dens_x1.0,79000,93552,2,188.477,1945903.955,640.850
+dens_x1.0,79000,93552,3,184.641,1945903.955,640.850
+dens_x1.0,92000,126776,1,220.118,2500509.887,823.501
+dens_x1.0,92000,126776,2,274.676,2500509.887,823.501
+dens_x1.0,92000,126776,3,223.749,2500509.887,823.501
+dens_x1.0,108000,175020,1,287.105,3280169.679,1080.268
+dens_x1.0,108000,175020,2,312.388,3280169.679,1080.268
+dens_x1.0,108000,175020,3,281.471,3280169.679,1080.268
+dens_x1.0,127000,241838,1,374.493,4334562.918,1427.515
+dens_x1.0,127000,241838,2,358.238,4334562.918,1427.515
+dens_x1.0,127000,241838,3,408.865,4334562.918,1427.515
+dens_x1.0,149000,333094,1,487.476,5742559.864,1891.215
+dens_x1.0,149000,333094,2,564.073,5742559.864,1891.215
+dens_x1.0,149000,333094,3,462.038,5742559.864,1891.215
+dens_x1.0,174000,454003,1,625.079,7577993.244,2495.684
+dens_x1.0,174000,454003,2,720.059,7577993.244,2495.684
+dens_x1.0,174000,454003,3,753.227,7577993.244,2495.684
+dens_x1.0,204000,623499,1,1540.281,10116899.562,3331.829
+dens_x1.0,204000,623499,2,1002.571,10116899.562,3331.829
+dens_x1.0,204000,623499,3,1068.854,10116899.562,3331.829
+dens_x1.0,240000,862431,1,1298.144,13657349.809,4497.817
+dens_x1.0,240000,862431,2,1314.402,13657349.809,4497.817
+dens_x1.0,240000,862431,3,1310.847,13657349.809,4497.817
+dens_x1.0,281000,1182168,1,2025.112,18357066.601,6045.589
+dens_x1.0,281000,1182168,2,2083.471,18357066.601,6045.589
+dens_x1.0,281000,1182168,3,1986.419,18357066.601,6045.589
+dens_x1.0,329000,1623777,1,3467.966,24807713.897,8170.000
+dens_x1.0,329000,1623777,2,2775.287,24807713.897,8170.000
+dens_x1.0,329000,1623777,3,2566.352,24807713.897,8170.000
+dens_x1.0,386000,2232671,1,4668.748,33685517.024,11093.754
+dens_x1.0,386000,2232671,2,3965.444,33685517.024,11093.754
+dens_x1.0,386000,2232671,3,4808.141,33685517.024,11093.754
+dens_x1.0,452000,3055874,1,7017.716,45677561.904,15043.131
+dens_x1.0,452000,3055874,2,5722.622,45677561.904,15043.131
+dens_x1.0,452000,3055874,3,6966.270,45677561.904,15043.131
+dens_x1.0,530000,4208065,1,9001.121,62450692.510,20567.077
+dens_x1.0,530000,4208065,2,10557.556,62450692.510,20567.077
+dens_x1.0,530000,4208065,3,10740.295,62450692.510,20567.077
+dens_x1.0,621000,5774507,1,14674.393,85310220.195,28095.475
+dens_x1.0,621000,5774507,2,16115.354,85310220.195,28095.475
+dens_x1.0,621000,5774507,3,15775.575,85310220.195,28095.475
+dens_x1.0,728000,7935888,1,25390.090,116945648.236,38514.067
+dens_x1.0,728000,7935888,2,20846.516,116945648.236,38514.067
+dens_x1.0,728000,7935888,3,21598.478,116945648.236,38514.067
+dens_x1.0,853000,10900416,1,41442.053,160510699.866,52861.478
+dens_x1.0,853000,10900416,2,42161.875,160510699.866,52861.478
+dens_x1.0,853000,10900416,3,42103.177,160510699.866,52861.478
+dens_x1.0,1000000,14986435,1,53008.511,220860761.527,72736.748
+dens_x1.0,1000000,14986435,2,52939.177,220860761.527,72736.748
+dens_x1.0,1000000,14986435,3,51968.340,220860761.527,72736.748
+dens_x2.0,10000,3007,1,25.919,119798.897,39.454
+dens_x2.0,10000,3007,2,19.887,119798.897,39.454
+dens_x2.0,10000,3007,3,23.027,119798.897,39.454
+dens_x2.0,12000,4365,1,27.838,153710.912,50.622
+dens_x2.0,12000,4365,2,29.499,153710.912,50.622
+dens_x2.0,12000,4365,3,28.616,153710.912,50.622
+dens_x2.0,14000,5830,1,37.540,189313.294,62.347
+dens_x2.0,14000,5830,2,31.680,189313.294,62.347
+dens_x2.0,14000,5830,3,35.764,189313.294,62.347
+dens_x2.0,16000,7624,1,32.257,228688.447,75.315
+dens_x2.0,16000,7624,2,39.717,228688.447,75.315
+dens_x2.0,16000,7624,3,31.113,228688.447,75.315
+dens_x2.0,19000,10837,1,41.738,293959.920,96.811
+dens_x2.0,19000,10837,2,39.956,293959.920,96.811
+dens_x2.0,19000,10837,3,43.791,293959.920,96.811
+dens_x2.0,22000,14556,1,49.073,365516.050,120.377
+dens_x2.0,22000,14556,2,49.029,365516.050,120.377
+dens_x2.0,22000,14556,3,48.099,365516.050,120.377
+dens_x2.0,26000,20395,1,76.849,471644.695,155.328
+dens_x2.0,26000,20395,2,76.294,471644.695,155.328
+dens_x2.0,26000,20395,3,54.920,471644.695,155.328
+dens_x2.0,30000,26790,1,91.247,585445.422,192.807
+dens_x2.0,30000,26790,2,91.468,585445.422,192.807
+dens_x2.0,30000,26790,3,90.817,585445.422,192.807
+dens_x2.0,36000,38800,1,130.781,784747.311,258.443
+dens_x2.0,36000,38800,2,128.351,784747.311,258.443
+dens_x2.0,36000,38800,3,130.023,784747.311,258.443
+dens_x2.0,42000,53391,1,157.755,1015477.726,334.430
+dens_x2.0,42000,53391,2,156.454,1015477.726,334.430
+dens_x2.0,42000,53391,3,172.249,1015477.726,334.430
+dens_x2.0,49000,71945,1,231.629,1306154.668,430.160
+dens_x2.0,49000,71945,2,201.577,1306154.668,430.160
+dens_x2.0,49000,71945,3,229.360,1306154.668,430.160
+dens_x2.0,57000,97182,1,165.994,1688417.255,556.052
+dens_x2.0,57000,97182,2,249.906,1688417.255,556.052
+dens_x2.0,57000,97182,3,226.709,1688417.255,556.052
+dens_x2.0,67000,134082,1,268.547,2234513.248,735.899
+dens_x2.0,67000,134082,2,211.378,2234513.248,735.899
+dens_x2.0,67000,134082,3,217.940,2234513.248,735.899
+dens_x2.0,79000,186876,1,268.171,2998337.660,987.452
+dens_x2.0,79000,186876,2,270.091,2998337.660,987.452
+dens_x2.0,79000,186876,3,366.576,2998337.660,987.452
+dens_x2.0,92000,253371,1,379.773,3947432.991,1300.020
+dens_x2.0,92000,253371,2,457.191,3947432.991,1300.020
+dens_x2.0,92000,253371,3,438.751,3947432.991,1300.020
+dens_x2.0,108000,349108,1,489.690,5297829.841,1744.750
+dens_x2.0,108000,349108,2,546.835,5297829.841,1744.750
+dens_x2.0,108000,349108,3,592.548,5297829.841,1744.750
+dens_x2.0,127000,484181,1,648.164,7182563.887,2365.456
+dens_x2.0,127000,484181,2,661.210,7182563.887,2365.456
+dens_x2.0,127000,484181,3,683.688,7182563.887,2365.456
+dens_x2.0,149000,666491,1,941.120,9713885.437,3199.104
+dens_x2.0,149000,666491,2,939.268,9713885.437,3199.104
+dens_x2.0,149000,666491,3,1079.986,9713885.437,3199.104
+dens_x2.0,174000,907603,1,1457.898,13051498.522,4298.290
+dens_x2.0,174000,907603,2,1491.741,13051498.522,4298.290
+dens_x2.0,174000,907603,3,1509.796,13051498.522,4298.290
+dens_x2.0,204000,1247723,1,1941.400,17748584.329,5845.195
+dens_x2.0,204000,1247723,2,1883.289,17748584.329,5845.195
+dens_x2.0,204000,1247723,3,2148.763,17748584.329,5845.195
+dens_x2.0,240000,1725541,1,3554.892,24349896.729,8019.226
+dens_x2.0,240000,1725541,2,3500.571,24349896.729,8019.226
+dens_x2.0,240000,1725541,3,2982.789,24349896.729,8019.226
+dens_x2.0,281000,2365586,1,5271.622,33204358.944,10935.293
+dens_x2.0,281000,2365586,2,5262.823,33204358.944,10935.293
+dens_x2.0,281000,2365586,3,5355.604,33204358.944,10935.293
+dens_x2.0,329000,3242756,1,6945.261,45374920.412,14943.461
+dens_x2.0,329000,3242756,2,8169.639,45374920.412,14943.461
+dens_x2.0,329000,3242756,3,8000.158,45374920.412,14943.461
+dens_x2.0,386000,4461354,1,12528.673,62354387.279,20535.361
+dens_x2.0,386000,4461354,2,12562.673,62354387.279,20535.361
+dens_x2.0,386000,4461354,3,12035.052,62354387.279,20535.361
+dens_x2.0,452000,6119886,1,20186.465,85575402.535,28182.808
+dens_x2.0,452000,6119886,2,19916.812,85575402.535,28182.808
+dens_x2.0,452000,6119886,3,15977.399,85575402.535,28182.808
+dens_x2.0,530000,8413791,1,32317.410,117884820.410,38823.367
+dens_x2.0,530000,8413791,2,31720.112,117884820.410,38823.367
+dens_x2.0,530000,8413791,3,31302.338,117884820.410,38823.367
+dens_x2.0,621000,11552079,1,53605.814,162377752.059,53476.360
+dens_x2.0,621000,11552079,2,52721.452,162377752.059,53476.360
+dens_x2.0,621000,11552079,3,53208.675,162377752.059,53476.360
+dens_x2.0,728000,15870417,1,88661.318,224046367.608,73785.874
+dens_x2.0,728000,15870417,2,89084.669,224046367.608,73785.874
+dens_x2.0,728000,15870417,3,92250.894,224046367.608,73785.874
+dens_x2.0,853000,21800415,1,103414.772,309366697.818,101884.678
+dens_x2.0,853000,21800415,2,108232.648,309366697.818,101884.678
+dens_x2.0,853000,21800415,3,109812.620,309366697.818,101884.678
+dens_x2.0,1000000,29958277,1,177792.954,427704402.750,140857.196
+dens_x2.0,1000000,29958277,2,167656.235,427704402.750,140857.196
+dens_x2.0,1000000,29958277,3,168312.820,427704402.750,140857.196

results_nosquared2.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,15.467,99112.473,26.023
+dens_x0.5,10000,761,2,15.346,99112.473,26.023
+dens_x0.5,10000,761,3,15.207,99112.473,26.023
+dens_x0.5,12000,1077,1,18.413,122827.840,32.250
+dens_x0.5,12000,1077,2,19.956,122827.840,32.250
+dens_x0.5,12000,1077,3,18.865,122827.840,32.250
+dens_x0.5,14000,1455,1,28.436,147545.989,38.740
+dens_x0.5,14000,1455,2,21.738,147545.989,38.740
+dens_x0.5,14000,1455,3,22.054,147545.989,38.740
+dens_x0.5,16000,1937,1,23.350,173636.330,45.590
+dens_x0.5,16000,1937,2,24.948,173636.330,45.590
+dens_x0.5,16000,1937,3,23.249,173636.330,45.590
+dens_x0.5,19000,2715,1,28.115,213940.398,56.172
+dens_x0.5,19000,2715,2,29.197,213940.398,56.172
+dens_x0.5,19000,2715,3,29.275,213940.398,56.172
+dens_x0.5,22000,3611,1,33.547,256079.209,67.236
+dens_x0.5,22000,3611,2,36.908,256079.209,67.236
+dens_x0.5,22000,3611,3,35.231,256079.209,67.236
+dens_x0.5,26000,5194,1,38.924,317113.582,83.261
+dens_x0.5,26000,5194,2,42.115,317113.582,83.261
+dens_x0.5,26000,5194,3,39.713,317113.582,83.261
+dens_x0.5,30000,6763,1,67.958,378988.027,99.507
+dens_x0.5,30000,6763,2,42.524,378988.027,99.507
+dens_x0.5,30000,6763,3,43.359,378988.027,99.507
+dens_x0.5,36000,9660,1,46.321,479031.581,125.774
+dens_x0.5,36000,9660,2,45.298,479031.581,125.774
+dens_x0.5,36000,9660,3,42.455,479031.581,125.774
+dens_x0.5,42000,13346,1,56.408,589181.686,154.695
+dens_x0.5,42000,13346,2,55.323,589181.686,154.695
+dens_x0.5,42000,13346,3,59.205,589181.686,154.695
+dens_x0.5,49000,17900,1,60.583,722491.606,189.697
+dens_x0.5,49000,17900,2,59.983,722491.606,189.697
+dens_x0.5,49000,17900,3,59.318,722491.606,189.697
+dens_x0.5,57000,24363,1,73.324,890990.473,233.938
+dens_x0.5,57000,24363,2,73.899,890990.473,233.938
+dens_x0.5,57000,24363,3,74.315,890990.473,233.938
+dens_x0.5,67000,33829,1,138.137,1120457.009,294.187
+dens_x0.5,67000,33829,2,137.270,1120457.009,294.187
+dens_x0.5,67000,33829,3,137.367,1120457.009,294.187
+dens_x0.5,79000,46917,1,184.130,1419991.587,372.832
+dens_x0.5,79000,46917,2,194.899,1419991.587,372.832
+dens_x0.5,79000,46917,3,221.479,1419991.587,372.832
+dens_x0.5,92000,63864,1,233.290,1781454.424,467.738
+dens_x0.5,92000,63864,2,284.284,1781454.424,467.738
+dens_x0.5,92000,63864,3,311.293,1781454.424,467.738
+dens_x0.5,108000,87350,1,263.641,2264084.329,594.457
+dens_x0.5,108000,87350,2,266.367,2264084.329,594.457
+dens_x0.5,108000,87350,3,280.940,2264084.329,594.457
+dens_x0.5,127000,120912,1,271.834,2913447.536,764.954
+dens_x0.5,127000,120912,2,284.661,2913447.536,764.954
+dens_x0.5,127000,120912,3,257.204,2913447.536,764.954
+dens_x0.5,149000,166365,1,335.326,3756533.770,986.314
+dens_x0.5,149000,166365,2,320.044,3756533.770,986.314
+dens_x0.5,149000,166365,3,302.948,3756533.770,986.314
+dens_x0.5,174000,226551,1,398.037,4833373.044,1269.049
+dens_x0.5,174000,226551,2,381.909,4833373.044,1269.049
+dens_x0.5,174000,226551,3,395.745,4833373.044,1269.049
+dens_x0.5,204000,312325,1,533.436,6312525.050,1657.414
+dens_x0.5,204000,312325,2,518.033,6312525.050,1657.414
+dens_x0.5,204000,312325,3,533.411,6312525.050,1657.414
+dens_x0.5,240000,431142,1,654.672,8314371.662,2183.018
+dens_x0.5,240000,431142,2,636.705,8314371.662,2183.018
+dens_x0.5,240000,431142,3,646.470,8314371.662,2183.018
+dens_x0.5,281000,589828,1,928.709,10925503.834,2868.596
+dens_x0.5,281000,589828,2,938.183,10925503.834,2868.596
+dens_x0.5,281000,589828,3,893.858,10925503.834,2868.596
+dens_x0.5,329000,811844,1,1234.116,14493068.872,3805.295
+dens_x0.5,329000,811844,2,1227.491,14493068.872,3805.295
+dens_x0.5,329000,811844,3,1175.747,14493068.872,3805.295
+dens_x0.5,386000,1115568,1,1811.607,19315559.086,5071.486
+dens_x0.5,386000,1115568,2,1734.140,19315559.086,5071.486
+dens_x0.5,386000,1115568,3,1746.189,19315559.086,5071.486
+dens_x0.5,452000,1531464,1,2372.135,25827552.428,6781.272
+dens_x0.5,452000,1531464,2,2490.243,25827552.428,6781.272
+dens_x0.5,452000,1531464,3,2315.387,25827552.428,6781.272
+dens_x0.5,530000,2103874,1,3490.322,34716124.680,9115.052
+dens_x0.5,530000,2103874,2,4043.773,34716124.680,9115.052
+dens_x0.5,530000,2103874,3,3668.140,34716124.680,9115.052
+dens_x0.5,621000,2887681,1,5439.786,46802598.872,12288.472
+dens_x0.5,621000,2887681,2,5475.330,46802598.872,12288.472
+dens_x0.5,621000,2887681,3,5776.755,46802598.872,12288.472
+dens_x0.5,728000,3971764,1,8193.403,63437679.196,16656.172
+dens_x0.5,728000,3971764,2,8824.932,63437679.196,16656.172
+dens_x0.5,728000,3971764,3,8664.769,63437679.196,16656.172
+dens_x0.5,853000,5448500,1,15321.894,86056528.179,22594.969
+dens_x0.5,853000,5448500,2,14151.323,86056528.179,22594.969
+dens_x0.5,853000,5448500,3,11713.548,86056528.179,22594.969
+dens_x0.5,1000000,7489974,1,19408.272,117293325.434,30796.490
+dens_x0.5,1000000,7489974,2,19446.200,117293325.434,30796.490
+dens_x0.5,1000000,7489974,3,23624.300,117293325.434,30796.490
+dens_x1.0,10000,1550,1,32.667,106379.431,27.931
+dens_x1.0,10000,1550,2,41.566,106379.431,27.931
+dens_x1.0,10000,1550,3,42.463,106379.431,27.931
+dens_x1.0,12000,2135,1,61.629,132765.276,34.859
+dens_x1.0,12000,2135,2,58.200,132765.276,34.859
+dens_x1.0,12000,2135,3,42.940,132765.276,34.859
+dens_x1.0,14000,2963,1,57.881,161942.582,42.520
+dens_x1.0,14000,2963,2,62.641,161942.582,42.520
+dens_x1.0,14000,2963,3,59.948,161942.582,42.520
+dens_x1.0,16000,3912,1,74.806,192755.010,50.610
+dens_x1.0,16000,3912,2,70.515,192755.010,50.610
+dens_x1.0,16000,3912,3,66.497,192755.010,50.610
+dens_x1.0,19000,5413,1,83.372,240521.618,63.151
+dens_x1.0,19000,5413,2,59.112,240521.618,63.151
+dens_x1.0,19000,5413,3,70.868,240521.618,63.151
+dens_x1.0,22000,7290,1,33.001,292864.786,76.894
+dens_x1.0,22000,7290,2,65.463,292864.786,76.894
+dens_x1.0,22000,7290,3,34.928,292864.786,76.894
+dens_x1.0,26000,10078,1,57.807,366763.602,96.297
+dens_x1.0,26000,10078,2,59.212,366763.602,96.297
+dens_x1.0,26000,10078,3,62.940,366763.602,96.297
+dens_x1.0,30000,13336,1,50.504,446748.773,117.298
+dens_x1.0,30000,13336,2,83.799,446748.773,117.298
+dens_x1.0,30000,13336,3,55.835,446748.773,117.298
+dens_x1.0,36000,19373,1,138.905,580933.327,152.530
+dens_x1.0,36000,19373,2,92.379,580933.327,152.530
+dens_x1.0,36000,19373,3,145.058,580933.327,152.530
+dens_x1.0,42000,26057,1,147.388,724495.682,190.223
+dens_x1.0,42000,26057,2,207.063,724495.682,190.223
+dens_x1.0,42000,26057,3,174.600,724495.682,190.223
+dens_x1.0,49000,35903,1,227.166,916916.365,240.745
+dens_x1.0,49000,35903,2,205.647,916916.365,240.745
+dens_x1.0,49000,35903,3,195.526,916916.365,240.745
+dens_x1.0,57000,48648,1,251.880,1156930.810,303.763
+dens_x1.0,57000,48648,2,356.017,1156930.810,303.763
+dens_x1.0,57000,48648,3,266.212,1156930.810,303.763
+dens_x1.0,67000,67188,1,390.684,1491157.159,391.518
+dens_x1.0,67000,67188,2,228.733,1491157.159,391.518
+dens_x1.0,67000,67188,3,207.511,1491157.159,391.518
+dens_x1.0,79000,93552,1,544.276,1945903.955,510.916
+dens_x1.0,79000,93552,2,299.406,1945903.955,510.916
+dens_x1.0,79000,93552,3,566.521,1945903.955,510.916
+dens_x1.0,92000,126776,1,248.202,2500509.887,656.533
+dens_x1.0,92000,126776,2,375.889,2500509.887,656.533
+dens_x1.0,92000,126776,3,338.922,2500509.887,656.533
+dens_x1.0,108000,175020,1,364.955,3280169.679,861.240
+dens_x1.0,108000,175020,2,309.446,3280169.679,861.240
+dens_x1.0,108000,175020,3,323.999,3280169.679,861.240
+dens_x1.0,127000,241838,1,704.959,4334562.918,1138.081
+dens_x1.0,127000,241838,2,494.727,4334562.918,1138.081
+dens_x1.0,127000,241838,3,464.965,4334562.918,1138.081
+dens_x1.0,149000,333094,1,550.220,5742559.864,1507.764
+dens_x1.0,149000,333094,2,519.299,5742559.864,1507.764
+dens_x1.0,149000,333094,3,539.635,5742559.864,1507.764
+dens_x1.0,174000,454003,1,786.343,7577993.244,1989.675
+dens_x1.0,174000,454003,2,704.575,7577993.244,1989.675
+dens_x1.0,174000,454003,3,724.878,7577993.244,1989.675
+dens_x1.0,204000,623499,1,965.397,10116899.562,2656.289
+dens_x1.0,204000,623499,2,910.950,10116899.562,2656.289
+dens_x1.0,204000,623499,3,934.879,10116899.562,2656.289
+dens_x1.0,240000,862431,1,1269.612,13657349.809,3585.868
+dens_x1.0,240000,862431,2,1262.191,13657349.809,3585.868
+dens_x1.0,240000,862431,3,1509.854,13657349.809,3585.868
+dens_x1.0,281000,1182168,1,1911.024,18357066.601,4819.824
+dens_x1.0,281000,1182168,2,1828.471,18357066.601,4819.824
+dens_x1.0,281000,1182168,3,1921.693,18357066.601,4819.824
+dens_x1.0,329000,1623777,1,2749.311,24807713.897,6513.504
+dens_x1.0,329000,1623777,2,2523.897,24807713.897,6513.504
+dens_x1.0,329000,1623777,3,3497.880,24807713.897,6513.504
+dens_x1.0,386000,2232671,1,3995.353,33685517.024,8844.456
+dens_x1.0,386000,2232671,2,4067.755,33685517.024,8844.456
+dens_x1.0,386000,2232671,3,4026.271,33685517.024,8844.456
+dens_x1.0,452000,3055874,1,5745.135,45677561.904,11993.083
+dens_x1.0,452000,3055874,2,6786.590,45677561.904,11993.083
+dens_x1.0,452000,3055874,3,5896.035,45677561.904,11993.083
+dens_x1.0,530000,4208065,1,10912.807,62450692.510,16397.030
+dens_x1.0,530000,4208065,2,9795.368,62450692.510,16397.030
+dens_x1.0,530000,4208065,3,9179.829,62450692.510,16397.030
+dens_x1.0,621000,5774507,1,12696.188,85310220.195,22399.018
+dens_x1.0,621000,5774507,2,15804.396,85310220.195,22399.018
+dens_x1.0,621000,5774507,3,12631.341,85310220.195,22399.018
+dens_x1.0,728000,7935888,1,20186.080,116945648.236,30705.204
+dens_x1.0,728000,7935888,2,19884.957,116945648.236,30705.204
+dens_x1.0,728000,7935888,3,20055.111,116945648.236,30705.204
+dens_x1.0,853000,10900416,1,31703.026,160510699.866,42143.627
+dens_x1.0,853000,10900416,2,31489.327,160510699.866,42143.627
+dens_x1.0,853000,10900416,3,33382.287,160510699.866,42143.627
+dens_x1.0,1000000,14986435,1,53515.256,220860761.527,57989.116
+dens_x1.0,1000000,14986435,2,43407.301,220860761.527,57989.116
+dens_x1.0,1000000,14986435,3,48171.900,220860761.527,57989.116
+dens_x2.0,10000,3007,1,37.511,119798.897,31.454
+dens_x2.0,10000,3007,2,21.882,119798.897,31.454
+dens_x2.0,10000,3007,3,25.285,119798.897,31.454
+dens_x2.0,12000,4365,1,37.716,153710.912,40.358
+dens_x2.0,12000,4365,2,32.105,153710.912,40.358
+dens_x2.0,12000,4365,3,45.543,153710.912,40.358
+dens_x2.0,14000,5830,1,47.403,189313.294,49.706
+dens_x2.0,14000,5830,2,51.361,189313.294,49.706
+dens_x2.0,14000,5830,3,28.873,189313.294,49.706
+dens_x2.0,16000,7624,1,55.510,228688.447,60.044
+dens_x2.0,16000,7624,2,25.997,228688.447,60.044
+dens_x2.0,16000,7624,3,29.947,228688.447,60.044
+dens_x2.0,19000,10837,1,51.411,293959.920,77.182
+dens_x2.0,19000,10837,2,33.435,293959.920,77.182
+dens_x2.0,19000,10837,3,43.471,293959.920,77.182
+dens_x2.0,22000,14556,1,51.011,365516.050,95.970
+dens_x2.0,22000,14556,2,59.701,365516.050,95.970
+dens_x2.0,22000,14556,3,63.187,365516.050,95.970
+dens_x2.0,26000,20395,1,522.381,471644.695,123.835
+dens_x2.0,26000,20395,2,73.043,471644.695,123.835
+dens_x2.0,26000,20395,3,57.832,471644.695,123.835
+dens_x2.0,30000,26790,1,173.512,585445.422,153.714
+dens_x2.0,30000,26790,2,137.431,585445.422,153.714
+dens_x2.0,30000,26790,3,137.061,585445.422,153.714
+dens_x2.0,36000,38800,1,148.907,784747.311,206.043
+dens_x2.0,36000,38800,2,268.881,784747.311,206.043
+dens_x2.0,36000,38800,3,198.411,784747.311,206.043
+dens_x2.0,42000,53391,1,217.649,1015477.726,266.623
+dens_x2.0,42000,53391,2,351.625,1015477.726,266.623
+dens_x2.0,42000,53391,3,195.803,1015477.726,266.623
+dens_x2.0,49000,71945,1,267.586,1306154.668,342.943
+dens_x2.0,49000,71945,2,252.261,1306154.668,342.943
+dens_x2.0,49000,71945,3,260.151,1306154.668,342.943
+dens_x2.0,57000,97182,1,164.837,1688417.255,443.310
+dens_x2.0,57000,97182,2,164.714,1688417.255,443.310
+dens_x2.0,57000,97182,3,172.108,1688417.255,443.310
+dens_x2.0,67000,134082,1,216.360,2234513.248,586.693
+dens_x2.0,67000,134082,2,223.478,2234513.248,586.693
+dens_x2.0,67000,134082,3,209.673,2234513.248,586.693
+dens_x2.0,79000,186876,1,384.747,2998337.660,787.242
+dens_x2.0,79000,186876,2,314.466,2998337.660,787.242
+dens_x2.0,79000,186876,3,348.417,2998337.660,787.242
+dens_x2.0,92000,253371,1,389.063,3947432.991,1036.436
+dens_x2.0,92000,253371,2,376.706,3947432.991,1036.436
+dens_x2.0,92000,253371,3,409.028,3947432.991,1036.436
+dens_x2.0,108000,349108,1,569.904,5297829.841,1390.996
+dens_x2.0,108000,349108,2,509.953,5297829.841,1390.996
+dens_x2.0,108000,349108,3,614.219,5297829.841,1390.996
+dens_x2.0,127000,484181,1,661.230,7182563.887,1885.851
+dens_x2.0,127000,484181,2,736.545,7182563.887,1885.851
+dens_x2.0,127000,484181,3,690.100,7182563.887,1885.851
+dens_x2.0,149000,666491,1,889.956,9713885.437,2550.474
+dens_x2.0,149000,666491,2,912.750,9713885.437,2550.474
+dens_x2.0,149000,666491,3,944.557,9713885.437,2550.474
+dens_x2.0,174000,907603,1,1342.386,13051498.522,3426.796
+dens_x2.0,174000,907603,2,1332.430,13051498.522,3426.796
+dens_x2.0,174000,907603,3,1345.460,13051498.522,3426.796
+dens_x2.0,204000,1247723,1,1799.321,17748584.329,4660.061
+dens_x2.0,204000,1247723,2,1801.665,17748584.329,4660.061
+dens_x2.0,204000,1247723,3,1804.948,17748584.329,4660.061
+dens_x2.0,240000,1725541,1,2803.778,24349896.729,6393.299
+dens_x2.0,240000,1725541,2,2813.138,24349896.729,6393.299
+dens_x2.0,240000,1725541,3,2724.773,24349896.729,6393.299
+dens_x2.0,281000,2365586,1,4194.937,33204358.944,8718.124
+dens_x2.0,281000,2365586,2,4184.496,33204358.944,8718.124
+dens_x2.0,281000,2365586,3,4189.026,33204358.944,8718.124
+dens_x2.0,329000,3242756,1,6341.712,45374920.412,11913.622
+dens_x2.0,329000,3242756,2,6127.021,45374920.412,11913.622
+dens_x2.0,329000,3242756,3,6071.201,45374920.412,11913.622
+dens_x2.0,386000,4461354,1,9538.741,62354387.279,16371.744
+dens_x2.0,386000,4461354,2,9270.728,62354387.279,16371.744
+dens_x2.0,386000,4461354,3,9227.539,62354387.279,16371.744
+dens_x2.0,452000,6119886,1,14518.451,85575402.535,22468.645
+dens_x2.0,452000,6119886,2,17278.281,85575402.535,22468.645
+dens_x2.0,452000,6119886,3,16358.872,85575402.535,22468.645
+dens_x2.0,530000,8413791,1,25026.906,117884820.410,30951.793
+dens_x2.0,530000,8413791,2,26927.986,117884820.410,30951.793
+dens_x2.0,530000,8413791,3,28834.072,117884820.410,30951.793
+dens_x2.0,621000,11552079,1,35260.396,162377752.059,42633.840
+dens_x2.0,621000,11552079,2,35232.220,162377752.059,42633.840
+dens_x2.0,621000,11552079,3,35913.410,162377752.059,42633.840
+dens_x2.0,728000,15870417,1,59449.804,224046367.608,58825.527
+dens_x2.0,728000,15870417,2,55741.168,224046367.608,58825.527
+dens_x2.0,728000,15870417,3,57488.308,224046367.608,58825.527
+dens_x2.0,853000,21800415,1,86590.882,309366697.818,81227.200
+dens_x2.0,853000,21800415,2,86308.070,309366697.818,81227.200
+dens_x2.0,853000,21800415,3,84987.572,309366697.818,81227.200
+dens_x2.0,1000000,29958277,1,135554.293,427704402.750,112297.902
+dens_x2.0,1000000,29958277,2,134537.849,427704402.750,112297.902
+dens_x2.0,1000000,29958277,3,151123.123,427704402.750,112297.902

results_qing.csv

@@ -0,0 +1,271 @@
+series,n,m,trial,ms,theo_x,normalized_theory_ms
+dens_x0.5,10000,761,1,17.752,99112.473,50.448
+dens_x0.5,10000,761,2,17.676,99112.473,50.448
+dens_x0.5,10000,761,3,17.359,99112.473,50.448
+dens_x0.5,12000,1077,1,19.911,122827.840,62.519
+dens_x0.5,12000,1077,2,18.349,122827.840,62.519
+dens_x0.5,12000,1077,3,17.963,122827.840,62.519
+dens_x0.5,14000,1455,1,21.243,147545.989,75.101
+dens_x0.5,14000,1455,2,21.438,147545.989,75.101
+dens_x0.5,14000,1455,3,21.547,147545.989,75.101
+dens_x0.5,16000,1937,1,33.463,173636.330,88.381
+dens_x0.5,16000,1937,2,25.635,173636.330,88.381
+dens_x0.5,16000,1937,3,24.989,173636.330,88.381
+dens_x0.5,19000,2715,1,29.835,213940.398,108.895
+dens_x0.5,19000,2715,2,29.484,213940.398,108.895
+dens_x0.5,19000,2715,3,29.684,213940.398,108.895
+dens_x0.5,22000,3611,1,30.203,256079.209,130.344
+dens_x0.5,22000,3611,2,30.522,256079.209,130.344
+dens_x0.5,22000,3611,3,33.004,256079.209,130.344
+dens_x0.5,26000,5194,1,41.147,317113.582,161.410
+dens_x0.5,26000,5194,2,40.076,317113.582,161.410
+dens_x0.5,26000,5194,3,36.815,317113.582,161.410
+dens_x0.5,30000,6763,1,44.788,378988.027,192.904
+dens_x0.5,30000,6763,2,44.445,378988.027,192.904
+dens_x0.5,30000,6763,3,42.679,378988.027,192.904
+dens_x0.5,36000,9660,1,45.425,479031.581,243.826
+dens_x0.5,36000,9660,2,46.989,479031.581,243.826
+dens_x0.5,36000,9660,3,44.779,479031.581,243.826
+dens_x0.5,42000,13346,1,50.098,589181.686,299.892
+dens_x0.5,42000,13346,2,51.428,589181.686,299.892
+dens_x0.5,42000,13346,3,53.184,589181.686,299.892
+dens_x0.5,49000,17900,1,64.897,722491.606,367.747
+dens_x0.5,49000,17900,2,57.590,722491.606,367.747
+dens_x0.5,49000,17900,3,62.995,722491.606,367.747
+dens_x0.5,57000,24363,1,110.618,890990.473,453.512
+dens_x0.5,57000,24363,2,109.511,890990.473,453.512
+dens_x0.5,57000,24363,3,107.511,890990.473,453.512
+dens_x0.5,67000,33829,1,157.316,1120457.009,570.310
+dens_x0.5,67000,33829,2,152.683,1120457.009,570.310
+dens_x0.5,67000,33829,3,149.917,1120457.009,570.310
+dens_x0.5,79000,46917,1,217.186,1419991.587,722.773
+dens_x0.5,79000,46917,2,218.315,1419991.587,722.773
+dens_x0.5,79000,46917,3,214.770,1419991.587,722.773
+dens_x0.5,92000,63864,1,254.246,1781454.424,906.757
+dens_x0.5,92000,63864,2,223.552,1781454.424,906.757
+dens_x0.5,92000,63864,3,225.867,1781454.424,906.757
+dens_x0.5,108000,87350,1,301.682,2264084.329,1152.414
+dens_x0.5,108000,87350,2,218.064,2264084.329,1152.414
+dens_x0.5,108000,87350,3,215.160,2264084.329,1152.414
+dens_x0.5,127000,120912,1,315.482,2913447.536,1482.939
+dens_x0.5,127000,120912,2,293.023,2913447.536,1482.939
+dens_x0.5,127000,120912,3,379.251,2913447.536,1482.939
+dens_x0.5,149000,166365,1,338.435,3756533.770,1912.068
+dens_x0.5,149000,166365,2,346.454,3756533.770,1912.068
+dens_x0.5,149000,166365,3,346.208,3756533.770,1912.068
+dens_x0.5,174000,226551,1,455.336,4833373.044,2460.177
+dens_x0.5,174000,226551,2,504.231,4833373.044,2460.177
+dens_x0.5,174000,226551,3,522.258,4833373.044,2460.177
+dens_x0.5,204000,312325,1,696.115,6312525.050,3213.062
+dens_x0.5,204000,312325,2,850.734,6312525.050,3213.062
+dens_x0.5,204000,312325,3,922.932,6312525.050,3213.062
+dens_x0.5,240000,431142,1,1927.235,8314371.662,4231.998
+dens_x0.5,240000,431142,2,1138.207,8314371.662,4231.998
+dens_x0.5,240000,431142,3,1115.626,8314371.662,4231.998
+dens_x0.5,281000,589828,1,1827.282,10925503.834,5561.059
+dens_x0.5,281000,589828,2,1339.802,10925503.834,5561.059
+dens_x0.5,281000,589828,3,1169.025,10925503.834,5561.059
+dens_x0.5,329000,811844,1,1522.195,14493068.872,7376.943
+dens_x0.5,329000,811844,2,1501.547,14493068.872,7376.943
+dens_x0.5,329000,811844,3,1607.400,14493068.872,7376.943
+dens_x0.5,386000,1115568,1,2344.641,19315559.086,9831.580
+dens_x0.5,386000,1115568,2,2432.236,19315559.086,9831.580
+dens_x0.5,386000,1115568,3,2322.046,19315559.086,9831.580
+dens_x0.5,452000,1531464,1,3316.101,25827552.428,13146.172
+dens_x0.5,452000,1531464,2,3377.502,25827552.428,13146.172
+dens_x0.5,452000,1531464,3,3354.746,25827552.428,13146.172
+dens_x0.5,530000,2103874,1,5013.327,34716124.680,17670.437
+dens_x0.5,530000,2103874,2,4984.979,34716124.680,17670.437
+dens_x0.5,530000,2103874,3,5053.679,34716124.680,17670.437
+dens_x0.5,621000,2887681,1,7653.489,46802598.872,23822.428
+dens_x0.5,621000,2887681,2,7610.639,46802598.872,23822.428
+dens_x0.5,621000,2887681,3,7519.181,46802598.872,23822.428
+dens_x0.5,728000,3971764,1,12015.660,63437679.196,32289.650
+dens_x0.5,728000,3971764,2,11577.906,63437679.196,32289.650
+dens_x0.5,728000,3971764,3,11991.380,63437679.196,32289.650
+dens_x0.5,853000,5448500,1,17834.890,86056528.179,43802.598
+dens_x0.5,853000,5448500,2,17710.405,86056528.179,43802.598
+dens_x0.5,853000,5448500,3,18224.804,86056528.179,43802.598
+dens_x0.5,1000000,7489974,1,26985.940,117293325.434,59702.065
+dens_x0.5,1000000,7489974,2,28216.802,117293325.434,59702.065
+dens_x0.5,1000000,7489974,3,27674.139,117293325.434,59702.065
+dens_x1.0,10000,1550,1,17.827,106379.431,54.147
+dens_x1.0,10000,1550,2,16.952,106379.431,54.147
+dens_x1.0,10000,1550,3,17.015,106379.431,54.147
+dens_x1.0,12000,2135,1,20.646,132765.276,67.577
+dens_x1.0,12000,2135,2,21.123,132765.276,67.577
+dens_x1.0,12000,2135,3,20.674,132765.276,67.577
+dens_x1.0,14000,2963,1,24.111,161942.582,82.428
+dens_x1.0,14000,2963,2,25.719,161942.582,82.428
+dens_x1.0,14000,2963,3,25.629,161942.582,82.428
+dens_x1.0,16000,3912,1,24.766,192755.010,98.112
+dens_x1.0,16000,3912,2,28.343,192755.010,98.112
+dens_x1.0,16000,3912,3,26.101,192755.010,98.112
+dens_x1.0,19000,5413,1,27.698,240521.618,122.425
+dens_x1.0,19000,5413,2,30.179,240521.618,122.425
+dens_x1.0,19000,5413,3,29.468,240521.618,122.425
+dens_x1.0,22000,7290,1,38.453,292864.786,149.068
+dens_x1.0,22000,7290,2,35.754,292864.786,149.068
+dens_x1.0,22000,7290,3,34.257,292864.786,149.068
+dens_x1.0,26000,10078,1,44.338,366763.602,186.682
+dens_x1.0,26000,10078,2,45.863,366763.602,186.682
+dens_x1.0,26000,10078,3,43.600,366763.602,186.682
+dens_x1.0,30000,13336,1,45.214,446748.773,227.394
+dens_x1.0,30000,13336,2,46.397,446748.773,227.394
+dens_x1.0,30000,13336,3,45.272,446748.773,227.394
+dens_x1.0,36000,19373,1,89.256,580933.327,295.694
+dens_x1.0,36000,19373,2,89.571,580933.327,295.694
+dens_x1.0,36000,19373,3,61.459,580933.327,295.694
+dens_x1.0,42000,26057,1,111.164,724495.682,368.767
+dens_x1.0,42000,26057,2,112.365,724495.682,368.767
+dens_x1.0,42000,26057,3,111.181,724495.682,368.767
+dens_x1.0,49000,35903,1,147.373,916916.365,466.709
+dens_x1.0,49000,35903,2,160.224,916916.365,466.709
+dens_x1.0,49000,35903,3,149.107,916916.365,466.709
+dens_x1.0,57000,48648,1,204.262,1156930.810,588.875
+dens_x1.0,57000,48648,2,187.035,1156930.810,588.875
+dens_x1.0,57000,48648,3,213.253,1156930.810,588.875
+dens_x1.0,67000,67188,1,236.565,1491157.159,758.996
+dens_x1.0,67000,67188,2,233.706,1491157.159,758.996
+dens_x1.0,67000,67188,3,234.320,1491157.159,758.996
+dens_x1.0,79000,93552,1,197.042,1945903.955,990.461
+dens_x1.0,79000,93552,2,268.478,1945903.955,990.461
+dens_x1.0,79000,93552,3,308.769,1945903.955,990.461
+dens_x1.0,92000,126776,1,273.573,2500509.887,1272.754
+dens_x1.0,92000,126776,2,250.997,2500509.887,1272.754
+dens_x1.0,92000,126776,3,275.845,2500509.887,1272.754
+dens_x1.0,108000,175020,1,379.493,3280169.679,1669.600
+dens_x1.0,108000,175020,2,309.677,3280169.679,1669.600
+dens_x1.0,108000,175020,3,306.916,3280169.679,1669.600
+dens_x1.0,127000,241838,1,468.439,4334562.918,2206.284
+dens_x1.0,127000,241838,2,446.181,4334562.918,2206.284
+dens_x1.0,127000,241838,3,458.373,4334562.918,2206.284
+dens_x1.0,149000,333094,1,566.457,5742559.864,2922.951
+dens_x1.0,149000,333094,2,579.989,5742559.864,2922.951
+dens_x1.0,149000,333094,3,577.801,5742559.864,2922.951
+dens_x1.0,174000,454003,1,785.385,7577993.244,3857.183
+dens_x1.0,174000,454003,2,768.331,7577993.244,3857.183
+dens_x1.0,174000,454003,3,799.821,7577993.244,3857.183
+dens_x1.0,204000,623499,1,1114.363,10116899.562,5149.481
+dens_x1.0,204000,623499,2,1120.554,10116899.562,5149.481
+dens_x1.0,204000,623499,3,1099.929,10116899.562,5149.481
+dens_x1.0,240000,862431,1,1680.754,13657349.809,6951.563
+dens_x1.0,240000,862431,2,1659.639,13657349.809,6951.563
+dens_x1.0,240000,862431,3,1635.171,13657349.809,6951.563
+dens_x1.0,281000,1182168,1,2464.097,18357066.601,9343.710
+dens_x1.0,281000,1182168,2,2520.973,18357066.601,9343.710
+dens_x1.0,281000,1182168,3,2521.537,18357066.601,9343.710
+dens_x1.0,329000,1623777,1,3657.489,24807713.897,12627.076
+dens_x1.0,329000,1623777,2,3822.215,24807713.897,12627.076
+dens_x1.0,329000,1623777,3,3771.764,24807713.897,12627.076
+dens_x1.0,386000,2232671,1,5777.777,33685517.024,17145.860
+dens_x1.0,386000,2232671,2,5853.952,33685517.024,17145.860
+dens_x1.0,386000,2232671,3,5741.782,33685517.024,17145.860
+dens_x1.0,452000,3055874,1,8791.834,45677561.904,23249.786
+dens_x1.0,452000,3055874,2,8823.808,45677561.904,23249.786
+dens_x1.0,452000,3055874,3,8832.307,45677561.904,23249.786
+dens_x1.0,530000,4208065,1,13664.587,62450692.510,31787.276
+dens_x1.0,530000,4208065,2,13613.387,62450692.510,31787.276
+dens_x1.0,530000,4208065,3,13421.059,62450692.510,31787.276
+dens_x1.0,621000,5774507,1,20937.405,85310220.195,43422.729
+dens_x1.0,621000,5774507,2,21621.346,85310220.195,43422.729
+dens_x1.0,621000,5774507,3,21139.229,85310220.195,43422.729
+dens_x1.0,728000,7935888,1,34245.137,116945648.236,59525.098
+dens_x1.0,728000,7935888,2,33147.218,116945648.236,59525.098
+dens_x1.0,728000,7935888,3,34011.001,116945648.236,59525.098
+dens_x1.0,853000,10900416,1,54191.487,160510699.866,81699.621
+dens_x1.0,853000,10900416,2,54442.605,160510699.866,81699.621
+dens_x1.0,853000,10900416,3,55017.857,160510699.866,81699.621
+dens_x1.0,1000000,14986435,1,86648.996,220860761.527,112417.680
+dens_x1.0,1000000,14986435,2,88775.035,220860761.527,112417.680
+dens_x1.0,1000000,14986435,3,87850.833,220860761.527,112417.680
+dens_x2.0,10000,3007,1,21.643,119798.897,60.977
+dens_x2.0,10000,3007,2,21.430,119798.897,60.977
+dens_x2.0,10000,3007,3,21.235,119798.897,60.977
+dens_x2.0,12000,4365,1,25.837,153710.912,78.239
+dens_x2.0,12000,4365,2,23.999,153710.912,78.239
+dens_x2.0,12000,4365,3,23.257,153710.912,78.239
+dens_x2.0,14000,5830,1,26.238,189313.294,96.360
+dens_x2.0,14000,5830,2,25.462,189313.294,96.360
+dens_x2.0,14000,5830,3,25.834,189313.294,96.360
+dens_x2.0,16000,7624,1,30.745,228688.447,116.402
+dens_x2.0,16000,7624,2,34.669,228688.447,116.402
+dens_x2.0,16000,7624,3,29.403,228688.447,116.402
+dens_x2.0,19000,10837,1,36.609,293959.920,149.625
+dens_x2.0,19000,10837,2,35.257,293959.920,149.625
+dens_x2.0,19000,10837,3,33.859,293959.920,149.625
+dens_x2.0,22000,14556,1,48.189,365516.050,186.047
+dens_x2.0,22000,14556,2,47.185,365516.050,186.047
+dens_x2.0,22000,14556,3,44.113,365516.050,186.047
+dens_x2.0,26000,20395,1,85.873,471644.695,240.066
+dens_x2.0,26000,20395,2,87.164,471644.695,240.066
+dens_x2.0,26000,20395,3,85.546,471644.695,240.066
+dens_x2.0,30000,26790,1,104.921,585445.422,297.991
+dens_x2.0,30000,26790,2,108.649,585445.422,297.991
+dens_x2.0,30000,26790,3,106.060,585445.422,297.991
+dens_x2.0,36000,38800,1,154.243,784747.311,399.435
+dens_x2.0,36000,38800,2,151.319,784747.311,399.435
+dens_x2.0,36000,38800,3,158.272,784747.311,399.435
+dens_x2.0,42000,53391,1,204.388,1015477.726,516.876
+dens_x2.0,42000,53391,2,198.844,1015477.726,516.876
+dens_x2.0,42000,53391,3,197.630,1015477.726,516.876
+dens_x2.0,49000,71945,1,250.340,1306154.668,664.830
+dens_x2.0,49000,71945,2,279.343,1306154.668,664.830
+dens_x2.0,49000,71945,3,280.425,1306154.668,664.830
+dens_x2.0,57000,97182,1,296.197,1688417.255,859.401
+dens_x2.0,57000,97182,2,182.517,1688417.255,859.401
+dens_x2.0,57000,97182,3,181.647,1688417.255,859.401
+dens_x2.0,67000,134082,1,334.311,2234513.248,1137.363
+dens_x2.0,67000,134082,2,268.165,2234513.248,1137.363
+dens_x2.0,67000,134082,3,282.074,2234513.248,1137.363
+dens_x2.0,79000,186876,1,326.573,2998337.660,1526.148
+dens_x2.0,79000,186876,2,326.265,2998337.660,1526.148
+dens_x2.0,79000,186876,3,343.059,2998337.660,1526.148
+dens_x2.0,92000,253371,1,468.680,3947432.991,2009.235
+dens_x2.0,92000,253371,2,456.861,3947432.991,2009.235
+dens_x2.0,92000,253371,3,474.351,3947432.991,2009.235
+dens_x2.0,108000,349108,1,660.224,5297829.841,2696.585
+dens_x2.0,108000,349108,2,637.197,5297829.841,2696.585
+dens_x2.0,108000,349108,3,651.868,5297829.841,2696.585
+dens_x2.0,127000,484181,1,877.044,7182563.887,3655.910
+dens_x2.0,127000,484181,2,953.163,7182563.887,3655.910
+dens_x2.0,127000,484181,3,986.232,7182563.887,3655.910
+dens_x2.0,149000,666491,1,1297.659,9713885.437,4944.348
+dens_x2.0,149000,666491,2,1257.054,9713885.437,4944.348
+dens_x2.0,149000,666491,3,1284.346,9713885.437,4944.348
+dens_x2.0,174000,907603,1,1867.315,13051498.522,6643.186
+dens_x2.0,174000,907603,2,1903.876,13051498.522,6643.186
+dens_x2.0,174000,907603,3,1907.572,13051498.522,6643.186
+dens_x2.0,204000,1247723,1,3091.032,17748584.329,9033.993
+dens_x2.0,204000,1247723,2,2802.598,17748584.329,9033.993
+dens_x2.0,204000,1247723,3,2826.897,17748584.329,9033.993
+dens_x2.0,240000,1725541,1,4427.498,24349896.729,12394.048
+dens_x2.0,240000,1725541,2,4409.478,24349896.729,12394.048
+dens_x2.0,240000,1725541,3,4625.596,24349896.729,12394.048
+dens_x2.0,281000,2365586,1,6853.596,33204358.944,16900.951
+dens_x2.0,281000,2365586,2,6972.828,33204358.944,16900.951
+dens_x2.0,281000,2365586,3,6824.291,33204358.944,16900.951
+dens_x2.0,329000,3242756,1,10753.997,45374920.412,23095.743
+dens_x2.0,329000,3242756,2,10933.854,45374920.412,23095.743
+dens_x2.0,329000,3242756,3,10953.280,45374920.412,23095.743
+dens_x2.0,386000,4461354,1,16880.353,62354387.279,31738.257
+dens_x2.0,386000,4461354,2,17044.714,62354387.279,31738.257
+dens_x2.0,386000,4461354,3,17421.910,62354387.279,31738.257
+dens_x2.0,452000,6119886,1,27124.239,85575402.535,43557.706
+dens_x2.0,452000,6119886,2,27560.953,85575402.535,43557.706
+dens_x2.0,452000,6119886,3,27165.046,85575402.535,43557.706
+dens_x2.0,530000,8413791,1,43964.374,117884820.410,60003.135
+dens_x2.0,530000,8413791,2,43212.872,117884820.410,60003.135
+dens_x2.0,530000,8413791,3,44285.243,117884820.410,60003.135
+dens_x2.0,621000,11552079,1,73741.057,162377752.059,82649.947
+dens_x2.0,621000,11552079,2,70827.593,162377752.059,82649.947
+dens_x2.0,621000,11552079,3,66956.073,162377752.059,82649.947
+dens_x2.0,728000,15870417,1,107879.018,224046367.608,114039.147
+dens_x2.0,728000,15870417,2,108503.864,224046367.608,114039.147
+dens_x2.0,728000,15870417,3,107531.729,224046367.608,114039.147
+dens_x2.0,853000,21800415,1,174361.084,309366697.818,157467.022
+dens_x2.0,853000,21800415,2,173178.964,309366697.818,157467.022
+dens_x2.0,853000,21800415,3,186579.072,309366697.818,157467.022
+dens_x2.0,1000000,29958277,1,291293.554,427704402.750,217700.674
+dens_x2.0,1000000,29958277,2,278506.552,427704402.750,217700.674
+dens_x2.0,1000000,29958277,3,286580.368,427704402.750,217700.674

seeds_diam_1e-2.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3cdc07c7a3e60521a69a41eea8d44b6a40583dc9ce0d777011e549ef201f05ae
+size 13140898

seeds_diam_1e-3.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50b8aab046e0bf1e7ccb275424ff97a64eb0a1473c5f569e01b7121d13a84359
+size 13142060

seeds_diam_1e-4.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddcc4f54f5261f92098e476e78a3fa541ffc6c63243b0aa24ab9cebf65e31812
+size 13140779

seeds_diam_1e-6_coarsen.json

@@ -0,0 +1 @@
+{"clusters": [{"cluster_id": 0, "members": [0, 633, 926, 1310, 1862, 1866, 2582, 2692]}, {"cluster_id": 1, "members": [1, 203, 247, 470, 652, 854, 857, 1097, 1246, 2583]}, {"cluster_id": 2, "members": [2, 49, 1454, 1662, 1666, 2381]}, {"cluster_id": 3, "members": [3, 594, 749, 1333, 2195, 2312, 2544]}, {"cluster_id": 4, "members": [4, 982, 1016, 1256, 1563, 2176, 2633]}, {"cluster_id": 5, "members": [5, 689, 796, 1148, 1173, 1250, 1292, 1317, 1629, 1659, 2220, 2221, 2507, 2525]}, {"cluster_id": 6, "members": [6, 61, 149, 1008, 1416, 1468, 1922, 1925, 2575]}, {"cluster_id": 7, "members": [7, 173, 208, 290, 480, 510, 649, 681, 687, 987, 1019, 1033, 1171, 1175, 1189, 1298, 1426, 1445, 1446, 1456, 1457, 1498, 1586, 1710, 2359, 2372, 2398, 2489, 2634, 2635, 2636, 2693, 2703]}, {"cluster_id": 8, "members": [8, 14, 258, 435, 751]}, {"cluster_id": 9, "members": [9, 64, 369, 385, 494, 723, 986, 1209, 1644, 2483, 2484, 2614, 2616]}, {"cluster_id": 10, "members": [10, 192, 193, 303, 308, 420, 519, 719, 1066, 1253, 1346, 1574, 1670, 2143, 2545]}, {"cluster_id": 11, "members": [11, 77, 354, 659, 767, 770, 873, 1053, 1121, 1131, 1215, 1405, 1410, 1479, 1508, 1655, 1708, 1803, 1839, 1857, 1894, 2118, 2282, 2313, 2314, 2384, 2453, 2460]}, {"cluster_id": 12, "members": [12, 1001, 1910, 2506, 2661, 2662]}, {"cluster_id": 13, "members": [13, 224, 310, 568, 829, 875, 892, 962, 1004, 1062, 1107, 1116, 1223, 1241, 1331, 1489, 1576, 1581, 1806, 1808, 1809, 1810, 1811, 1814, 1816, 1819, 1822]}, {"cluster_id": 14, "members": [15, 395, 599, 640, 765, 894, 1090, 1093, 1147, 1271, 1598, 1845, 1865, 2367, 2368, 2370, 2371, 2420]}, {"cluster_id": 15, "members": [16, 46, 466, 970, 1305, 1738, 2366, 2444]}, {"cluster_id": 16, "members": [17, 24, 927, 1247, 1315, 1316, 1335, 1356, 1613, 1679, 1856, 2096, 2139, 2140, 2141, 2142]}, {"cluster_id": 17, "members": [18, 317, 943, 1196, 1624, 1785, 1786, 2082, 2300]}, {"cluster_id": 18, "members": [19, 216, 295, 872, 990, 1176, 1177, 1542, 1607, 1931, 1935, 1939]}, {"cluster_id": 19, "members": [20, 128, 1404, 2269, 2270, 2374, 2375, 2416, 2417]}, {"cluster_id": 20, "members": [21, 313, 722, 813, 835, 865, 1043, 1092, 1106, 1134, 1321, 1567, 1654, 1874, 1920, 2296, 2310, 2376, 2422, 2499, 2500, 2501, 2625, 2626, 2653]}, {"cluster_id": 21, "members": [22, 275, 463, 706, 789, 805, 1234, 1371, 1393, 1702, 1703, 1967, 1969, 1970, 2238, 2239]}, {"cluster_id": 22, "members": [23, 92, 108, 898, 1155, 1327, 1328, 1835, 1836, 2157, 2158, 2159, 2160, 2161, 2209]}, {"cluster_id": 23, "members": [25, 1301, 1344, 1447, 1928, 2011, 2315, 2317, 2475]}, {"cluster_id": 24, "members": [26, 99, 122, 123, 127, 177, 1236, 1437, 1639, 1688, 2454, 2455, 2479, 2578, 2604, 2679, 2682]}, {"cluster_id": 25, "members": [27, 288, 531, 606, 618, 938, 2027, 2230, 2360, 2361]}, {"cluster_id": 26, "members": [28, 54, 422, 545, 700, 877, 1010, 1061, 1069, 1111, 1182, 1569, 1578, 1582, 1651, 1683, 1684, 1685, 1687, 1690, 1691, 1726, 1736, 1909, 2513, 2677]}, {"cluster_id": 27, "members": [29, 43, 152, 375, 396, 963, 1094, 1369, 1530, 1653, 1988, 2240, 2399, 2400, 2401, 2434, 2435, 2645]}, {"cluster_id": 28, "members": [30, 405, 697, 738, 1080, 1275, 2059, 2162, 2343, 2344]}, {"cluster_id": 29, "members": [31, 76, 737, 1594, 1974, 1979, 2178]}, {"cluster_id": 30, "members": [32, 518, 666, 2151, 2423]}, {"cluster_id": 31, "members": [33, 196, 286, 332, 442, 588, 627, 698, 911, 1051, 1798, 2040, 2120, 2122, 2362, 2615, 2618, 2619, 2637, 2696]}, {"cluster_id": 32, "members": [34, 59, 101, 111, 228, 322, 326, 333, 346, 357, 557, 569, 628, 684, 686, 757, 764, 839, 849, 882, 919, 929, 951, 977, 999, 1040, 1114, 1149, 1150, 1238, 1243, 1272, 1281, 1339, 1358, 1384, 1389, 1451, 1471, 1492, 1517, 1520, 1546, 1562, 1565, 1604, 1612, 1646, 1694, 1696, 1709, 1711, 1713, 1716, 1717, 1720, 1725, 1730, 1734, 1737, 1744, 1746, 1747, 1752, 1753, 1754, 1757, 1759, 1760, 1764, 1930, 1948, 1982, 2041, 2153, 2334, 2421, 2512, 2560, 2561, 2597, 2627, 2643]}, {"cluster_id": 33, "members": [35, 504, 702, 728, 777, 895, 945, 1187, 1188, 1296, 1377, 1436, 1913, 2069, 2070, 2388, 2640]}, {"cluster_id": 34, "members": [36, 530, 1146, 1505, 1552, 1699, 1788, 2086, 2094, 2106, 2254, 2591]}, {"cluster_id": 35, "members": [37, 60, 331, 355, 1527, 1668, 2274, 2480]}, {"cluster_id": 36, "members": [38, 226, 429, 523, 617, 705, 794, 863, 949, 1160, 1166, 1493, 1807, 2043, 2406]}, {"cluster_id": 37, "members": [39, 1522, 1697, 1965, 2253, 2594, 2680]}, {"cluster_id": 38, "members": [40, 866, 1049, 1364, 2610]}, {"cluster_id": 39, "members": [41, 96, 175, 398, 596, 644, 730, 955, 992, 2135, 2164, 2177, 2217]}, {"cluster_id": 40, "members": [42, 69, 87, 106, 262, 604, 907, 1351, 1362, 1372, 1535, 1916, 2461]}, {"cluster_id": 41, "members": [44, 2247, 2263, 2624, 2701]}, {"cluster_id": 42, "members": [45, 174, 183, 214, 220, 367, 440, 454, 459, 469, 476, 482, 486, 490, 496, 498, 499, 501, 503, 535, 539, 541, 542, 551, 561, 563, 577, 578, 579, 582, 605, 663, 676, 913, 1014, 1056, 1091, 1105, 1122, 1140, 1233, 1705, 1792, 1917, 1952, 2029, 2179, 2322, 2354, 2393, 2433, 2457, 2458, 2548, 2549]}, {"cluster_id": 43, "members": [47, 163, 448, 472, 559, 587, 844, 1032, 1183, 1579, 1825]}, {"cluster_id": 44, "members": [48, 1031, 2205, 2206, 2471]}, {"cluster_id": 45, "members": [50, 241, 590, 648, 750, 1060, 1128, 1355, 1441, 1619, 1700, 1707, 1954, 1957, 2268, 2271, 2272, 2273, 2275, 2445, 2630, 2699]}, {"cluster_id": 46, "members": [51, 421, 457, 708, 710, 768, 1392, 1723, 1724, 2121, 2212, 2214, 2215, 2216]}, {"cluster_id": 47, "members": [52, 538, 636, 858, 1139, 1286, 1467, 1515, 1838, 2072, 2172, 2182, 2233, 2234, 2502]}, {"cluster_id": 48, "members": [53, 1601, 1739, 1750, 1766]}, {"cluster_id": 49, "members": [55, 815, 1079, 2018, 2020]}, {"cluster_id": 50, "members": [56, 412, 447, 581, 1616, 2638]}, {"cluster_id": 51, "members": [57, 731, 901, 1279, 2073, 2186, 2203, 2330, 2345, 2418, 2511]}, {"cluster_id": 52, "members": [58, 246, 744, 936, 1154, 1458, 1714, 1715, 2439, 2514, 2515, 2530, 2641]}, {"cluster_id": 53, "members": [62, 373, 475, 485, 624, 1025, 1042, 1047, 1198, 1926, 2051, 2052, 2219, 2333]}, {"cluster_id": 54, "members": [63, 72, 202, 268, 488, 613, 616, 745, 771, 1020, 1257, 1322, 1345, 1719, 2053, 2054, 2087, 2091, 2099, 2107, 2165, 2175, 2181, 2213, 2259, 2261, 2337, 2363, 2496, 2539, 2540, 2571, 2577]}, {"cluster_id": 55, "members": [65, 543, 619, 1293, 2021]}, {"cluster_id": 56, "members": [66, 411, 1096, 1465, 1488, 1633, 1799, 2601, 2609, 2631]}, {"cluster_id": 57, "members": [67, 113, 282, 540, 747, 2527, 2628]}, {"cluster_id": 58, "members": [68, 307, 391, 493, 976, 978, 991, 998, 1431, 2365]}, {"cluster_id": 59, "members": [70, 344, 441, 1455, 1537, 2184, 2323]}, {"cluster_id": 60, "members": [71, 73, 206, 558, 876, 881, 1199, 1214, 1589, 1706, 1745, 1751, 2584, 2585, 2596, 2632, 2691]}, {"cluster_id": 61, "members": [74, 497, 1118, 1376, 1652, 1848, 1972, 2419, 2620]}, {"cluster_id": 62, "members": [75, 143, 205, 232, 235, 316, 335, 343, 416, 532, 547, 664, 669, 766, 845, 869, 920, 1052, 1075, 1100, 1124, 1212, 1297, 1299, 1323, 1440, 1701, 1820, 1853, 1854, 1855, 1859, 1860, 1861, 1863, 1864, 1868, 1872, 1875, 1877, 1942, 2299, 2573]}, {"cluster_id": 63, "members": [78, 994, 1219, 1329, 1418]}, {"cluster_id": 64, "members": [79, 522, 603, 716, 975, 1057, 1821, 2097, 2098, 2409]}, {"cluster_id": 65, "members": [80, 257, 434, 964, 965, 1117, 1524, 1693, 2049, 2055, 2198, 2408]}, {"cluster_id": 66, "members": [81, 315, 347, 408, 423, 527, 647, 827, 1921, 2232]}, {"cluster_id": 67, "members": [82, 194, 673, 1050, 1081, 1217, 1230, 1320, 1342, 1349, 1379, 1380, 1491, 1634, 2523, 2563, 2564, 2646, 2684]}, {"cluster_id": 68, "members": [83, 449, 1280, 1516, 2581, 2612]}, {"cluster_id": 69, "members": [84, 284, 583, 677, 786, 947, 954, 1112, 2222, 2223, 2224, 2225, 2226]}, {"cluster_id": 70, "members": [85, 401, 864, 1065, 2210, 2211, 2487, 2488]}, {"cluster_id": 71, "members": [86, 280, 897, 993, 1141, 1152, 1203, 1983, 2155, 2293, 2294, 2295]}, {"cluster_id": 72, "members": [88, 161, 191, 272, 382, 516, 841, 842, 851, 914, 1174, 1197, 1288, 1309, 1551, 1658, 1677, 1741, 1791, 1882, 1989, 2010, 2012, 2014, 2015, 2016, 2104, 2593]}, {"cluster_id": 73, "members": [89, 761, 884, 1087, 1153, 1157, 1401]}, {"cluster_id": 74, "members": [90, 155, 156, 221, 376, 672, 817, 1689, 1763, 2456]}, {"cluster_id": 75, "members": [91, 330, 574, 576, 665, 1018, 2001, 2002, 2044, 2123, 2348, 2378, 2379, 2380]}, {"cluster_id": 76, "members": [93, 100, 104, 117, 135, 139, 146, 150, 151, 154, 162, 164, 166, 169, 172, 178, 180, 182, 190, 198, 199, 212, 222, 223, 229, 233, 234, 237, 244, 245, 254, 259, 261, 264, 265, 266, 271, 273, 281, 292, 298, 300, 311, 319, 325, 328, 334, 351, 353, 356, 358, 360, 362, 368, 372, 374, 392, 413, 455, 460, 491, 550, 593, 595, 626, 654, 655, 690, 718, 758, 782, 821, 1036, 1132, 1201, 1259, 1680, 2056, 2149, 2311, 2342, 2377, 2532, 2537, 2562, 2576, 2617]}, {"cluster_id": 77, "members": [94, 195, 560, 585, 675, 774, 934, 1649, 2355, 2356, 2357, 2424, 2490, 2526]}, {"cluster_id": 78, "members": [95, 525, 861, 1397, 1540, 1628, 2057, 2180]}, {"cluster_id": 79, "members": [97, 661, 1028, 1353, 1883, 1884, 1885, 2134, 2644]}, {"cluster_id": 80, "members": [98, 772, 908, 1013, 1156, 1435, 1521, 1566, 1625, 1661, 1729, 1840, 1851, 1895, 2050, 2071, 2485]}, {"cluster_id": 81, "members": [102, 505, 937, 1180, 1251, 1304, 1448, 1560, 1871, 1878, 2256]}, {"cluster_id": 82, "members": [103, 134, 350, 406, 484, 608, 1775, 2081, 2505]}, {"cluster_id": 83, "members": [105, 1205, 1721, 1761, 2476, 2651]}, {"cluster_id": 84, "members": [107, 388, 971, 1113, 1360, 1478, 1650, 1899, 1902, 1904, 1949, 1960, 2491, 2510, 2602, 2603]}, {"cluster_id": 85, "members": [109, 133, 153, 318, 384, 452, 508, 556, 660, 1045, 1210, 1258, 1337, 1561, 1571, 1623, 1648, 1767, 1768, 1773, 1774, 1779, 1780, 1783, 1784, 1787, 1991, 2092, 2095, 2318, 2325]}, {"cluster_id": 86, "members": [110, 1161, 1599, 2279, 2441]}, {"cluster_id": 87, "members": [112, 426, 487, 1564, 1772, 1782, 1789, 2026, 2469]}, {"cluster_id": 88, "members": [114, 610, 826, 1263, 1407, 2287, 2288, 2481]}, {"cluster_id": 89, "members": [115, 277, 451, 630, 671, 880, 935, 973, 2019]}, {"cluster_id": 90, "members": [116, 130, 211, 1076, 1394, 1487, 1501, 2017, 2137, 2467, 2468]}, {"cluster_id": 91, "members": [118, 255, 260, 365, 404, 446, 554, 685, 743, 781, 795, 836, 903, 967, 989, 1029, 1170, 1343, 1375, 1476, 1645, 2030, 2031, 2112, 2166, 2316, 2402, 2403, 2436, 2572, 2586, 2659]}, {"cluster_id": 92, "members": [119, 132, 479, 517, 646, 904, 952, 1312, 1959, 2105, 2621]}, {"cluster_id": 93, "members": [120, 483, 514, 816, 1842, 1980, 2281, 2405]}, {"cluster_id": 94, "members": [121, 250, 589, 980, 1158, 2251, 2252, 2429]}, {"cluster_id": 95, "members": [124, 236, 287, 294, 306, 329, 417, 474, 656, 699, 773, 852, 940, 958, 1009, 1070, 1072, 1181, 1193, 1245, 1367, 1411, 1459, 1572, 1584, 1640, 1656, 1771, 1778, 1781, 1800, 1801, 1804, 1805, 1829, 2047, 2080, 2088, 2089, 2090, 2218, 2289, 2327, 2328, 2389, 2392, 2451, 2452, 2478, 2598, 2705]}, {"cluster_id": 96, "members": [125, 458, 1022, 1276, 1555, 1953, 1962, 1963]}, {"cluster_id": 97, "members": [126, 215, 1248, 1852, 2045, 2046, 2048, 2079]}, {"cluster_id": 98, "members": [129, 393, 701, 735, 739, 1237, 1543, 1881, 1956, 1958, 2060, 2061, 2062, 2649]}, {"cluster_id": 99, "members": [131, 683, 834, 1547, 2168, 2169]}, {"cluster_id": 100, "members": [136, 741, 831, 1011, 1088, 1638, 1733, 2093]}, {"cluster_id": 101, "members": [137, 397, 879, 2144, 2329, 2331, 2504]}, {"cluster_id": 102, "members": [138, 337, 784, 960, 1499, 1735, 1776, 2078, 2324, 2650]}, {"cluster_id": 103, "members": [140, 444, 584, 623, 2463]}, {"cluster_id": 104, "members": [141, 622, 740, 788, 1002, 1951, 2033, 2204, 2521]}, {"cluster_id": 105, "members": [142, 227, 638, 808, 823, 924, 1283, 1318, 2382, 2383]}, {"cluster_id": 106, "members": [144, 145, 213, 537, 1165, 1593, 1698, 2192, 2622]}, {"cluster_id": 107, "members": [147, 242, 279, 285, 304, 502, 838, 1058, 1172, 1195, 1976, 2280]}, {"cluster_id": 108, "members": [148, 378, 380, 381, 450, 477, 602, 930, 2518, 2538, 2569]}, {"cluster_id": 109, "members": [157, 639, 811, 855, 900, 2493]}, {"cluster_id": 110, "members": [158, 184, 293, 520, 691, 791, 1270, 1336, 1615, 1665, 1889, 1890, 1891, 1893, 2032, 2034, 2035, 2037, 2038, 2039, 2042]}, {"cluster_id": 111, "members": [159, 176, 240, 256, 291, 379, 512, 591, 642, 756, 1017, 1078, 1423, 1472, 1514, 1692, 2187, 2302, 2308, 2472, 2559, 2652]}, {"cluster_id": 112, "members": [160, 267, 670, 1059, 1600, 2000, 2606]}, {"cluster_id": 113, "members": [165, 302, 320, 428, 432, 521, 598, 637, 968, 1126, 1442, 1473, 1597, 1603, 1823, 2462, 2707]}, {"cluster_id": 114, "members": [167, 168, 2437, 2438, 2482]}, {"cluster_id": 115, "members": [170, 609, 762, 859, 860, 1030, 1041, 1067, 1427, 1466, 1712, 2446, 2448]}, {"cluster_id": 116, "members": [171, 775, 790, 1548, 2075, 2076, 2077, 2667, 2668]}, {"cluster_id": 117, "members": [179, 197, 231, 387, 2191, 2412]}, {"cluster_id": 118, "members": [181, 712, 850, 1359, 1464, 2278]}, {"cluster_id": 119, "members": [185, 201, 297, 570, 874, 899, 1858, 2430, 2528]}, {"cluster_id": 120, "members": [186, 1228, 1319, 1391, 1536, 1672, 2580, 2702]}, {"cluster_id": 121, "members": [187, 492, 688, 726, 956, 1109, 1206, 1208, 1533, 2671, 2672, 2673, 2674, 2675, 2678]}, {"cluster_id": 122, "members": [188, 301, 439, 524, 552, 721, 853, 1034, 1169, 1642, 1727, 1731, 1755, 2492, 2657, 2658]}, {"cluster_id": 123, "members": [189, 564, 592, 1144, 1262, 1509, 2669]}, {"cluster_id": 124, "members": [200, 500, 792, 843, 1191, 1439, 1558, 2241, 2242, 2339, 2340, 2410, 2411, 2676]}, {"cluster_id": 125, "members": [204, 415, 818, 966, 1843, 1847, 1850, 1999]}, {"cluster_id": 126, "members": [207, 345, 390, 431, 571, 625, 696, 833, 1024, 1108, 1130, 1570, 1678, 1762, 2036, 2495, 2595, 2694, 2695]}, {"cluster_id": 127, "members": [209, 361, 856, 1390, 1449, 2681, 2683]}, {"cluster_id": 128, "members": [210, 566, 674, 717, 1614, 1626, 1671, 1906, 1907, 1968]}, {"cluster_id": 129, "members": [217, 243, 473, 1532, 1686, 2431, 2432, 2473, 2474, 2647, 2648]}, {"cluster_id": 130, "members": [218, 575, 1348, 1382, 2283, 2605]}, {"cluster_id": 131, "members": [219, 251, 253, 402, 507, 1211, 1300, 1932, 1933, 1936, 1940, 1941, 2508]}, {"cluster_id": 132, "members": [225, 694, 941, 974, 1496, 2255, 2516, 2517, 2533]}, {"cluster_id": 133, "members": [230, 511, 549, 1000, 1095, 1325, 1497, 2613]}, {"cluster_id": 134, "members": [238, 653, 1231, 1409, 1553, 1826, 1827, 1828]}, {"cluster_id": 135, "members": [239, 887, 910, 1220, 1274, 1490, 2085]}, {"cluster_id": 136, "members": [248, 443, 946, 1264, 1531, 1964, 1966, 1971, 2623]}, {"cluster_id": 137, "members": [249, 424, 433, 436, 621, 668, 878, 885, 1039, 1142, 1289, 1332, 1494, 1526, 1622, 1879, 1977, 1978, 1981, 2326, 2425, 2426]}, {"cluster_id": 138, "members": [252, 414, 711, 1285, 2111]}, {"cluster_id": 139, "members": [263, 296, 632, 732, 1103, 1123, 1430, 1480, 1718, 1748, 1886, 1887, 1888, 2685]}, {"cluster_id": 140, "members": [269, 321, 418, 1083, 1225, 1232, 1450, 2543, 2551, 2579]}, {"cluster_id": 141, "members": [270, 276, 314, 339, 348, 467, 495, 729, 804, 1073, 1086, 1136, 1244, 1568, 1575, 1585, 1588, 1590, 1596, 1602, 1636, 1641, 1643, 1647, 1945, 1993, 2074, 2369, 2546, 2607]}, {"cluster_id": 142, "members": [274, 748, 1145, 1374, 1550, 1617]}, {"cluster_id": 143, "members": [278, 1185, 1460, 1992, 2358, 2503, 2589]}, {"cluster_id": 144, "members": [283, 471, 600, 724, 1027, 1138, 1260, 1399, 1420, 1559, 2013, 2228, 2250, 2394, 2395, 2396, 2397, 2665, 2666, 2706]}, {"cluster_id": 145, "members": [289, 312, 489, 513, 657, 1254, 1722, 1740, 2390, 2391, 2440]}, {"cluster_id": 146, "members": [299, 1046, 1162, 1484, 1510, 2386, 2387, 2498, 2587]}, {"cluster_id": 147, "members": [305, 769, 942, 1218, 1923, 1924, 1927, 2116]}, {"cluster_id": 148, "members": [309, 377, 383, 481, 715, 733, 759, 862, 1164, 1192, 1265, 1287, 1424, 1591, 1621, 1704, 1813, 2231, 2291, 2292, 2303, 2541]}, {"cluster_id": 149, "members": [323, 847, 2022, 2188, 2189, 2190, 2193, 2194]}, {"cluster_id": 150, "members": [324, 342, 438, 478, 651, 787, 1512, 1667, 1984, 1985, 2023, 2024, 2145, 2156]}, {"cluster_id": 151, "members": [327, 692, 928, 1326, 1777, 1802, 2063, 2064, 2494, 2629]}, {"cluster_id": 152, "members": [336, 352, 1294, 1432, 1605, 1815, 1944, 1946, 1947, 1950, 2663]}, {"cluster_id": 153, "members": [338, 776, 1340, 1363, 1413, 1438, 1934, 1943, 2664]}, {"cluster_id": 154, "members": [340, 341, 529, 709, 1417, 1419, 2260]}, {"cluster_id": 155, "members": [349, 419, 465, 896, 1403, 2427, 2428]}, {"cluster_id": 156, "members": [359, 389, 905, 969, 1282, 1743]}, {"cluster_id": 157, "members": [363, 536, 567, 631, 760, 848, 985, 1003, 1373, 2459, 2554, 2689]}, {"cluster_id": 158, "members": [364, 650, 819, 1306, 1486, 1606, 1632, 1756, 2642]}, {"cluster_id": 159, "members": [366, 553, 572, 1110, 1120, 1127, 1357, 1583, 1995, 2025, 2028, 2170, 2171]}, {"cluster_id": 160, "members": [370, 371, 400, 720, 1354, 1408, 1414, 1415]}, {"cluster_id": 161, "members": [386, 597, 725, 1104, 1137, 2114, 2115, 2264, 2509, 2519, 2588]}, {"cluster_id": 162, "members": [394, 906, 1063, 1064, 1129, 1518, 2351, 2352, 2353]}, {"cluster_id": 163, "members": [399, 456, 544, 680, 1089, 1580, 1769, 2199, 2200, 2201, 2347]}, {"cluster_id": 164, "members": [403, 806, 1037, 1383, 2447, 2654, 2655]}, {"cluster_id": 165, "members": [407, 409, 667, 679, 695, 948, 1005, 1012, 1395, 1541, 1681, 1682, 1797, 2084, 2266, 2267, 2319, 2320, 2321, 2364]}, {"cluster_id": 166, "members": [410, 612, 802, 807, 1168, 1179, 1226, 1955, 2235]}, {"cluster_id": 167, "members": [425, 464, 526, 533, 580, 615, 1007, 1216, 1742, 1749, 1758, 2307, 2443, 2466]}, {"cluster_id": 168, "members": [427, 565, 1249, 1361, 1398, 1528, 1610, 2556, 2557, 2558]}, {"cluster_id": 169, "members": [430, 912, 983, 1261, 2130, 2249]}, {"cluster_id": 170, "members": [437, 645, 755, 891, 950, 1267, 1277, 1495, 1674, 1898, 1900, 2608]}, {"cluster_id": 171, "members": [445, 909, 959, 1023, 1347, 1434, 1453, 1870, 1873, 1876, 1986, 1987, 1990, 1994, 1996, 1997, 1998, 2003, 2004, 2005, 2007, 2008, 2301, 2529, 2568]}, {"cluster_id": 172, "members": [453, 635, 1278, 1378, 1544, 1818, 2058, 2150, 2183, 2305, 2306, 2552]}, {"cluster_id": 173, "members": [461, 1021, 1268, 1869, 2309]}, {"cluster_id": 174, "members": [462, 611, 714, 783, 810, 893, 984, 1015, 1048, 1068, 1125, 1143, 1324, 1385, 1425, 1452, 1474, 1481, 1618, 1669, 1892, 2265, 2332, 2335, 2690]}, {"cluster_id": 175, "members": [468, 573, 643, 1167, 1311, 1313, 2407]}, {"cluster_id": 176, "members": [506, 778, 779, 822, 921, 1074, 1163, 1352, 1370, 1402, 1482, 1485, 1525, 1539, 1592, 1637, 1675, 1676, 1905, 1914, 1918, 1919, 1973, 2132]}, {"cluster_id": 177, "members": [509, 846, 939, 1101, 1242, 2136, 2146, 2148, 2173, 2174]}, {"cluster_id": 178, "members": [515, 798, 828, 1102, 1207, 1793, 1795, 1817, 2290, 2464, 2465, 2547]}, {"cluster_id": 179, "members": [528, 678, 803, 830, 2100, 2101]}, {"cluster_id": 180, "members": [534, 629, 883, 933, 1038]}, {"cluster_id": 181, "members": [546, 870, 1119, 1235, 1273, 1462, 1608, 1611, 1663, 2124, 2125, 2126, 2127, 2128, 2129]}, {"cluster_id": 182, "members": [548, 825, 944, 1523, 1765, 1770, 1937, 1938, 2138, 2449]}, {"cluster_id": 183, "members": [555, 1463, 1695, 2131, 2133]}, {"cluster_id": 184, "members": [562, 742, 754, 902, 1077, 1284, 1620, 2450]}, {"cluster_id": 185, "members": [586, 682, 780, 925, 1222, 1224, 1291, 1587, 1660, 1732, 2119, 2154, 2341]}, {"cluster_id": 186, "members": [601, 713, 809, 812, 814, 1044, 1365, 2229]}, {"cluster_id": 187, "members": [607, 658, 785, 1084, 1213, 1290, 1513, 1664, 1812, 2147, 2258, 2486, 2550]}, {"cluster_id": 188, "members": [614, 888, 1475, 1556, 2470]}, {"cluster_id": 189, "members": [620, 923, 1186, 2009, 2534, 2535, 2574]}, {"cluster_id": 190, "members": [634, 753, 890, 1227, 1269, 1314, 1630, 1880, 2227]}, {"cluster_id": 191, "members": [641, 868, 988, 1341, 1844, 2276, 2704]}, {"cluster_id": 192, "members": [662, 832, 932, 1469, 1929, 2346, 2349, 2590, 2600]}, {"cluster_id": 193, "members": [693, 707, 763, 1511, 1519, 1635, 2108]}, {"cluster_id": 194, "members": [703, 801, 1915, 1961, 2497]}, {"cluster_id": 195, "members": [704, 1412, 1483, 2113, 2208]}, {"cluster_id": 196, "members": [727, 1151, 1350, 1386, 2570]}, {"cluster_id": 197, "members": [734, 736, 1006, 1302, 1388, 1631, 1841, 2196, 2197, 2202, 2277]}, {"cluster_id": 198, "members": [746, 916, 1026, 1303, 1849, 2006, 2163, 2477, 2531]}, {"cluster_id": 199, "members": [752, 957, 1557, 2083, 2167]}, {"cluster_id": 200, "members": [793, 840, 1896, 1897, 1903, 2592]}, {"cluster_id": 201, "members": [797, 800, 820, 824, 889, 915, 931, 1035, 1433, 1830, 1831, 1832, 1833, 1834]}, {"cluster_id": 202, "members": [799, 961, 1728, 2257, 2338, 2555, 2599]}, {"cluster_id": 203, "members": [837, 996, 1071, 1135, 1194, 1255, 1422, 1461, 1545, 2248, 2670]}, {"cluster_id": 204, "members": [867, 871, 1229, 1252, 1308, 1504, 1506, 1507, 1627, 2442, 2522, 2656]}, {"cluster_id": 205, "members": [886, 1295, 1400, 1573, 2697, 2698, 2700]}, {"cluster_id": 206, "members": [917, 1846, 1867, 2117, 2639]}, {"cluster_id": 207, "members": [918, 1098, 1477, 2065, 2066, 2067]}, {"cluster_id": 208, "members": [922, 1307, 1406, 1444, 2536]}, {"cluster_id": 209, "members": [953, 972, 2565, 2566, 2567]}, {"cluster_id": 210, "members": [979, 1200, 1366, 2207, 2373, 2413, 2414, 2415]}, {"cluster_id": 211, "members": [981, 1054, 1055, 1099, 1159, 1221, 1334, 1577]}, {"cluster_id": 212, "members": [995, 1338, 1609, 1790, 1912, 2102, 2103, 2152, 2304, 2385, 2404]}, {"cluster_id": 213, "members": [997, 1204, 1330, 1837, 2068]}, {"cluster_id": 214, "members": [1082, 1202, 1428, 1429, 1534, 2336, 2520, 2524]}, {"cluster_id": 215, "members": [1085, 1396, 1421, 1975, 2110]}, {"cluster_id": 216, "members": [1115, 1178, 1266, 1529, 1673, 1908, 2350, 2660]}, {"cluster_id": 217, "members": [1133, 1500, 1538, 2109, 2185]}, {"cluster_id": 218, "members": [1184, 1190, 1239, 1387, 2611]}, {"cluster_id": 219, "members": [1240, 1443, 2236, 2237, 2262, 2542]}, {"cluster_id": 220, "members": [1368, 1470, 1824, 1911, 2553]}, {"cluster_id": 221, "members": [1381, 1794, 1796, 2297, 2298]}, {"cluster_id": 222, "members": [1502, 1503, 1595, 2243, 2244, 2245, 2246]}, {"cluster_id": 223, "members": [1549, 1901, 2284, 2285, 2286]}, {"cluster_id": 224, "members": [1554, 1657, 2686, 2687, 2688]}]}

seeds_diam_1e-8.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38713bab14cb6b5def749b13d955bd7c5b69d807eea4a3491f2a2b203a37acec
+size 13141004

seeds_diam_1e0.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3720e74552fda80710fa03addc8b6586425185da4ed281151fe65498a30aa514
+size 13146594

seeds_diam_1e3.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46249fc1669b3e62bafbd256451fae4e9253c7e65a568349cf17d8dc4ea73919
+size 13156713

src/.ipynb_checkpoints/2.2_lrmc_bilevel-checkpoint.py

@@ -0,0 +1,325 @@
+# 2.1_lrmc_bilevel.py
+# Top-1 LRMC ablation with debug guards so seeds differences are visible.
+# Requires: torch, torch_geometric, torch_scatter, torch_sparse
+
+import argparse, json, hashlib
+from pathlib import Path
+from typing import List, Tuple, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch import Tensor
+
+from torch_scatter import scatter_mean
+from torch_sparse import coalesce, spspmm
+from torch_geometric.datasets import Planetoid
+from torch_geometric.nn import GCNConv
+
+
+# ---------------------------
+# Utilities: edges and seeds
+# ---------------------------
+
+def add_scaled_self_loops(edge_index: Tensor,
+                          edge_weight: Optional[Tensor],
+                          num_nodes: int,
+                          scale: float = 1.0) -> Tuple[Tensor, Tensor]:
+    if scale == 0.0:
+        if edge_weight is None:
+            edge_weight = torch.ones(edge_index.size(1), device=edge_index.device)
+        return edge_index, edge_weight
+    device = edge_index.device
+    self_loops = torch.arange(num_nodes, device=device)
+    self_index = torch.stack([self_loops, self_loops], dim=0)
+    self_weight = torch.full((num_nodes,), float(scale), device=device)
+    base_w = edge_weight if edge_weight is not None else torch.ones(edge_index.size(1), device=device)
+    ei = torch.cat([edge_index, self_index], dim=1)
+    ew = torch.cat([base_w, self_weight], dim=0)
+    ei, ew = coalesce(ei, ew, num_nodes, num_nodes, op='add')
+    return ei, ew
+
+
+def adjacency_power(edge_index: Tensor, num_nodes: int, k: int = 2) -> Tensor:
+    # A^2 using spspmm; return binary, coalesced, no self loops
+    row, col = edge_index
+    val = torch.ones(row.numel(), device=edge_index.device)
+    Ai, Av = edge_index, val
+    Ri, _ = spspmm(Ai, Av, Ai, Av, num_nodes, num_nodes, num_nodes)
+    mask = Ri[0] != Ri[1]
+    Ri = Ri[:, mask]
+    Ri, _ = coalesce(Ri, torch.ones(Ri.size(1), device=edge_index.device), num_nodes, num_nodes, op='add')
+    return Ri
+
+
+def build_cluster_graph(edge_index: Tensor,
+                        num_nodes: int,
+                        node2cluster: Tensor,
+                        weight_per_edge: Optional[Tensor] = None,
+                        num_clusters: Optional[int] = None
+                        ) -> Tuple[Tensor, Tensor, int]:
+    K = int(node2cluster.max().item()) + 1 if num_clusters is None else num_clusters
+    src, dst = edge_index
+    csrc = node2cluster[src]
+    cdst = node2cluster[dst]
+    edge_c = torch.stack([csrc, cdst], dim=0)
+    w = weight_per_edge if weight_per_edge is not None else torch.ones(edge_c.size(1), device=edge_c.device)
+    edge_c, w = coalesce(edge_c, w, K, K, op='add')
+    return edge_c, w, K
+
+
+# -----
+# Seeds
+# -----
+
+def _md5(path: Path) -> str:
+    h = hashlib.md5()
+    with path.open('rb') as f:
+        for chunk in iter(lambda: f.read(8192), b''):
+            h.update(chunk)
+    return h.hexdigest()
+
+
+def _extract_members(cluster_obj: dict) -> List[int]:
+    """
+    Try 'members' first, then 'seed_nodes'. Raise if neither works.
+    """
+    m = cluster_obj.get("members", None)
+    if isinstance(m, list) and len(m) > 0:
+        return list(dict.fromkeys(int(x) for x in m))  # dedupe/preserve order
+    m2 = cluster_obj.get("seed_nodes", None)
+    if isinstance(m2, list) and len(m2) > 0:
+        return list(dict.fromkeys(int(x) for x in m2))
+    # If both present but empty, return empty; caller will handle.
+    if isinstance(m, list) or isinstance(m2, list):
+        return []
+    raise KeyError("Cluster object has neither 'members' nor 'seed_nodes'.")
+
+
+def _pick_top1_cluster(obj: dict) -> List[int]:
+    """
+    From {"clusters":[{..., "score":float, "members" or "seed_nodes"}, ...]},
+    choose max by (score, size). Returns deduped member list.
+    """
+    clusters = obj.get("clusters", [])
+    if not isinstance(clusters, list) or len(clusters) == 0:
+        return []
+    def keyfun(c):
+        score = float(c.get("score", 0.0))
+        try:
+            mem = _extract_members(c)
+        except KeyError:
+            mem = []
+        return (score, len(mem))
+    best = max(clusters, key=keyfun)
+    try:
+        members = _extract_members(best)
+    except KeyError:
+        members = []
+    return sorted(set(int(x) for x in members))
+
+
+def load_top1_assignment(seeds_json: str, n_nodes: int, debug: bool = False) -> Tuple[Tensor, Tensor, dict]:
+    """
+    Hard assignment for top-1 LRMC cluster:
+      cluster 0 = top cluster; others are singletons.
+    Returns node2cluster[N], cluster_scores[K,1], and a small debug dict.
+    """
+    p = Path(seeds_json)
+    text = p.read_text(encoding='utf-8')
+    obj = json.loads(text)
+
+    C_star = _pick_top1_cluster(obj)
+    # if len(C_star) > 0 and max(C_star) == n_nodes:
+    # Looks 1-indexed (since max == N, not N-1) → shift down by 1
+    C_star = [u - 1 for u in C_star]
+    C_star = torch.tensor(C_star, dtype=torch.long)
+
+    # C_star = _pick_top1_cluster(obj)
+    # C_star = torch.tensor(C_star, dtype=torch.long)
+    node2cluster = torch.full((n_nodes,), -1, dtype=torch.long)
+
+    if C_star.numel() == 0:
+        # FAIL LOUDLY instead of silently falling back to identity
+        raise RuntimeError(
+            f"No members found for top-1 cluster in {seeds_json}. "
+            f"Expected 'members' or 'seed_nodes' to be non-empty."
+        )
+
+    node2cluster[C_star] = 0
+    outside = torch.tensor(sorted(set(range(n_nodes)) - set(C_star.tolist())), dtype=torch.long)
+    if outside.numel() > 0:
+        node2cluster[outside] = torch.arange(1, 1 + outside.numel(), dtype=torch.long)
+    assert int(node2cluster.min()) >= 0
+
+    K = 1 + outside.numel()
+    cluster_scores = torch.zeros(K, 1, dtype=torch.float32)
+    cluster_scores[0, 0] = 1.0
+
+    info = {
+        "json_md5": _md5(p),
+        "top_cluster_size": int(C_star.numel()),
+        "K": int(K),
+        "n_outside": int(outside.numel()),
+        "first_members": [int(x) for x in C_star[:10].tolist()],
+    }
+    if debug:
+        print(f"[LRMC] Loaded {seeds_json} (md5={info['json_md5']}) | "
+              f"top_size={info['top_cluster_size']} K={info['K']} outside={info['n_outside']} "
+              f"first10={info['first_members']}")
+    return node2cluster, cluster_scores, info
+
+
+# --------------------------
+# Models (baseline + pooled)
+# --------------------------
+
+class GCN2(nn.Module):
+    def __init__(self, in_dim, hid, out_dim, dropout=0.5):
+        super().__init__()
+        self.conv1 = GCNConv(in_dim, hid)
+        self.conv2 = GCNConv(hid, out_dim)
+        self.dropout = dropout
+    def forward(self, x, edge_index):
+        x = F.relu(self.conv1(x, edge_index))
+        x = F.dropout(x, p=self.dropout, training=self.training)
+        x = self.conv2(x, edge_index)
+        return x
+
+
+class OneClusterPool(nn.Module):
+    def __init__(self,
+                 in_dim: int,
+                 hid: int,
+                 out_dim: int,
+                 node2cluster: Tensor,
+                 edge_index_node: Tensor,
+                 num_nodes: int,
+                 self_loop_scale: float = 0.0,
+                 use_a2_for_clusters: bool = False,
+                 debug_header: str = ""):
+        super().__init__()
+        self.n2c = node2cluster.long()
+        self.K = int(self.n2c.max().item()) + 1
+
+        # Node graph (A + λI if desired)
+        ei_node = edge_index_node
+        ei_node, ew_node = add_scaled_self_loops(ei_node, None, num_nodes, scale=self_loop_scale)
+        self.register_buffer("edge_index_node", ei_node)
+        self.register_buffer("edge_weight_node", ew_node)
+
+        # Cluster graph
+        ei_for_c = adjacency_power(edge_index_node, num_nodes, k=2) if use_a2_for_clusters else edge_index_node
+        edge_index_c, edge_weight_c, K = build_cluster_graph(ei_for_c, num_nodes, self.n2c)
+        self.register_buffer("edge_index_c", edge_index_c)
+        self.register_buffer("edge_weight_c", edge_weight_c)
+        self.K = K
+
+        if debug_header:
+            print(f"[POOL] {debug_header} | cluster_edges={edge_index_c.size(1)} (K={K})")
+
+        # Layers
+        self.gcn_node1 = GCNConv(in_dim, hid, add_self_loops=False, normalize=True)
+        self.gcn_cluster = GCNConv(hid, hid, add_self_loops=True,  normalize=True)
+        self.gcn_node2 = GCNConv(hid * 2, out_dim)  # concat [h_node, h_broadcast]
+
+    def forward(self, x: Tensor, edge_index_node: Tensor) -> Tensor:
+        h1 = F.relu(self.gcn_node1(x, self.edge_index_node, self.edge_weight_node))
+        z  = scatter_mean(h1, self.n2c, dim=0, dim_size=self.K)          # [K, H]
+        z2 = F.relu(self.gcn_cluster(z, self.edge_index_c, self.edge_weight_c))
+        hb = z2[self.n2c]                                                # [N, H]
+        hcat = torch.cat([h1, hb], dim=1)                                # [N, 2H]
+        out = self.gcn_node2(hcat, edge_index_node)
+        return out
+
+
+# -------------
+# Training glue
+# -------------
+
+@torch.no_grad()
+def accuracy(logits: Tensor, y: Tensor, mask: Tensor) -> float:
+    pred = logits[mask].argmax(dim=1)
+    return (pred == y[mask]).float().mean().item()
+
+
+def run_train_eval(model: nn.Module, data, epochs=200, lr=0.01, wd=5e-4):
+    opt = torch.optim.Adam(model.parameters(), lr=lr, weight_decay=wd)
+    best_val, best_state = 0.0, None
+    for ep in range(1, epochs + 1):
+        model.train()
+        opt.zero_grad(set_to_none=True)
+        logits = model(data.x, data.edge_index)
+        loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+        loss.backward(); opt.step()
+
+        model.eval()
+        logits = model(data.x, data.edge_index)
+        val_acc = accuracy(logits, data.y, data.val_mask)
+        if val_acc > best_val:
+            best_val, best_state = val_acc, {k: v.detach().clone() for k, v in model.state_dict().items()}
+        if ep % 20 == 0:
+            tr = accuracy(logits, data.y, data.train_mask)
+            te = accuracy(logits, data.y, data.test_mask)
+            print(f"[{ep:04d}] loss={loss.item():.4f}  train={tr:.3f}  val={val_acc:.3f}  test={te:.3f}")
+
+    if best_state is not None:
+        model.load_state_dict(best_state)
+    model.eval()
+    logits = model(data.x, data.edge_index)
+    return {"val": accuracy(logits, data.y, data.val_mask),
+            "test": accuracy(logits, data.y, data.test_mask)}
+
+
+# -----------
+# Entrypoint
+# -----------
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--dataset", required=True, choices=["Cora", "Citeseer", "Pubmed"])
+    ap.add_argument("--seeds", required=True, help="Path to LRMC seeds JSON (single large graph).")
+    ap.add_argument("--variant", choices=["baseline", "pool"], default="pool")
+    ap.add_argument("--hidden", type=int, default=128)
+    ap.add_argument("--epochs", type=int, default=200)
+    ap.add_argument("--lr", type=float, default=0.01)
+    ap.add_argument("--wd", type=float, default=5e-4)
+    ap.add_argument("--dropout", type=float, default=0.5)  # baseline only
+    ap.add_argument("--self_loop_scale", type=float, default=0.0)
+    ap.add_argument("--use_a2", action="store_true", help="Use A^2 for the cluster graph.")
+    ap.add_argument("--seed", type=int, default=42)
+    ap.add_argument("--debug", action="store_true", help="Print seeds md5, cluster size, K, etc.")
+    args = ap.parse_args()
+
+    torch.manual_seed(args.seed)
+
+    ds = Planetoid(root=f"./data/{args.dataset}", name=args.dataset)
+    data = ds[0]
+    in_dim, out_dim, n = ds.num_node_features, ds.num_classes, data.num_nodes
+
+    if args.variant == "baseline":
+        model = GCN2(in_dim, args.hidden, out_dim, dropout=args.dropout)
+        res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+        print(f"Baseline GCN: val={res['val']:.4f}  test={res['test']:.4f}")
+        return
+
+    # pool variant
+    node2cluster, _, info = load_top1_assignment(args.seeds, n, debug=args.debug)
+    dbg_header = f"seeds_md5={info['json_md5']} top_size={info['top_cluster_size']} K={info['K']}"
+
+    model = OneClusterPool(in_dim=in_dim,
+                           hid=args.hidden,
+                           out_dim=out_dim,
+                           node2cluster=node2cluster,
+                           edge_index_node=data.edge_index,
+                           num_nodes=n,
+                           self_loop_scale=args.self_loop_scale,
+                           use_a2_for_clusters=args.use_a2,
+                           debug_header=dbg_header)
+
+    res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+    print(f"L-RMC (top-1 pool): val={res['val']:.4f}  test={res['test']:.4f}")
+
+
+if __name__ == "__main__":
+    main()

src/.ipynb_checkpoints/2.3_lrmc_bilevel-checkpoint.py

@@ -0,0 +1,436 @@
+# 2.3_lrmc_bilevel.py
+# Top-1 LRMC ablation with: cluster refinement (k-core), gated residual fusion,
+# sparsified cluster graph (drop self-loops + per-row top-k), and A + γA² mix.
+# Requires: torch, torch_geometric, torch_scatter, torch_sparse
+
+import argparse, json, hashlib
+from pathlib import Path
+from typing import List, Tuple, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch import Tensor
+
+from torch_scatter import scatter_mean
+from torch_sparse import coalesce, spspmm
+from torch_geometric.datasets import Planetoid
+from torch_geometric.nn import GCNConv
+
+
+# ---------------------------
+# Utilities: edges and seeds
+# ---------------------------
+
+def add_scaled_self_loops(edge_index: Tensor,
+                          edge_weight: Optional[Tensor],
+                          num_nodes: int,
+                          scale: float = 1.0) -> Tuple[Tensor, Tensor]:
+    if scale == 0.0:
+        if edge_weight is None:
+            edge_weight = torch.ones(edge_index.size(1), device=edge_index.device)
+        return edge_index, edge_weight
+    device = edge_index.device
+    self_loops = torch.arange(num_nodes, device=device)
+    self_index = torch.stack([self_loops, self_loops], dim=0)
+    self_weight = torch.full((num_nodes,), float(scale), device=device)
+    base_w = edge_weight if edge_weight is not None else torch.ones(edge_index.size(1), device=device)
+    ei = torch.cat([edge_index, self_index], dim=1)
+    ew = torch.cat([base_w, self_weight], dim=0)
+    ei, ew = coalesce(ei, ew, num_nodes, num_nodes, op='add')
+    return ei, ew
+
+
+def adjacency_power(edge_index: Tensor, num_nodes: int, k: int = 2) -> Tensor:
+    # A^2 using spspmm; return binary, coalesced, no self loops
+    row, col = edge_index
+    val = torch.ones(row.numel(), device=edge_index.device)
+    Ai, Av = edge_index, val
+    Ri, _ = spspmm(Ai, Av, Ai, Av, num_nodes, num_nodes, num_nodes)
+    mask = Ri[0] != Ri[1]
+    Ri = Ri[:, mask]
+    Ri, _ = coalesce(Ri, torch.ones(Ri.size(1), device=edge_index.device), num_nodes, num_nodes, op='add')
+    return Ri
+
+
+def _md5(path: Path) -> str:
+    h = hashlib.md5()
+    with path.open('rb') as f:
+        for chunk in iter(lambda: f.read(8192), b''):
+            h.update(chunk)
+    return h.hexdigest()
+
+
+# -----
+# Seeds
+# -----
+
+def _extract_members(cluster_obj: dict) -> List[int]:
+    m = cluster_obj.get("members", None)
+    if isinstance(m, list) and len(m) > 0:
+        return list(dict.fromkeys(int(x) for x in m))
+    m2 = cluster_obj.get("seed_nodes", None)
+    if isinstance(m2, list) and len(m2) > 0:
+        return list(dict.fromkeys(int(x) for x in m2))
+    if isinstance(m, list) or isinstance(m2, list):
+        return []
+    raise KeyError("Cluster object has neither 'members' nor 'seed_nodes'.")
+
+
+def _pick_top1_cluster(obj: dict) -> List[int]:
+    clusters = obj.get("clusters", [])
+    if not isinstance(clusters, list) or len(clusters) == 0:
+        return []
+    def keyfun(c):
+        score = float(c.get("score", 0.0))
+        try:
+            mem = _extract_members(c)
+        except KeyError:
+            mem = []
+        return (score, len(mem))
+    best = max(clusters, key=keyfun)
+    try:
+        members = _extract_members(best)
+    except KeyError:
+        members = []
+    return sorted(set(int(x) for x in members))
+
+
+def refine_k_core(C_star: List[int], edge_index: Tensor, k: int = 2, rounds: int = 50) -> List[int]:
+    """Refine cluster by taking a k-core of its induced subgraph (label-free purity boost)."""
+    if k <= 0 or len(C_star) == 0:
+        return C_star
+    device = edge_index.device
+    S = torch.tensor(sorted(set(C_star)), device=device, dtype=torch.long)
+    inS = torch.zeros(int(edge_index.max().item()) + 1, dtype=torch.bool, device=device)
+    inS[S] = True
+    ei = edge_index
+    for _ in range(rounds):
+        u, v = ei[0], ei[1]
+        mask_int = inS[u] & inS[v]
+        u_int, v_int = u[mask_int], v[mask_int]
+        if u_int.numel() == 0:
+            break
+        deg = torch.zeros_like(inS, dtype=torch.long)
+        deg.scatter_add_(0, u_int, torch.ones_like(u_int, dtype=torch.long))
+        deg.scatter_add_(0, v_int, torch.ones_like(v_int, dtype=torch.long))
+        keep = inS.clone()
+        kill = (deg < k) & inS
+        if not kill.any():
+            break
+        keep[kill] = False
+        if keep.sum() == inS.sum():
+            break
+        inS = keep
+    out = torch.nonzero(inS, as_tuple=False).view(-1).tolist()
+    # return only nodes that were originally in C_star
+    return sorted(set(out).intersection(set(C_star)))
+
+
+def load_top1_assignment(seeds_json: str, n_nodes: int,
+                         debug: bool = False,
+                         refine_k: int = 0,
+                         edge_index_for_refine: Optional[Tensor] = None) -> Tuple[Tensor, Tensor, dict]:
+    """
+    Hard assignment for top-1 LRMC cluster with optional k-core refinement.
+      cluster 0 = top cluster; others are singletons.
+    """
+    p = Path(seeds_json)
+    obj = json.loads(p.read_text(encoding='utf-8'))
+    C_star = _pick_top1_cluster(obj)
+    if len(C_star) > 0 and max(C_star) == n_nodes:
+        # 1-indexed → shift down
+        C_star = [u - 1 for u in C_star]
+
+    if refine_k > 0:
+        if edge_index_for_refine is None:
+            raise ValueError("--refine_k requires access to edge_index for refinement.")
+        C_star = refine_k_core(C_star, edge_index_for_refine, k=refine_k)
+
+    C = torch.tensor(C_star, dtype=torch.long)
+    if C.numel() == 0:
+        raise RuntimeError(
+            f"No members found for top-1 cluster in {seeds_json}. "
+            f"Expected 'members' or 'seed_nodes' to be non-empty."
+        )
+
+    node2cluster = torch.full((n_nodes,), -1, dtype=torch.long)
+    node2cluster[C] = 0
+    outside = torch.tensor(sorted(set(range(n_nodes)) - set(C.tolist())), dtype=torch.long)
+    if outside.numel() > 0:
+        node2cluster[outside] = torch.arange(1, 1 + outside.numel(), dtype=torch.long)
+
+    K = 1 + outside.numel()
+    cluster_scores = torch.zeros(K, 1, dtype=torch.float32)
+    cluster_scores[0, 0] = 1.0
+
+    info = {
+        "json_md5": _md5(p),
+        "top_cluster_size": int(C.numel()),
+        "K": int(K),
+        "n_outside": int(outside.numel()),
+        "first_members": [int(x) for x in C[:10].tolist()],
+    }
+    if debug:
+        print(f"[LRMC] Loaded {seeds_json} (md5={info['json_md5']}) | "
+              f"top_size={info['top_cluster_size']} K={info['K']} outside={info['n_outside']} "
+              f"first10={info['first_members']}")
+    return node2cluster, cluster_scores, info
+
+
+# ---------------------------
+# Cluster graph construction
+# ---------------------------
+
+def _sparsify_topk(edge_index: Tensor, edge_weight: Tensor, K: int, topk: int) -> Tuple[Tensor, Tensor]:
+    """Keep per-row top-k neighbors by weight; symmetrize and coalesce."""
+    if topk <= 0:
+        return edge_index, edge_weight
+    row, col = edge_index
+    keep = torch.zeros(edge_weight.numel(), dtype=torch.bool, device=edge_weight.device)
+    # simple per-row loop (K ~ 2k is fine)
+    for r in range(K):
+        idx = (row == r).nonzero(as_tuple=False).view(-1)
+        if idx.numel():
+            k = min(topk, idx.numel())
+            _, order = torch.topk(edge_weight[idx], k)
+            keep[idx[order]] = True
+    ei = edge_index[:, keep]
+    ew = edge_weight[keep]
+    # symmetrize
+    rev = torch.stack([ei[1], ei[0]], dim=0)
+    ei2 = torch.cat([ei, rev], dim=1)
+    ew2 = torch.cat([ew, ew], dim=0)
+    ei2, ew2 = coalesce(ei2, ew2, K, K, op='max')
+    return ei2, ew2
+
+
+def build_cluster_graph_mixed(edge_index_node: Tensor,
+                              num_nodes: int,
+                              node2cluster: Tensor,
+                              use_a2: bool,
+                              a2_gamma: float,
+                              drop_self_loops: bool,
+                              topk_per_row: int) -> Tuple[Tensor, Tensor, int]:
+    """
+    Build A_c = S^T (A + γ A²) S, optionally drop diag, then per-row top-k sparsify.
+    """
+    device = edge_index_node.device
+    # combine A and γA² at node level
+    row, col = edge_index_node
+    wA = torch.ones(row.numel(), device=device)
+    e_all = edge_index_node
+    w_all = wA
+    if use_a2 and a2_gamma > 0.0:
+        A2 = adjacency_power(edge_index_node, num_nodes, k=2)
+        wA2 = torch.full((A2.size(1),), float(a2_gamma), device=device)
+        e_all = torch.cat([e_all, A2], dim=1)
+        w_all = torch.cat([w_all, wA2], dim=0)
+
+    # project to clusters: S^T * (⋅) * S
+    K = int(node2cluster.max().item()) + 1
+    src, dst = e_all
+    csrc = node2cluster[src]
+    cdst = node2cluster[dst]
+    eC = torch.stack([csrc, cdst], dim=0)
+    eC, wC = coalesce(eC, w_all, K, K, op='add')
+
+    if drop_self_loops:
+        mask = eC[0] != eC[1]
+        eC, wC = eC[:, mask], wC[mask]
+
+    if topk_per_row > 0:
+        eC, wC = _sparsify_topk(eC, wC, K, topk_per_row)
+
+    return eC, wC, K
+
+
+# --------------------------
+# Models (baseline + pooled)
+# --------------------------
+
+class GCN2(nn.Module):
+    def __init__(self, in_dim, hid, out_dim, dropout=0.5):
+        super().__init__()
+        self.conv1 = GCNConv(in_dim, hid)
+        self.conv2 = GCNConv(hid, out_dim)
+        self.dropout = dropout
+    def forward(self, x, edge_index):
+        x = F.relu(self.conv1(x, edge_index))
+        x = F.dropout(x, p=self.dropout, training=self.training)
+        x = self.conv2(x, edge_index)
+        return x
+
+
+class OneClusterPoolGated(nn.Module):
+    """
+    Node-GCN -> pool (means) -> Cluster-GCN over sparsified A_c -> residual gate -> Node-GCN -> logits
+    """
+    def __init__(self,
+                 in_dim: int,
+                 hid: int,
+                 out_dim: int,
+                 node2cluster: Tensor,
+                 edge_index_node: Tensor,
+                 num_nodes: int,
+                 self_loop_scale: float = 0.0,
+                 use_a2_for_clusters: bool = False,
+                 a2_gamma: float = 0.2,
+                 drop_cluster_self_loops: bool = True,
+                 cluster_topk: int = 24,
+                 debug_header: str = ""):
+        super().__init__()
+        self.n2c = node2cluster.long()
+        self.K = int(self.n2c.max().item()) + 1
+
+        # Node graph (A + λI)
+        ei_node = edge_index_node
+        ei_node, ew_node = add_scaled_self_loops(ei_node, None, num_nodes, scale=self_loop_scale)
+        self.register_buffer("edge_index_node", ei_node)
+        self.register_buffer("edge_weight_node", ew_node)
+
+        # Cluster graph: A_c = S^T (A + γA²) S → drop diag → per-row top-k
+        eC, wC, K = build_cluster_graph_mixed(
+            edge_index_node, num_nodes, self.n2c,
+            use_a2=use_a2_for_clusters, a2_gamma=a2_gamma,
+            drop_self_loops=drop_cluster_self_loops, topk_per_row=cluster_topk
+        )
+        self.register_buffer("edge_index_c", eC)
+        self.register_buffer("edge_weight_c", wC)
+        self.K = K
+
+        if debug_header:
+            print(f"[POOL] {debug_header} | cluster_edges={eC.size(1)} (K={K})")
+
+        # Layers: gated residual fusion
+        self.gcn_node1 = GCNConv(in_dim, hid, add_self_loops=False, normalize=True)
+        self.gcn_cluster = GCNConv(hid, hid, add_self_loops=True,  normalize=True)
+        self.down = nn.Linear(hid, hid)
+        self.gate = nn.Sequential(nn.Linear(2*hid, hid//2), nn.ReLU(), nn.Linear(hid//2, 1))
+        self.lambda_logit = nn.Parameter(torch.tensor(0.0))
+        self.gcn_node2 = GCNConv(hid, out_dim)  # final node conv on gated residual
+
+    def forward(self, x: Tensor, edge_index_node: Tensor) -> Tensor:
+        # node step
+        h1 = F.relu(self.gcn_node1(x, self.edge_index_node, self.edge_weight_node))
+        # pool
+        z  = scatter_mean(h1, self.n2c, dim=0, dim_size=self.K)          # [K, H]
+        # cluster step
+        z2 = F.relu(self.gcn_cluster(z, self.edge_index_c, self.edge_weight_c))
+        # broadcast + gated residual
+        hb = z2[self.n2c]                                                # [N, H]
+        inj = self.down(hb)
+        gate_dyn = torch.sigmoid(self.gate(torch.cat([h1, inj], dim=1))) # [N,1]
+        lam = torch.sigmoid(self.lambda_logit)                           # scalar in (0,1)
+        alpha = lam * 1.0 + (1.0 - lam) * gate_dyn
+        h2 = h1 + alpha * inj
+        h2 = F.dropout(h2, p=0.5, training=self.training)
+        # final node conv (use same weighted adjacency)
+        out = self.gcn_node2(h2, self.edge_index_node, self.edge_weight_node)
+        return out
+
+
+# -------------
+# Training glue
+# -------------
+
+@torch.no_grad()
+def accuracy(logits: Tensor, y: Tensor, mask: Tensor) -> float:
+    pred = logits[mask].argmax(dim=1)
+    return (pred == y[mask]).float().mean().item()
+
+
+def run_train_eval(model: nn.Module, data, epochs=200, lr=0.01, wd=5e-4):
+    opt = torch.optim.Adam(model.parameters(), lr=lr, weight_decay=wd)
+    best_val, best_state = 0.0, None
+    for ep in range(1, epochs + 1):
+        model.train()
+        opt.zero_grad(set_to_none=True)
+        logits = model(data.x, data.edge_index)
+        loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+        loss.backward(); opt.step()
+
+        model.eval()
+        logits = model(data.x, data.edge_index)
+        val_acc = accuracy(logits, data.y, data.val_mask)
+        if val_acc > best_val:
+            best_val, best_state = val_acc, {k: v.detach().clone() for k, v in model.state_dict().items()}
+        if ep % 20 == 0:
+            tr = accuracy(logits, data.y, data.train_mask)
+            te = accuracy(logits, data.y, data.test_mask)
+            print(f"[{ep:04d}] loss={loss.item():.4f}  train={tr:.3f}  val={val_acc:.3f}  test={te:.3f}")
+
+    if best_state is not None:
+        model.load_state_dict(best_state)
+    model.eval()
+    logits = model(data.x, data.edge_index)
+    return {"val": accuracy(logits, data.y, data.val_mask),
+            "test": accuracy(logits, data.y, data.test_mask)}
+
+
+# -----------
+# Entrypoint
+# -----------
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--dataset", required=True, choices=["Cora", "Citeseer", "Pubmed"])
+    ap.add_argument("--seeds", required=True, help="Path to LRMC seeds JSON (single large graph).")
+    ap.add_argument("--variant", choices=["baseline", "pool"], default="pool")
+    ap.add_argument("--hidden", type=int, default=128)
+    ap.add_argument("--epochs", type=int, default=200)
+    ap.add_argument("--lr", type=float, default=0.01)
+    ap.add_argument("--wd", type=float, default=5e-4)
+    ap.add_argument("--dropout", type=float, default=0.5)  # baseline only
+    ap.add_argument("--self_loop_scale", type=float, default=0.0)
+
+    # NEW knobs for cluster graph & refinement
+    ap.add_argument("--use_a2", action="store_true", help="Include A^2 in cluster graph.")
+    ap.add_argument("--a2_gamma", type=float, default=0.2, help="Weight for A^2 in A + γA^2.")
+    ap.add_argument("--cluster_topk", type=int, default=24, help="Top-k neighbors per cluster row to keep.")
+    ap.add_argument("--drop_cluster_self_loops", action="store_true", help="Drop (c,c) in cluster graph.")
+    ap.add_argument("--refine_k", type=int, default=0, help="k-core refinement on the top cluster (e.g., 2).")
+
+    ap.add_argument("--seed", type=int, default=42)
+    ap.add_argument("--debug", action="store_true")
+    args = ap.parse_args()
+
+    torch.manual_seed(args.seed)
+
+    ds = Planetoid(root=f"./data/{args.dataset}", name=args.dataset)
+    data = ds[0]
+    in_dim, out_dim, n = ds.num_node_features, ds.num_classes, data.num_nodes
+
+    if args.variant == "baseline":
+        model = GCN2(in_dim, args.hidden, out_dim, dropout=args.dropout)
+        res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+        print(f"Baseline GCN: val={res['val']:.4f}  test={res['test']:.4f}")
+        return
+
+    # pool variant
+    node2cluster, _, info = load_top1_assignment(
+        args.seeds, n, debug=args.debug, refine_k=args.refine_k, edge_index_for_refine=data.edge_index
+    )
+    dbg_header = f"seeds_md5={info['json_md5']} top_size={info['top_cluster_size']} K={info['K']}"
+
+    model = OneClusterPoolGated(
+        in_dim=in_dim,
+        hid=args.hidden,
+        out_dim=out_dim,
+        node2cluster=node2cluster,
+        edge_index_node=data.edge_index,
+        num_nodes=n,
+        self_loop_scale=args.self_loop_scale,
+        use_a2_for_clusters=args.use_a2,
+        a2_gamma=args.a2_gamma,
+        drop_cluster_self_loops=args.drop_cluster_self_loops,
+        cluster_topk=args.cluster_topk,
+        debug_header=dbg_header
+    )
+
+    res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+    print(f"L-RMC (top-1 pool, gated): val={res['val']:.4f}  test={res['test']:.4f}")
+
+
+if __name__ == "__main__":
+    main()

src/.ipynb_checkpoints/2.4_lrmc_bilevel-checkpoint.py

@@ -0,0 +1,413 @@
+# lrmc_bilevel.py
+# Top-1 LRMC ablation: one-cluster pooling vs. plain GCN on Planetoid (e.g., Cora)
+# Requires: torch, torch_geometric, torch_scatter, torch_sparse
+#
+# Usage examples:
+#   python lrmc_bilevel.py --dataset Cora --seeds /path/to/lrmc_seeds.json --variant baseline
+#   python lrmc_bilevel.py --dataset Cora --seeds /path/to/lrmc_seeds.json --variant pool --lrmc_inv_weight 0.01 --lrmc_gamma 0.7
+#
+# Notes:
+#  - We read your LRMC JSON, pick the single cluster with the highest 'score',
+#    assign it to cluster id 0, and make all other nodes singletons (1..K-1).
+#  - For --variant pool: Node-GCN -> pool (means) -> Cluster-GCN -> broadcast + skip -> Node-GCN -> classifier
+#    This variant now also applies the L-RMC stability tricks.
+#  - For --variant baseline: Standard 2-layer GCN.
+#  - Keep flags like --self_loop_scale and --use_a2 if you want A+λI / A^2 augmentation.
+
+import argparse, json
+from pathlib import Path
+from typing import List, Tuple, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from torch import Tensor
+from torch_scatter import scatter_add, scatter_mean
+from torch_sparse import coalesce, spspmm
+
+from torch_geometric.datasets import Planetoid
+from torch_geometric.nn import GCNConv
+from torch_geometric.utils import subgraph, degree # Added for stability score
+
+# ---------------------------
+# Utilities: edges and seeds
+# ---------------------------
+
+def add_scaled_self_loops(edge_index: Tensor,
+                          edge_weight: Optional[Tensor],
+                          num_nodes: int,
+                          scale: float = 1.0) -> Tuple[Tensor, Tensor]:
+    """Add self-loops with chosen weight (scale). If scale=0, return unchanged (and create weights if None)."""
+    if scale == 0.0:
+        if edge_weight is None:
+            edge_weight = torch.ones(edge_index.size(1), device=edge_index.device)
+        return edge_index, edge_weight
+    device = edge_index.device
+    self_loops = torch.arange(num_nodes, device=device)
+    self_index = torch.stack([self_loops, self_loops], dim=0)
+    self_weight = torch.full((num_nodes,), float(scale), device=device)
+    base_w = edge_weight if edge_weight is not None else torch.ones(edge_index.size(1), device=device)
+    ei = torch.cat([edge_index, self_index], dim=1)
+    ew = torch.cat([base_w, self_weight], dim=0)
+    ei, ew = coalesce(ei, ew, num_nodes, num_nodes, op='add')
+    return ei, ew
+
+
+def adjacency_power(edge_index: Tensor, num_nodes: int, k: int = 2) -> Tensor:
+    """
+    Compute (binary) k-th power adjacency using sparse matmul (torch_sparse.spspmm).
+    Here we use k=2. Returns coalesced edge_index without weights.
+    """
+    row, col = edge_index
+    val = torch.ones(row.numel(), device=edge_index.device)
+    Ai, Av = edge_index, val
+    # A^2
+    Ri, Rv = spspmm(Ai, Av, Ai, Av, num_nodes, num_nodes, num_nodes)
+    mask = Ri[0] != Ri[1]  # drop diagonal; add custom self-loops later if desired
+    Ri = Ri[:, mask]
+    Ri, _ = coalesce(Ri, torch.ones(Ri.size(1), device=edge_index.device), num_nodes, num_nodes, op='add')
+    return Ri
+
+
+def build_cluster_graph(edge_index: Tensor,
+                        num_nodes: int,
+                        node2cluster: Tensor,
+                        weight_per_edge: Optional[Tensor] = None,
+                        num_clusters: Optional[int] = None
+                        ) -> Tuple[Tensor, Tensor, int]:
+    """
+    Build cluster graph A_c = S^T A S with summed multiplicities as weights.
+    node2cluster: [N] long tensor mapping each node -> cluster id.
+    """
+    K = int(node2cluster.max().item()) + 1 if num_clusters is None else num_clusters
+    src, dst = edge_index
+    csrc = node2cluster[src]
+    cdst = node2cluster[dst]
+    edge_c = torch.stack([csrc, cdst], dim=0)
+    w = weight_per_edge if weight_per_edge is not None else torch.ones(edge_c.size(1), device=edge_c.device)
+    edge_c, w = coalesce(edge_c, w, K, K, op='add')  # sum multiplicities
+    return edge_c, w, K
+
+
+# -----
+# Seeds
+# -----
+
+def _pick_top1_cluster(obj: dict) -> List[int]:
+    """
+    From LRMC JSON with structure: {"clusters":[{"members":[...], "score":float, ...}, ...]}
+    choose the cluster with max (score, size) and return its members.
+    """
+    clusters = obj.get("clusters", [])
+    if not clusters:
+        return []
+    # choose by highest score, then by size (tiebreaker)
+    best = max(clusters, key=lambda c: (float(c.get("score", 0.0)), len(c.get("members", []))))
+    return list(best.get("members", []))
+
+
+def load_top1_assignment(seeds_json: str, n_nodes: int) -> Tuple[Tensor, Tensor, Tensor]:
+    """
+    Create a hard assignment for top-1 LRMC cluster:
+      - cluster 0 = top-1 LRMC set
+      - nodes outside are singletons (1..K-1)
+    Returns:
+      node2cluster: [N] long
+      cluster_scores: [K,1] with 1.0 for top cluster, 0.0 for singletons
+      core_nodes: [|C|] long, original indices of nodes in the top-1 LRMC cluster
+    """
+    obj = json.loads(Path(seeds_json).read_text())
+    C_star_list = _pick_top1_cluster(obj)
+    C_star = torch.tensor(sorted(set(C_star_list)), dtype=torch.long) # Original indices of core nodes
+
+    node2cluster = torch.full((n_nodes,), -1, dtype=torch.long)
+    node2cluster[C_star] = 0
+    outside = torch.tensor(sorted(set(range(n_nodes)) - set(C_star.tolist())), dtype=torch.long)
+    if outside.numel() > 0:
+        node2cluster[outside] = torch.arange(1, 1 + outside.numel(), dtype=torch.long)
+    assert int(node2cluster.min()) >= 0, "All nodes must be assigned."
+
+    K = 1 + outside.numel()
+    cluster_scores = torch.zeros(K, 1, dtype=torch.float32)
+    if C_star.numel() > 0:
+        cluster_scores[0, 0] = 1.0  # emphasize the supercluster
+    return node2cluster, cluster_scores, C_star
+
+
+# --------------------------
+# Models (baseline + pooled)
+# --------------------------
+
+class GCN2(nn.Module):
+    """Plain 2-layer GCN baseline."""
+    def __init__(self, in_dim, hid, out_dim, dropout_p: float = 0.5):
+        super().__init__()
+        self.conv1 = GCNConv(in_dim, hid)
+        self.conv2 = GCNConv(hid, out_dim)
+        self.dropout_p = dropout_p
+
+    def forward(self, x, edge_index):
+        x = F.relu(self.conv1(x, edge_index))
+        x = F.dropout(x, p=self.dropout_p, training=self.training)
+        x = self.conv2(x, edge_index)
+        return x
+
+
+class OneClusterPool(nn.Module):
+    """
+    Node-GCN -> pool to one-cluster + singletons -> Cluster-GCN -> broadcast + skip -> Node-GCN -> classifier
+    This version includes L-RMC stability tricks:
+    1. Backbone-invariance regularizer (loss computed in forward).
+    2. Boundary damping on node graph.
+    """
+    def __init__(self,
+                 in_dim: int,
+                 hid: int,
+                 out_dim: int,
+                 node2cluster: Tensor,
+                 core_nodes: Tensor, # New: explicit core nodes
+                 edge_index_node: Tensor,
+                 num_nodes: int,
+                 self_loop_scale: float = 0.0,
+                 use_a2_for_clusters: bool = False,
+                 lrmc_gamma: float = 1.0,  # New: damping factor (1.0 means no damping)
+                 dropout_p: float = 0.5):
+        super().__init__()
+        self.n2c = node2cluster.long()
+        self.K = int(self.n2c.max().item()) + 1
+        self.core_nodes = core_nodes # Store original indices of core nodes
+        self.lrmc_gamma = lrmc_gamma
+        self.dropout_p = dropout_p
+
+        # Node graph (A + λI if desired)
+        ei_node = edge_index_node
+        ew_node_base = None # Will be created by add_scaled_self_loops if None
+        ei_node, ew_node = add_scaled_self_loops(ei_node, ew_node_base, num_nodes, scale=self_loop_scale)
+
+        # --- Apply Boundary Damping ---
+        if self.lrmc_gamma < 1.0 and self.core_nodes.numel() > 0:
+            is_core = torch.zeros(num_nodes, dtype=torch.bool, device=ei_node.device)
+            is_core[self.core_nodes] = True
+
+            src_is_core = is_core[ei_node[0]]
+            dst_is_core = is_core[ei_node[1]]
+            cross_boundary_mask = (src_is_core != dst_is_core)
+
+            ew_node[cross_boundary_mask] *= self.lrmc_gamma
+        # --- End Boundary Damping ---
+
+        self.register_buffer("edge_index_node", ei_node)
+        self.register_buffer("edge_weight_node", ew_node)
+
+        # Cluster graph from A or A^2
+        ei_for_c = adjacency_power(edge_index_node, num_nodes, k=2) if use_a2_for_clusters else edge_index_node
+        edge_index_c, edge_weight_c, K = build_cluster_graph(ei_for_c, num_nodes, self.n2c)
+        self.register_buffer("edge_index_c", edge_index_c)
+        self.register_buffer("edge_weight_c", edge_weight_c)
+        self.K = K
+
+        # Layers
+        self.gcn_node1 = GCNConv(in_dim, hid, add_self_loops=False, normalize=True)
+        self.gcn_cluster = GCNConv(hid, hid, add_self_loops=True,  normalize=True)
+        self.gcn_node2 = GCNConv(hid * 2, out_dim)  # on concatenated [h_node, h_broadcast]
+
+    def forward(self, x: Tensor, edge_index_node: Tensor) -> Tuple[Tensor, Optional[Tensor]]:
+        # Node GCN (uses stored weights)
+        h1 = F.relu(self.gcn_node1(x, self.edge_index_node, self.edge_weight_node))
+        h1 = F.dropout(h1, p=self.dropout_p, training=self.training) # Consistent dropout
+
+        # --- Backbone-invariance regularizer ---
+        lrmc_inv_loss = None
+        # Apply only if core nodes exist AND regularizer weight is positive (handled by run_train_eval)
+        if self.core_nodes.numel() > 0:
+            core_embeddings = h1[self.core_nodes]
+            # Calculate mean embedding of the core, keepdim=True for broadcasting
+            avg_embedding = core_embeddings.mean(dim=0, keepdim=True)
+            # MSE between each core embedding and the average core embedding
+            lrmc_inv_loss = F.mse_loss(core_embeddings, avg_embedding.expand_as(core_embeddings), reduction='mean')
+        # --- End Backbone-invariance regularizer ---
+
+        # Pool to clusters: mean per cluster
+        z = scatter_mean(h1, self.n2c, dim=0, dim_size=self.K)  # [K, H]
+
+        # Cluster GCN
+        z2 = F.relu(self.gcn_cluster(z, self.edge_index_c, self.edge_weight_c))
+
+        # Broadcast back + skip concat
+        hb = z2[self.n2c]                    # [N, H]
+        hcat = torch.cat([h1, hb], dim=1)    # [N, 2H]
+
+        # Final node GCN head -> logits
+        out = self.gcn_node2(hcat, edge_index_node)
+        return out, lrmc_inv_loss
+
+
+# -------------
+# Training glue
+# -------------
+
+@torch.no_grad()
+def accuracy(logits: Tensor, y: Tensor, mask: Tensor) -> float:
+    pred = logits[mask].argmax(dim=1)
+    return (pred == y[mask]).float().mean().item()
+
+
+def run_train_eval(model: nn.Module, data, epochs=200, lr=0.01, wd=5e-4, lrmc_inv_weight: float = 0.0):
+    opt = torch.optim.Adam(model.parameters(), lr=lr, weight_decay=wd)
+    best_val, best_state = 0.0, None
+    for ep in range(1, epochs + 1):
+        model.train()
+        opt.zero_grad(set_to_none=True)
+
+        # Model output depends on its type
+        res = model(data.x, data.edge_index)
+
+        current_lrmc_inv_loss = None
+        if isinstance(model, OneClusterPool): # OneClusterPool returns (logits, lrmc_inv_loss)
+            logits, current_lrmc_inv_loss = res
+            loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+            if current_lrmc_inv_loss is not None and lrmc_inv_weight > 0:
+                loss += lrmc_inv_weight * current_lrmc_inv_loss
+        else: # GCN2 (baseline) returns only logits
+            logits = res
+            loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+
+        loss.backward(); opt.step()
+
+        # track best on val
+        model.eval()
+        # For evaluation, we only need logits. If it's OneClusterPool, ignore the loss.
+        logits_eval, _ = model(data.x, data.edge_index) if isinstance(model, OneClusterPool) else (model(data.x, data.edge_index), None)
+
+        val_acc = accuracy(logits_eval, data.y, data.val_mask)
+        if val_acc > best_val:
+            best_val, best_state = val_acc, {k: v.detach().clone() for k, v in model.state_dict().items()}
+        if ep % 20 == 0:
+            tr = accuracy(logits_eval, data.y, data.train_mask)
+            te = accuracy(logits_eval, data.y, data.test_mask)
+            lrmc_loss_str = f" inv_l={current_lrmc_inv_loss.item():.4f}" if current_lrmc_inv_loss is not None else ""
+            print(f"[{ep:04d}] loss={loss.item():.4f}{lrmc_loss_str} train={tr:.3f}  val={val_acc:.3f}  test={te:.3f}")
+
+    # test @ best val
+    if best_state is not None:
+        model.load_state_dict(best_state)
+    model.eval()
+    logits_final, _ = model(data.x, data.edge_index) if isinstance(model, OneClusterPool) else (model(data.x, data.edge_index), None)
+    return {
+        "val": accuracy(logits_final, data.y, data.val_mask),
+        "test": accuracy(logits_final, data.y, data.test_mask)
+    }
+
+# --------------------------
+# LRMC Stability Score (new)
+# --------------------------
+
+def compute_lrmc_stability_score(core_nodes: Tensor, edge_index: Tensor, num_nodes: int, epsilon: float = 1e-6) -> float:
+    """
+    Computes the L-RMC stability score S_L(C) = |C| / (d^T L_C d + epsilon).
+    Here, d^T L_C d is interpreted as sum of squared degree differences over edges within the core:
+    sum_{(u,v) in E_C} (deg_C(u) - deg_C(v))^2.
+    """
+    if core_nodes.numel() == 0:
+        return 0.0
+
+    # Get the induced subgraph for core_nodes, relabeling nodes to [0, ..., |C|-1]
+    # `subgraph` returns (sub_edge_index, edge_attr, mapping_nodes, edge_mask)
+    sub_edge_index, _, _, _ = subgraph(core_nodes, edge_index, relabel_nodes=True, num_nodes=num_nodes)
+
+    num_core_nodes = core_nodes.numel()
+
+    if sub_edge_index.numel() == 0:
+        # If the core has nodes but no internal edges, there's no degree variability over edges.
+        # d^T L_C d = 0 in this interpretation, leading to max stability score.
+        return float(num_core_nodes) / epsilon
+
+    # Compute degrees of nodes *within the induced subgraph*
+    # `degree` sums the occurrences of nodes in `sub_edge_index[0]` (sources).
+    # For undirected graphs, total degree is (in_degree + out_degree).
+    # Since `sub_edge_index` contains both (u,v) and (v,u) if relabel_nodes=False, we just need one pass.
+    # With relabel_nodes=True, it's a new adjacency matrix for `num_core_nodes`.
+    # Summing degrees from both source and destination to get total degree in undirected graph.
+    degrees_in_subgraph_full = degree(sub_edge_index[0], num_nodes=num_core_nodes, dtype=torch.float) + \
+                               degree(sub_edge_index[1], num_nodes=num_core_nodes, dtype=torch.float)
+
+    # Calculate sum of squared degree differences for each edge within the subgraph
+    # deg_u_relabel and deg_v_relabel are the degrees of source and destination nodes
+    # (after relabeling) within the induced subgraph.
+    deg_u_relabel = degrees_in_subgraph_full[sub_edge_index[0]]
+    deg_v_relabel = degrees_in_subgraph_full[sub_edge_index[1]]
+
+    degree_variability_sum = torch.sum((deg_u_relabel - deg_v_relabel)**2)
+
+    # Compute S_L(C)
+    score = float(num_core_nodes) / (degree_variability_sum.item() + epsilon)
+    return score
+
+
+# -----------
+# Entrypoint
+# -----------
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--dataset", required=True, choices=["Cora", "Citeseer", "Pubmed"])
+    ap.add_argument("--seeds", required=True, help="Path to LRMC seeds JSON (single large graph).")
+    ap.add_argument("--variant", choices=["baseline", "pool"], default="pool",
+                    help="baseline=plain GCN; pool=top-1 LRMC one-cluster pooling (with new L-RMC tricks)")
+    ap.add_argument("--hidden", type=int, default=128)
+    ap.add_argument("--epochs", type=int, default=200)
+    ap.add_argument("--lr", type=float, default=0.01)
+    ap.add_argument("--wd", type=float, default=5e-4)
+    ap.add_argument("--dropout", type=float, default=0.5, help="Dropout rate for GCN layers.")
+    ap.add_argument("--self_loop_scale", type=float, default=0.0, help="λ for A+λI on node graph (0 disables)")
+    ap.add_argument("--use_a2", action="store_true", help="Use A^2 to build the cluster graph (recommended for pool)")
+    ap.add_argument("--lrmc_inv_weight", type=float, default=0.0,
+                    help="Weight for the backbone-invariance regularizer (0 disables).")
+    ap.add_argument("--lrmc_gamma", type=float, default=1.0,
+                    help="Damping factor for cross-boundary edges (1.0 means no damping).")
+    ap.add_argument("--seed", type=int, default=42)
+    args = ap.parse_args()
+
+    torch.manual_seed(args.seed)
+
+    # Load dataset
+    ds = Planetoid(root=f"./data/{args.dataset}", name=args.dataset)
+    data = ds[0]
+    in_dim, out_dim, n = ds.num_node_features, ds.num_classes, data.num_nodes
+
+    if args.variant == "baseline":
+        model = GCN2(in_dim, args.hidden, out_dim, dropout_p=args.dropout)
+        # use default add_self_loops=True behavior inside convs
+        res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+        print(f"Baseline GCN: val={res['val']:.4f}  test={res['test']:.4f}")
+        return
+
+    # Top-1 LRMC assignment
+    node2cluster, _, core_nodes = load_top1_assignment(args.seeds, n)
+
+    # For informational purposes, compute and print the stability score for the found core
+    lrmc_score = compute_lrmc_stability_score(core_nodes, data.edge_index, n)
+    print(f"LRMC Top-1 Cluster Size: {core_nodes.numel()} nodes. Stability Score (S_L): {lrmc_score:.4f}")
+
+
+    # One-cluster pooled model with L-RMC tricks
+    model = OneClusterPool(in_dim=in_dim,
+                           hid=args.hidden,
+                           out_dim=out_dim,
+                           node2cluster=node2cluster,
+                           core_nodes=core_nodes, # Pass core nodes
+                           edge_index_node=data.edge_index,
+                           num_nodes=n,
+                           self_loop_scale=args.self_loop_scale,
+                           use_a2_for_clusters=args.use_a2,
+                           lrmc_gamma=args.lrmc_gamma, # Pass damping factor
+                           dropout_p=args.dropout) # Pass dropout rate
+
+    res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd,
+                         lrmc_inv_weight=args.lrmc_inv_weight) # Pass regularizer weight
+    print(f"L-RMC (top-1 pool with tricks): val={res['val']:.4f}  test={res['test']:.4f}")
+
+
+if __name__ == "__main__":
+    main()

src/.ipynb_checkpoints/export_edgelist-checkpoint.py

@@ -0,0 +1,158 @@
+#!/usr/bin/env python3
+"""
+This script extracts canonical – 0‑based, undirected, duplicate‑free – edgelists from
+PyTorch Geometric datasets.
+
+Usage
+-----
+
+    python export_edgelist.py [--data_root <root>] <dataset_name> <edges_out> [--dot <dot_file>]
+
+Arguments
+---------
+dataset_name : str
+    Name of a PyG dataset that can be loaded via `Planetoid(..)` or `TUDataset(..)`.
+
+edges_out : str
+    * For a single‑graph dataset (e.g. Planetoid Cora) – write to this file.
+    * For a multi‑graph dataset (e.g. TUDataset “Facebook”) – write to this
+      directory, one file per graph (`graph_000000.txt`, …).
+
+--dot path/to/file.dot : str, optional
+    If supplied, a GraphViz dot file is written that visualises all edgelists
+    (each graph gets its own subgraph in the file).
+
+The script prints *nothing* to stdout – the only side‑effect is the edge‑list
+(and, optionally, the .dot) output.
+
+"""
+
+from __future__ import annotations
+
+import argparse
+from pathlib import Path
+from typing import Iterable, Set, Tuple, List
+
+# ---------------------------------------------------------------------------
+
+def canonical_edges(edge_index) -> Set[Tuple[int, int]]:
+    """Return a set of undirected edges `(u, v)` with `u < v` and `u != v`."""
+    seen: Set[Tuple[int, int]] = set()
+    for u, v in edge_index.t().tolist():
+        if u == v:
+            continue
+        if u > v:
+            u, v = v, u
+        seen.add((u, v))
+    return seen
+
+
+def write_edges(out_file: Path, edges: Iterable[Tuple[int, int]]) -> None:
+    """Write edges to `out_file` in `u v\n` format."""
+    out_file.parent.mkdir(parents=True, exist_ok=True)
+    with out_file.open("w", encoding="utf-8") as f:
+        for u, v in sorted(edges):
+            f.write(f"{u} {v}\n")
+
+
+def dump_dot(dot_path: Path, graph_edges: List[Set[Tuple[int, int]]]) -> None:
+    """Create a GraphViz dot file from a list of edge sets."""
+    dot_lines: List[str] = ["graph G {"]
+    for i, edges in enumerate(graph_edges):
+        dot_lines.append(f"    /* graph {i} */")
+        dot_lines.append(f"    subgraph cluster_{i} {{")
+        dot_lines.append("        label = \"{}\";".format(i))
+        # add nodes
+        nodes: Set[int] = {n for e in edges for n in e}
+        for n in sorted(nodes):
+            dot_lines.append(f"        {n};")
+        # add edges
+        for u, v in sorted(edges):
+            dot_lines.append(f"        {u} -- {v};")
+        dot_lines.append("    }")
+    dot_lines.append("}")
+    dot_path.parent.mkdir(parents=True, exist_ok=True)
+    dot_path.write_text("\n".join(dot_lines), encoding="utf-8")
+
+
+# ---------------------------------------------------------------------------
+
+def process_planetoid(
+    root: Path, name: str, out_path: Path, dot_file: Path | None
+) -> None:
+    """Single‑graph dataset (Planetoid)."""
+    from torch_geometric.datasets import Planetoid
+
+    ds = Planetoid(root=str(root), name=name)
+    edges = canonical_edges(ds[0].edge_index)
+
+    if out_path.is_dir():
+        out_file = out_path / "graph_000000.txt"
+    else:
+        out_file = out_path
+
+    write_edges(out_file, edges)
+
+    if dot_file:
+        dump_dot(dot_file, [edges])
+
+
+def process_tudataset(
+    root: Path, name: str, out_dir: Path, dot_file: Path | None
+) -> None:
+    """Multi‑graph dataset (TUDataset)."""
+    from torch_geometric.datasets import TUDataset
+
+    ds = TUDataset(root=str(root), name=name)
+    out_dir.mkdir(parents=True, exist_ok=True)
+
+    if dot_file:
+        all_edges: List[Set[Tuple[int, int]]] = []
+
+    for i, data in enumerate(ds):
+        edges = canonical_edges(data.edge_index)
+        out_file = out_dir / f"graph_{i:06d}.txt"
+        write_edges(out_file, edges)
+
+        if dot_file:
+            all_edges.append(edges)
+
+    if dot_file:
+        dump_dot(dot_file, all_edges)
+
+# ---------------------------------------------------------------------------
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__.strip(), formatter_class=argparse.RawTextHelpFormatter)
+    parser.add_argument("--data_root", default="./data", help="Root directory of PyG datasets")
+    parser.add_argument("dataset_name", help="PyG dataset name (e.g. Cora, WikipediaNetwork)")
+    parser.add_argument("edges_out", help="File (single‑graph) or directory (many‑graph) for edgelists")
+    parser.add_argument("--dot", dest="dot_file", type=str, default="", help="Optional .dot file for visualization")
+    args = parser.parse_args()
+
+    root = Path(args.data_root)
+    out_path = Path(args.edges_out)
+    dot_path = Path(args.dot_file) if args.dot_file else None
+
+    # Decide whether the dataset is Planetoid (single graph) or TUDataset (many)
+    try:
+        from torch_geometric.datasets import Planetoid
+
+        Planetoid(root=str(root), name=args.dataset_name)
+        dataset_type = "planetoid"
+    except Exception:  # pragma: no cover
+        from torch_geometric.datasets import TUDataset
+
+        TUDataset(root=str(root), name=args.dataset_name)
+        dataset_type = "tudataset"
+
+    if dataset_type == "planetoid":
+        process_planetoid(root, args.dataset_name, out_path, dot_path)
+    else:  # tudataset
+        if out_path.is_file():
+            raise ValueError("For multi‑graph datasets (--tudataset) the output must be a directory")
+        process_tudataset(root, args.dataset_name, out_path, dot_path)
+
+
+if __name__ == "__main__":
+    main()

src/.ipynb_checkpoints/export_reddit_edgelist-checkpoint.py

@@ -0,0 +1,41 @@
+# export_reddit_edgelist_canonical.py
+# Writes EACH undirected edge exactly once: "u v" with u < v (0-based), from PyG Reddit.
+# This halves the edge count relative to to_undirected and avoids duplication downstream.
+#
+# Usage:
+#   python export_reddit_edgelist_canonical.py --out reddit_edges.txt --root ./data/Reddit
+
+import argparse
+from pathlib import Path
+import torch
+from torch_geometric.datasets import Reddit
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--root", type=str, default="./data/Reddit")
+    ap.add_argument("--out", type=str, default="reddit_edges.txt")
+    args = ap.parse_args()
+
+    ds = Reddit(root=args.root); data = ds[0]
+    ei = data.edge_index  # directed; in this dataset it's effectively undirected
+    outp = Path(args.out); outp.parent.mkdir(parents=True, exist_ok=True)
+
+    # canonical pairs u<v; de-duplicate
+    seen = set()
+    with outp.open("w") as f:
+        E = ei.size(1)
+        for e in range(E):
+            u = int(ei[0, e]); v = int(ei[1, e])
+            if u == v:
+                continue
+            if u > v:
+                u, v = v, u
+            key = (u << 32) | v
+            if key in seen:
+                continue
+            seen.add(key)
+            f.write(f"{u} {v}\n")
+    print(f"Wrote {len(seen)} undirected edges to {outp} (nodes: {data.num_nodes})")
+
+if __name__ == "__main__":
+    main()

src/.ipynb_checkpoints/highlight_seeds_dot-checkpoint.py

@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+"""
+highlight_seeds_dot.py
+
+Highlight seed nodes (first cluster in a JSON seeds file) inside a GraphViz .dot file.
+
+Usage
+-----
+    python highlight_seeds_dot.py <input.dot> <seeds.json> <output.dot>
+
+Every node listed in seeds.json[clusters][0]["seed_nodes"] will be rendered
+[color="red", style="filled", fillcolor="red", fontcolor="white"]
+and all its incident edges colored red.
+The script prints *nothing* to stdout – it only writes the modified .dot file.
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import re
+from pathlib import Path
+
+
+def load_seeds(seeds_path: Path) -> set[int]:
+    """Return the seed-node indices from the first cluster."""
+    with seeds_path.open("r", encoding="utf-8") as f:
+        data = json.load(f)
+    return {n - 1 for n in data["clusters"][0]["seed_nodes"]}
+
+
+def highlight_dot(dot_path: Path, seeds: set[int], out_path: Path) -> None:
+    """Read dot_path, highlight seeds and their edges, write to out_path."""
+    content = dot_path.read_text(encoding="utf-8")
+
+    # 1) highlight seed nodes
+    def node_replace(m: re.Match) -> str:
+        node_id = int(m.group(1))
+        if node_id in seeds:
+            return f'{m.group(0).rstrip(";")} [color="red", style="filled", fillcolor="red", fontcolor="white"];'
+        return m.group(0)
+
+    content = re.sub(rf'^\s*({"|".join(map(str, seeds))})\s*;', node_replace, content, flags=re.MULTILINE)
+
+    # 2) highlight edges incident to seeds
+    def edge_replace(m: re.Match) -> str:
+        u, v = int(m.group(1)), int(m.group(2))
+        if u in seeds or v in seeds:
+            return f'{m.group(0).rstrip(";")} [color="red"];'
+        return m.group(0)
+
+    content = re.sub(rf'^\s*({"|".join(map(str, seeds))})\s*--\s*(\d+)\s*;', edge_replace, content, flags=re.MULTILINE)
+    content = re.sub(rf'^\s*(\d+)\s*--\s*({"|".join(map(str, seeds))})\s*;', edge_replace, content, flags=re.MULTILINE)
+
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+    out_path.write_text(content, encoding="utf-8")
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__.strip(), formatter_class=argparse.RawTextHelpFormatter)
+    parser.add_argument("input_dot", type=str, help="Original GraphViz .dot file")
+    parser.add_argument("seeds_json", type=str, help="Seeds JSON file")
+    parser.add_argument("output_dot", type=str, help="Path to write highlighted .dot file")
+    args = parser.parse_args()
+
+    seeds = load_seeds(Path(args.seeds_json))
+    highlight_dot(Path(args.input_dot), seeds, Path(args.output_dot))
+
+
+if __name__ == "__main__":
+    main()

src/1_build_lrmc_levels.py

@@ -0,0 +1,194 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+import argparse
+import glob
+import json
+import os
+import re
+import shutil
+import subprocess
+from collections import defaultdict
+from typing import Dict, List, Tuple, Iterable, Set
+
+def ensure_dir(p: str):
+    os.makedirs(p, exist_ok=True)
+
+def read_edgelist(path: str) -> Iterable[Tuple[int, int]]:
+    with open(path, 'r') as f:
+        for line in f:
+            s = line.strip()
+            if not s or s.startswith('#'):
+                continue
+            parts = s.split()
+            if len(parts) < 2:
+                continue
+            try:
+                u = int(parts[0]); v = int(parts[1])
+            except ValueError:
+                continue
+            if u == v:
+                continue
+            a, b = (u, v) if u < v else (v, u)
+            yield a, b
+
+def write_edgelist(path: str, edges: Iterable[Tuple[int, int]]):
+    with open(path, 'w') as f:
+        for u, v in edges:
+            f.write(f"{u} {v}\n")
+
+def parse_seeds(path: str) -> Tuple[Dict[int, int], List[int]]:
+    """
+    Return (node_to_cluster_index, sorted_cluster_ids).
+    The cluster indices are 0..C-1, sorted by cluster_id.
+    - On overlapping membership, choose the cluster with higher 'score', then smaller cluster_id.
+    """
+    with open(path, 'r') as f:
+        js = json.load(f)
+    clusters = js.get('clusters', [])
+    # Sort clusters by cluster_id for stable indexing
+    clusters_sorted = sorted(clusters, key=lambda c: c.get('cluster_id', 0))
+    cluster_id_list = [c.get('cluster_id', i) for i, c in enumerate(clusters_sorted)]
+    cluster_id_to_idx = {cid: i for i, cid in enumerate(cluster_id_list)}
+
+    # Build node->(best_cluster_idx, best_score, best_cid)
+    node_choice: Dict[int, Tuple[int, float, int]] = {}
+
+    for c in clusters_sorted:
+        cid = c.get('cluster_id', None)
+        if cid is None:
+            continue
+        idx = cluster_id_to_idx[cid]
+        members = c.get('members', [])
+        score = float(c.get('score', 0.0))
+        for u in members:
+            prev = node_choice.get(u, None)
+            if prev is None or (score > prev[1]) or (score == prev[1] and cid < prev[2]):
+                node_choice[u] = (idx, score, cid)
+
+    node_to_cluster = {u: idx for u, (idx, score, cid) in node_choice.items()}
+    return node_to_cluster, cluster_id_list
+
+def coarsen_edgelist(prev_edgelist: str, seeds_json: str, out_edgelist: str) -> int:
+    node_to_cluster, cluster_id_list = parse_seeds(seeds_json)
+    edges_set: Set[Tuple[int, int]] = set()
+    missing_nodes = 0
+    for u, v in read_edgelist(prev_edgelist):
+        cu = node_to_cluster.get(u, None)
+        cv = node_to_cluster.get(v, None)
+        if cu is None or cv is None:
+            # If a node isn't present in any cluster JSON, skip or count as missing
+            missing_nodes += 1
+            continue
+        if cu == cv:
+            continue
+        a, b = (cu, cv) if cu < cv else (cv, cu)
+        edges_set.add((a, b))
+
+    write_edgelist(out_edgelist, sorted(edges_set))
+    return missing_nodes
+
+def run_java(java_exec: str, class_name: str, edgelist_path: str, out_json_path: str,
+              epsilon: str, java_opts: List[str]) -> None:
+    cmd = [java_exec] + java_opts + [class_name, edgelist_path, out_json_path, epsilon]
+    print("[run]", " ".join(cmd))
+    subprocess.run(cmd, check=True)
+
+def build_single_graph_levels(args):
+    ensure_dir(args.out_dir)
+    # Stage 0 edgelist is the input; optionally copy for record
+    stage0_dir = os.path.join(args.out_dir, "stage0")
+    ensure_dir(stage0_dir)
+    e0_copy = os.path.join(stage0_dir, "edgelist_0.txt")
+    if args.copy_inputs:
+        shutil.copyfile(args.input_edgelist, e0_copy)
+
+    prev_edgelist = args.input_edgelist
+    for lvl in range(args.levels):
+        stage_dir = os.path.join(args.out_dir, f"stage{lvl}")
+        ensure_dir(stage_dir)
+        seeds_out = os.path.join(stage_dir, "seeds.json")
+        # Run Java to produce seeds at this level
+        run_java(args.java, args.class_name, prev_edgelist, seeds_out, args.epsilon, args.java_opts)
+
+        # Prepare next-level edgelist (unless last level)
+        if lvl < args.levels - 1:
+            next_stage_dir = os.path.join(args.out_dir, f"stage{lvl+1}")
+            ensure_dir(next_stage_dir)
+            next_edgelist = os.path.join(next_stage_dir, f"edgelist_{lvl+1}.txt")
+            missing = coarsen_edgelist(prev_edgelist, seeds_out, next_edgelist)
+            if missing > 0:
+                print(f"[warn] stage{lvl}: {missing} edges had nodes missing from seeds; skipped.")
+            prev_edgelist = next_edgelist
+
+def build_multigraph_levels(args):
+    ensure_dir(args.out_dir)
+    # Enumerate graph files
+    graph_files = sorted(glob.glob(os.path.join(args.graphs_dir, args.glob)))
+    if not graph_files:
+        raise SystemExit(f"No graph files found in {args.graphs_dir} with pattern {args.glob}")
+
+    pattern = re.compile(r'(.*?)(\d+)(\.\w+)$')  # capture numeric id
+    def graph_id_from_path(p: str) -> str:
+        base = os.path.basename(p)
+        m = pattern.match(base)
+        if m:
+            return m.group(2).zfill(6)  # zero-pad to 6 for consistency
+        # fallback: strip extension
+        stem = os.path.splitext(base)[0]
+        m2 = re.search(r'(\d+)$', stem)
+        return (m2.group(1).zfill(6) if m2 else stem)
+
+    # Stage 0: run Java for each graph
+    prev_stage_edgelists: Dict[str, str] = {}
+    for lvl in range(args.levels):
+        stage_dir = os.path.join(args.out_dir, f"stage{lvl}")
+        ensure_dir(stage_dir)
+
+        if lvl == 0:
+            for gpath in graph_files:
+                gid = graph_id_from_path(gpath)
+                seeds_out = os.path.join(stage_dir, f"graph_{gid}.json")
+                run_java(args.java, args.class_name, gpath, seeds_out, args.epsilon, args.java_opts)
+                prev_stage_edgelists[gid] = gpath
+        else:
+            # For each graph, coarsen previous edgelist using previous seeds, then run Java
+            for gpath in graph_files:
+                gid = graph_id_from_path(gpath)
+                prev_edgelist = prev_stage_edgelists[gid]
+                prev_seeds = os.path.join(args.out_dir, f"stage{lvl-1}", f"graph_{gid}.json")
+                next_edgelist = os.path.join(stage_dir, f"graph_{gid}.txt")
+                missing = coarsen_edgelist(prev_edgelist, prev_seeds, next_edgelist)
+                if missing > 0:
+                    print(f"[warn] stage{lvl-1} graph_{gid}: {missing} edges had nodes missing from seeds; skipped.")
+
+                seeds_out = os.path.join(stage_dir, f"graph_{gid}.json")
+                run_java(args.java, args.class_name, next_edgelist, seeds_out, args.epsilon, args.java_opts)
+                prev_stage_edgelists[gid] = next_edgelist
+
+
+def main():
+    ap = argparse.ArgumentParser(description="Build LRMC seeds across multiple levels by invoking the Java LRMC tool and coarsening between levels.")
+    mode = ap.add_mutually_exclusive_group(required=True)
+    mode.add_argument('--input_edgelist', type=str, help='Single-graph mode: path to original edgelist.txt')
+    mode.add_argument('--graphs_dir', type=str, help='Multi-graph mode: directory containing per-graph edgelist files (e.g., graph_000000.txt)')
+    ap.add_argument('--glob', type=str, default='graph_*.txt', help='Multi-graph mode: glob pattern for graph files (default: graph_*.txt)')
+    ap.add_argument('--out_dir', type=str, required=True, help='Output directory (stages will be created here)')
+    ap.add_argument('--levels', type=int, required=True, help='Number of levels to build (e.g., 3)')
+    # Java settings
+    ap.add_argument('--java', type=str, default='java', help='Java executable (default: java)')
+    ap.add_argument('--class_name', type=str, default='LRMCGenerateSingleCluster', help='Fully qualified Java class name')
+    ap.add_argument('--epsilon', type=str, default='1e6', help='Epsilon argument for the Java tool (default: 1e6)')
+    ap.add_argument('--java_opts', type=str, default='', help='Extra options for java (e.g., "-Xmx16g -cp my.jar")')
+    ap.add_argument('--copy_inputs', action='store_true', help='Copy original edgelist under stage0 for record (single-graph mode)')
+    args = ap.parse_args()
+
+    # Parse java_opts into a list if provided
+    args.java_opts = args.java_opts.split() if args.java_opts else []
+
+    if args.input_edgelist:
+        build_single_graph_levels(args)
+    else:
+        build_multigraph_levels(args)
+
+if __name__ == '__main__':
+    main()

src/2.1_lrmc_bilevel.py

@@ -0,0 +1,301 @@
+# lrmc_bilevel.py
+# Top-1 LRMC ablation: one-cluster pooling vs. plain GCN on Planetoid (e.g., Cora)
+# Requires: torch, torch_geometric, torch_scatter, torch_sparse
+#
+# Usage examples:
+#   python lrmc_bilevel.py --dataset Cora --seeds /path/to/lrmc_seeds.json --variant baseline
+#   python lrmc_bilevel.py --dataset Cora --seeds /path/to/lrmc_seeds.json --variant pool
+#
+# Notes:
+#  - We read your LRMC JSON, pick the single cluster with the highest 'score',
+#    assign it to cluster id 0, and make all other nodes singletons (1..K-1).
+#  - For --variant pool: Node-GCN -> pool (means) -> Cluster-GCN -> broadcast + skip -> Node-GCN -> classifier
+#  - For --variant baseline: Standard 2-layer GCN.
+#  - Keep flags like --self_loop_scale and --use_a2 if you want A+λI / A^2 augmentation.
+
+import argparse, json
+from pathlib import Path
+from typing import List, Tuple, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from torch import Tensor
+from torch_scatter import scatter_add, scatter_mean
+from torch_sparse import coalesce, spspmm
+
+from torch_geometric.datasets import Planetoid
+from torch_geometric.nn import GCNConv
+
+
+# ---------------------------
+# Utilities: edges and seeds
+# ---------------------------
+
+def add_scaled_self_loops(edge_index: Tensor,
+                          edge_weight: Optional[Tensor],
+                          num_nodes: int,
+                          scale: float = 1.0) -> Tuple[Tensor, Tensor]:
+    """Add self-loops with chosen weight (scale). If scale=0, return unchanged (and create weights if None)."""
+    if scale == 0.0:
+        if edge_weight is None:
+            edge_weight = torch.ones(edge_index.size(1), device=edge_index.device)
+        return edge_index, edge_weight
+    device = edge_index.device
+    self_loops = torch.arange(num_nodes, device=device)
+    self_index = torch.stack([self_loops, self_loops], dim=0)
+    self_weight = torch.full((num_nodes,), float(scale), device=device)
+    base_w = edge_weight if edge_weight is not None else torch.ones(edge_index.size(1), device=device)
+    ei = torch.cat([edge_index, self_index], dim=1)
+    ew = torch.cat([base_w, self_weight], dim=0)
+    ei, ew = coalesce(ei, ew, num_nodes, num_nodes, op='add')
+    return ei, ew
+
+
+def adjacency_power(edge_index: Tensor, num_nodes: int, k: int = 2) -> Tensor:
+    """
+    Compute (binary) k-th power adjacency using sparse matmul (torch_sparse.spspmm).
+    Here we use k=2. Returns coalesced edge_index without weights.
+    """
+    row, col = edge_index
+    val = torch.ones(row.numel(), device=edge_index.device)
+    Ai, Av = edge_index, val
+    # A^2
+    Ri, Rv = spspmm(Ai, Av, Ai, Av, num_nodes, num_nodes, num_nodes)
+    mask = Ri[0] != Ri[1]  # drop diagonal; add custom self-loops later if desired
+    Ri = Ri[:, mask]
+    Ri, _ = coalesce(Ri, torch.ones(Ri.size(1), device=edge_index.device), num_nodes, num_nodes, op='add')
+    return Ri
+
+
+def build_cluster_graph(edge_index: Tensor,
+                        num_nodes: int,
+                        node2cluster: Tensor,
+                        weight_per_edge: Optional[Tensor] = None,
+                        num_clusters: Optional[int] = None
+                        ) -> Tuple[Tensor, Tensor, int]:
+    """
+    Build cluster graph A_c = S^T A S with summed multiplicities as weights.
+    node2cluster: [N] long tensor mapping each node -> cluster id.
+    """
+    K = int(node2cluster.max().item()) + 1 if num_clusters is None else num_clusters
+    src, dst = edge_index
+    csrc = node2cluster[src]
+    cdst = node2cluster[dst]
+    edge_c = torch.stack([csrc, cdst], dim=0)
+    w = weight_per_edge if weight_per_edge is not None else torch.ones(edge_c.size(1), device=edge_c.device)
+    edge_c, w = coalesce(edge_c, w, K, K, op='add')  # sum multiplicities
+    return edge_c, w, K
+
+
+# -----
+# Seeds
+# -----
+
+def _pick_top1_cluster(obj: dict) -> List[int]:
+    """
+    From LRMC JSON with structure: {"clusters":[{"members":[...], "score":float, ...}, ...]}
+    choose the cluster with max (score, size) and return its members.
+    """
+    clusters = obj.get("clusters", [])
+    if not clusters:
+        return []
+    # choose by highest score, then by size (tiebreaker)
+    best = max(clusters, key=lambda c: (float(c.get("score", 0.0)), len(c.get("members", []))))
+    return list(best.get("members", []))
+
+
+def load_top1_assignment(seeds_json: str, n_nodes: int) -> Tuple[Tensor, Tensor]:
+    """
+    Create a hard assignment for top-1 LRMC cluster:
+      - cluster 0 = top-1 LRMC set
+      - nodes outside are singletons (1..K-1)
+    Returns:
+      node2cluster: [N] long
+      cluster_scores: [K,1] with 1.0 for top cluster, 0.0 for singletons
+    """
+    obj = json.loads(Path(seeds_json).read_text())
+    C_star = _pick_top1_cluster(obj)
+    C_star = torch.tensor(sorted(set(C_star)), dtype=torch.long)
+
+    node2cluster = torch.full((n_nodes,), -1, dtype=torch.long)
+    node2cluster[C_star] = 0
+    outside = torch.tensor(sorted(set(range(n_nodes)) - set(C_star.tolist())), dtype=torch.long)
+    if outside.numel() > 0:
+        node2cluster[outside] = torch.arange(1, 1 + outside.numel(), dtype=torch.long)
+    assert int(node2cluster.min()) >= 0, "All nodes must be assigned."
+
+    K = 1 + outside.numel()
+    cluster_scores = torch.zeros(K, 1, dtype=torch.float32)
+    if C_star.numel() > 0:
+        cluster_scores[0, 0] = 1.0  # emphasize the supercluster
+    return node2cluster, cluster_scores
+
+
+# --------------------------
+# Models (baseline + pooled)
+# --------------------------
+
+class GCN2(nn.Module):
+    """Plain 2-layer GCN baseline."""
+    def __init__(self, in_dim, hid, out_dim):
+        super().__init__()
+        self.conv1 = GCNConv(in_dim, hid)
+        self.conv2 = GCNConv(hid, out_dim)
+
+    def forward(self, x, edge_index):
+        x = F.relu(self.conv1(x, edge_index))
+        x = F.dropout(x, p=0.5, training=self.training)
+        x = self.conv2(x, edge_index)
+        return x
+
+
+class OneClusterPool(nn.Module):
+    """
+    Node-GCN -> pool to one-cluster + singletons -> Cluster-GCN -> broadcast + skip -> Node-GCN -> classifier
+    """
+    def __init__(self,
+                 in_dim: int,
+                 hid: int,
+                 out_dim: int,
+                 node2cluster: Tensor,
+                 edge_index_node: Tensor,
+                 num_nodes: int,
+                 self_loop_scale: float = 0.0,
+                 use_a2_for_clusters: bool = False):
+        super().__init__()
+        self.n2c = node2cluster.long()
+        self.K = int(self.n2c.max().item()) + 1
+
+        # Node graph (A + λI if desired)
+        ei_node = edge_index_node
+        ei_node, ew_node = add_scaled_self_loops(ei_node, None, num_nodes, scale=self_loop_scale)
+        self.register_buffer("edge_index_node", ei_node)
+        self.register_buffer("edge_weight_node", ew_node)
+
+        # Cluster graph from A or A^2
+        ei_for_c = adjacency_power(edge_index_node, num_nodes, k=2) if use_a2_for_clusters else edge_index_node
+        edge_index_c, edge_weight_c, K = build_cluster_graph(ei_for_c, num_nodes, self.n2c)
+        self.register_buffer("edge_index_c", edge_index_c)
+        self.register_buffer("edge_weight_c", edge_weight_c)
+        self.K = K
+
+        # Layers
+        self.gcn_node1 = GCNConv(in_dim, hid, add_self_loops=False, normalize=True)
+        self.gcn_cluster = GCNConv(hid, hid, add_self_loops=True,  normalize=True)
+        self.gcn_node2 = GCNConv(hid * 2, out_dim)  # on concatenated [h_node, h_broadcast]
+
+    def forward(self, x: Tensor, edge_index_node: Tensor) -> Tensor:
+        # Node GCN (uses stored weights)
+        h1 = F.relu(self.gcn_node1(x, self.edge_index_node, self.edge_weight_node))
+
+        # Pool to clusters: mean per cluster
+        z = scatter_mean(h1, self.n2c, dim=0, dim_size=self.K)  # [K, H]
+
+        # Cluster GCN
+        z2 = F.relu(self.gcn_cluster(z, self.edge_index_c, self.edge_weight_c))
+
+        # Broadcast back + skip concat
+        hb = z2[self.n2c]                    # [N, H]
+        hcat = torch.cat([h1, hb], dim=1)    # [N, 2H]
+
+        # Final node GCN head -> logits
+        out = self.gcn_node2(hcat, edge_index_node)
+        return out
+
+
+# -------------
+# Training glue
+# -------------
+
+@torch.no_grad()
+def accuracy(logits: Tensor, y: Tensor, mask: Tensor) -> float:
+    pred = logits[mask].argmax(dim=1)
+    return (pred == y[mask]).float().mean().item()
+
+
+def run_train_eval(model: nn.Module, data, epochs=200, lr=0.01, wd=5e-4):
+    opt = torch.optim.Adam(model.parameters(), lr=lr, weight_decay=wd)
+    best_val, best_state = 0.0, None
+    for ep in range(1, epochs + 1):
+        model.train()
+        opt.zero_grad(set_to_none=True)
+        logits = model(data.x, data.edge_index)
+        loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+        loss.backward(); opt.step()
+
+        # track best on val
+        model.eval()
+        logits = model(data.x, data.edge_index)
+        val_acc = accuracy(logits, data.y, data.val_mask)
+        if val_acc > best_val:
+            best_val, best_state = val_acc, {k: v.detach().clone() for k, v in model.state_dict().items()}
+        if ep % 20 == 0:
+            tr = accuracy(logits, data.y, data.train_mask)
+            te = accuracy(logits, data.y, data.test_mask)
+            print(f"[{ep:04d}] loss={loss.item():.4f}  train={tr:.3f}  val={val_acc:.3f}  test={te:.3f}")
+
+    # test @ best val
+    if best_state is not None:
+        model.load_state_dict(best_state)
+    model.eval()
+    logits = model(data.x, data.edge_index)
+    return {
+        "val": accuracy(logits, data.y, data.val_mask),
+        "test": accuracy(logits, data.y, data.test_mask)
+    }
+
+
+# -----------
+# Entrypoint
+# -----------
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--dataset", required=True, choices=["Cora", "Citeseer", "Pubmed"])
+    ap.add_argument("--seeds", required=True, help="Path to LRMC seeds JSON (single large graph).")
+    ap.add_argument("--variant", choices=["baseline", "pool"], default="pool",
+                    help="baseline=plain GCN; pool=top-1 LRMC one-cluster pooling")
+    ap.add_argument("--hidden", type=int, default=128)
+    ap.add_argument("--epochs", type=int, default=200)
+    ap.add_argument("--lr", type=float, default=0.01)
+    ap.add_argument("--wd", type=float, default=5e-4)
+    ap.add_argument("--dropout", type=float, default=0.5)  # used in baseline only
+    ap.add_argument("--self_loop_scale", type=float, default=0.0, help="λ for A+λI on node graph (0 disables)")
+    ap.add_argument("--use_a2", action="store_true", help="Use A^2 to build the cluster graph (recommended for pool)")
+    ap.add_argument("--seed", type=int, default=42)
+    args = ap.parse_args()
+
+    torch.manual_seed(args.seed)
+
+    # Load dataset
+    ds = Planetoid(root=f"./data/{args.dataset}", name=args.dataset)
+    data = ds[0]
+    in_dim, out_dim, n = ds.num_node_features, ds.num_classes, data.num_nodes
+
+    if args.variant == "baseline":
+        model = GCN2(in_dim, args.hidden, out_dim)
+        # use default add_self_loops=True behavior inside convs
+        res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+        print(f"Baseline GCN: val={res['val']:.4f}  test={res['test']:.4f}")
+        return
+
+    # Top-1 LRMC assignment
+    node2cluster, _ = load_top1_assignment(args.seeds, n)
+
+    # One-cluster pooled model
+    model = OneClusterPool(in_dim=in_dim,
+                           hid=args.hidden,
+                           out_dim=out_dim,
+                           node2cluster=node2cluster,
+                           edge_index_node=data.edge_index,
+                           num_nodes=n,
+                           self_loop_scale=args.self_loop_scale,
+                           use_a2_for_clusters=args.use_a2)
+    res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+    print(f"L-RMC (top-1 pool): val={res['val']:.4f}  test={res['test']:.4f}")
+
+
+if __name__ == "__main__":
+    main()

src/2.2_lrmc_bilevel.py

@@ -0,0 +1,325 @@
+# 2.1_lrmc_bilevel.py
+# Top-1 LRMC ablation with debug guards so seeds differences are visible.
+# Requires: torch, torch_geometric, torch_scatter, torch_sparse
+
+import argparse, json, hashlib
+from pathlib import Path
+from typing import List, Tuple, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch import Tensor
+
+from torch_scatter import scatter_mean
+from torch_sparse import coalesce, spspmm
+from torch_geometric.datasets import Planetoid
+from torch_geometric.nn import GCNConv
+
+
+# ---------------------------
+# Utilities: edges and seeds
+# ---------------------------
+
+def add_scaled_self_loops(edge_index: Tensor,
+                          edge_weight: Optional[Tensor],
+                          num_nodes: int,
+                          scale: float = 1.0) -> Tuple[Tensor, Tensor]:
+    if scale == 0.0:
+        if edge_weight is None:
+            edge_weight = torch.ones(edge_index.size(1), device=edge_index.device)
+        return edge_index, edge_weight
+    device = edge_index.device
+    self_loops = torch.arange(num_nodes, device=device)
+    self_index = torch.stack([self_loops, self_loops], dim=0)
+    self_weight = torch.full((num_nodes,), float(scale), device=device)
+    base_w = edge_weight if edge_weight is not None else torch.ones(edge_index.size(1), device=device)
+    ei = torch.cat([edge_index, self_index], dim=1)
+    ew = torch.cat([base_w, self_weight], dim=0)
+    ei, ew = coalesce(ei, ew, num_nodes, num_nodes, op='add')
+    return ei, ew
+
+
+def adjacency_power(edge_index: Tensor, num_nodes: int, k: int = 2) -> Tensor:
+    # A^2 using spspmm; return binary, coalesced, no self loops
+    row, col = edge_index
+    val = torch.ones(row.numel(), device=edge_index.device)
+    Ai, Av = edge_index, val
+    Ri, _ = spspmm(Ai, Av, Ai, Av, num_nodes, num_nodes, num_nodes)
+    mask = Ri[0] != Ri[1]
+    Ri = Ri[:, mask]
+    Ri, _ = coalesce(Ri, torch.ones(Ri.size(1), device=edge_index.device), num_nodes, num_nodes, op='add')
+    return Ri
+
+
+def build_cluster_graph(edge_index: Tensor,
+                        num_nodes: int,
+                        node2cluster: Tensor,
+                        weight_per_edge: Optional[Tensor] = None,
+                        num_clusters: Optional[int] = None
+                        ) -> Tuple[Tensor, Tensor, int]:
+    K = int(node2cluster.max().item()) + 1 if num_clusters is None else num_clusters
+    src, dst = edge_index
+    csrc = node2cluster[src]
+    cdst = node2cluster[dst]
+    edge_c = torch.stack([csrc, cdst], dim=0)
+    w = weight_per_edge if weight_per_edge is not None else torch.ones(edge_c.size(1), device=edge_c.device)
+    edge_c, w = coalesce(edge_c, w, K, K, op='add')
+    return edge_c, w, K
+
+
+# -----
+# Seeds
+# -----
+
+def _md5(path: Path) -> str:
+    h = hashlib.md5()
+    with path.open('rb') as f:
+        for chunk in iter(lambda: f.read(8192), b''):
+            h.update(chunk)
+    return h.hexdigest()
+
+
+def _extract_members(cluster_obj: dict) -> List[int]:
+    """
+    Try 'members' first, then 'seed_nodes'. Raise if neither works.
+    """
+    m = cluster_obj.get("members", None)
+    if isinstance(m, list) and len(m) > 0:
+        return list(dict.fromkeys(int(x) for x in m))  # dedupe/preserve order
+    m2 = cluster_obj.get("seed_nodes", None)
+    if isinstance(m2, list) and len(m2) > 0:
+        return list(dict.fromkeys(int(x) for x in m2))
+    # If both present but empty, return empty; caller will handle.
+    if isinstance(m, list) or isinstance(m2, list):
+        return []
+    raise KeyError("Cluster object has neither 'members' nor 'seed_nodes'.")
+
+
+def _pick_top1_cluster(obj: dict) -> List[int]:
+    """
+    From {"clusters":[{..., "score":float, "members" or "seed_nodes"}, ...]},
+    choose max by (score, size). Returns deduped member list.
+    """
+    clusters = obj.get("clusters", [])
+    if not isinstance(clusters, list) or len(clusters) == 0:
+        return []
+    def keyfun(c):
+        score = float(c.get("score", 0.0))
+        try:
+            mem = _extract_members(c)
+        except KeyError:
+            mem = []
+        return (score, len(mem))
+    best = max(clusters, key=keyfun)
+    try:
+        members = _extract_members(best)
+    except KeyError:
+        members = []
+    return sorted(set(int(x) for x in members))
+
+
+def load_top1_assignment(seeds_json: str, n_nodes: int, debug: bool = False) -> Tuple[Tensor, Tensor, dict]:
+    """
+    Hard assignment for top-1 LRMC cluster:
+      cluster 0 = top cluster; others are singletons.
+    Returns node2cluster[N], cluster_scores[K,1], and a small debug dict.
+    """
+    p = Path(seeds_json)
+    text = p.read_text(encoding='utf-8')
+    obj = json.loads(text)
+
+    C_star = _pick_top1_cluster(obj)
+    # if len(C_star) > 0 and max(C_star) == n_nodes:
+    # Looks 1-indexed (since max == N, not N-1) → shift down by 1
+    C_star = [u - 1 for u in C_star]
+    C_star = torch.tensor(C_star, dtype=torch.long)
+
+    # C_star = _pick_top1_cluster(obj)
+    # C_star = torch.tensor(C_star, dtype=torch.long)
+    node2cluster = torch.full((n_nodes,), -1, dtype=torch.long)
+
+    if C_star.numel() == 0:
+        # FAIL LOUDLY instead of silently falling back to identity
+        raise RuntimeError(
+            f"No members found for top-1 cluster in {seeds_json}. "
+            f"Expected 'members' or 'seed_nodes' to be non-empty."
+        )
+
+    node2cluster[C_star] = 0
+    outside = torch.tensor(sorted(set(range(n_nodes)) - set(C_star.tolist())), dtype=torch.long)
+    if outside.numel() > 0:
+        node2cluster[outside] = torch.arange(1, 1 + outside.numel(), dtype=torch.long)
+    assert int(node2cluster.min()) >= 0
+
+    K = 1 + outside.numel()
+    cluster_scores = torch.zeros(K, 1, dtype=torch.float32)
+    cluster_scores[0, 0] = 1.0
+
+    info = {
+        "json_md5": _md5(p),
+        "top_cluster_size": int(C_star.numel()),
+        "K": int(K),
+        "n_outside": int(outside.numel()),
+        "first_members": [int(x) for x in C_star[:10].tolist()],
+    }
+    if debug:
+        print(f"[LRMC] Loaded {seeds_json} (md5={info['json_md5']}) | "
+              f"top_size={info['top_cluster_size']} K={info['K']} outside={info['n_outside']} "
+              f"first10={info['first_members']}")
+    return node2cluster, cluster_scores, info
+
+
+# --------------------------
+# Models (baseline + pooled)
+# --------------------------
+
+class GCN2(nn.Module):
+    def __init__(self, in_dim, hid, out_dim, dropout=0.5):
+        super().__init__()
+        self.conv1 = GCNConv(in_dim, hid)
+        self.conv2 = GCNConv(hid, out_dim)
+        self.dropout = dropout
+    def forward(self, x, edge_index):
+        x = F.relu(self.conv1(x, edge_index))
+        x = F.dropout(x, p=self.dropout, training=self.training)
+        x = self.conv2(x, edge_index)
+        return x
+
+
+class OneClusterPool(nn.Module):
+    def __init__(self,
+                 in_dim: int,
+                 hid: int,
+                 out_dim: int,
+                 node2cluster: Tensor,
+                 edge_index_node: Tensor,
+                 num_nodes: int,
+                 self_loop_scale: float = 0.0,
+                 use_a2_for_clusters: bool = False,
+                 debug_header: str = ""):
+        super().__init__()
+        self.n2c = node2cluster.long()
+        self.K = int(self.n2c.max().item()) + 1
+
+        # Node graph (A + λI if desired)
+        ei_node = edge_index_node
+        ei_node, ew_node = add_scaled_self_loops(ei_node, None, num_nodes, scale=self_loop_scale)
+        self.register_buffer("edge_index_node", ei_node)
+        self.register_buffer("edge_weight_node", ew_node)
+
+        # Cluster graph
+        ei_for_c = adjacency_power(edge_index_node, num_nodes, k=2) if use_a2_for_clusters else edge_index_node
+        edge_index_c, edge_weight_c, K = build_cluster_graph(ei_for_c, num_nodes, self.n2c)
+        self.register_buffer("edge_index_c", edge_index_c)
+        self.register_buffer("edge_weight_c", edge_weight_c)
+        self.K = K
+
+        if debug_header:
+            print(f"[POOL] {debug_header} | cluster_edges={edge_index_c.size(1)} (K={K})")
+
+        # Layers
+        self.gcn_node1 = GCNConv(in_dim, hid, add_self_loops=False, normalize=True)
+        self.gcn_cluster = GCNConv(hid, hid, add_self_loops=True,  normalize=True)
+        self.gcn_node2 = GCNConv(hid * 2, out_dim)  # concat [h_node, h_broadcast]
+
+    def forward(self, x: Tensor, edge_index_node: Tensor) -> Tensor:
+        h1 = F.relu(self.gcn_node1(x, self.edge_index_node, self.edge_weight_node))
+        z  = scatter_mean(h1, self.n2c, dim=0, dim_size=self.K)          # [K, H]
+        z2 = F.relu(self.gcn_cluster(z, self.edge_index_c, self.edge_weight_c))
+        hb = z2[self.n2c]                                                # [N, H]
+        hcat = torch.cat([h1, hb], dim=1)                                # [N, 2H]
+        out = self.gcn_node2(hcat, edge_index_node)
+        return out
+
+
+# -------------
+# Training glue
+# -------------
+
+@torch.no_grad()
+def accuracy(logits: Tensor, y: Tensor, mask: Tensor) -> float:
+    pred = logits[mask].argmax(dim=1)
+    return (pred == y[mask]).float().mean().item()
+
+
+def run_train_eval(model: nn.Module, data, epochs=200, lr=0.01, wd=5e-4):
+    opt = torch.optim.Adam(model.parameters(), lr=lr, weight_decay=wd)
+    best_val, best_state = 0.0, None
+    for ep in range(1, epochs + 1):
+        model.train()
+        opt.zero_grad(set_to_none=True)
+        logits = model(data.x, data.edge_index)
+        loss = F.cross_entropy(logits[data.train_mask], data.y[data.train_mask])
+        loss.backward(); opt.step()
+
+        model.eval()
+        logits = model(data.x, data.edge_index)
+        val_acc = accuracy(logits, data.y, data.val_mask)
+        if val_acc > best_val:
+            best_val, best_state = val_acc, {k: v.detach().clone() for k, v in model.state_dict().items()}
+        if ep % 20 == 0:
+            tr = accuracy(logits, data.y, data.train_mask)
+            te = accuracy(logits, data.y, data.test_mask)
+            print(f"[{ep:04d}] loss={loss.item():.4f}  train={tr:.3f}  val={val_acc:.3f}  test={te:.3f}")
+
+    if best_state is not None:
+        model.load_state_dict(best_state)
+    model.eval()
+    logits = model(data.x, data.edge_index)
+    return {"val": accuracy(logits, data.y, data.val_mask),
+            "test": accuracy(logits, data.y, data.test_mask)}
+
+
+# -----------
+# Entrypoint
+# -----------
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--dataset", required=True, choices=["Cora", "Citeseer", "Pubmed"])
+    ap.add_argument("--seeds", required=True, help="Path to LRMC seeds JSON (single large graph).")
+    ap.add_argument("--variant", choices=["baseline", "pool"], default="pool")
+    ap.add_argument("--hidden", type=int, default=128)
+    ap.add_argument("--epochs", type=int, default=200)
+    ap.add_argument("--lr", type=float, default=0.01)
+    ap.add_argument("--wd", type=float, default=5e-4)
+    ap.add_argument("--dropout", type=float, default=0.5)  # baseline only
+    ap.add_argument("--self_loop_scale", type=float, default=0.0)
+    ap.add_argument("--use_a2", action="store_true", help="Use A^2 for the cluster graph.")
+    ap.add_argument("--seed", type=int, default=42)
+    ap.add_argument("--debug", action="store_true", help="Print seeds md5, cluster size, K, etc.")
+    args = ap.parse_args()
+
+    torch.manual_seed(args.seed)
+
+    ds = Planetoid(root=f"./data/{args.dataset}", name=args.dataset)
+    data = ds[0]
+    in_dim, out_dim, n = ds.num_node_features, ds.num_classes, data.num_nodes
+
+    if args.variant == "baseline":
+        model = GCN2(in_dim, args.hidden, out_dim, dropout=args.dropout)
+        res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+        print(f"Baseline GCN: val={res['val']:.4f}  test={res['test']:.4f}")
+        return
+
+    # pool variant
+    node2cluster, _, info = load_top1_assignment(args.seeds, n, debug=args.debug)
+    dbg_header = f"seeds_md5={info['json_md5']} top_size={info['top_cluster_size']} K={info['K']}"
+
+    model = OneClusterPool(in_dim=in_dim,
+                           hid=args.hidden,
+                           out_dim=out_dim,
+                           node2cluster=node2cluster,
+                           edge_index_node=data.edge_index,
+                           num_nodes=n,
+                           self_loop_scale=args.self_loop_scale,
+                           use_a2_for_clusters=args.use_a2,
+                           debug_header=dbg_header)
+
+    res = run_train_eval(model, data, epochs=args.epochs, lr=args.lr, wd=args.wd)
+    print(f"L-RMC (top-1 pool): val={res['val']:.4f}  test={res['test']:.4f}")
+
+
+if __name__ == "__main__":
+    main()