======================================================= RESTART [10-13 06:45:05] ======================================================= ======================================================= RESTART [10-13 06:45:05] ======================================================= ======================================================= RESTART [10-13 06:45:05] ======================================================= wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information. wandb: Currently logged in as: kaiqiu. Use `wandb login --relogin` to force relogin wandb: Tracking run with wandb version 0.18.3 wandb: Run data is saved locally in /home/user/VAR/wandb/run-20241012_154627-ldd3ziku wandb: Run `wandb offline` to turn off syncing. wandb: Syncing run laced-star-189 wandb: View project at https://wandb.ai/kaiqiu/VAR wandb: View run at https://wandb.ai/kaiqiu/VAR/runs/ldd3ziku /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): AUTOTUNE addmm(6840x1088, 6840x64, 64x1088) triton_mm_9 0.0326 ms 100.0% triton_mm_13 0.0331 ms 98.5% triton_mm_16 0.0338 ms 96.4% triton_mm_8 0.0342 ms 95.1% triton_mm_6 0.0352 ms 92.5% triton_mm_5 0.0356 ms 91.5% triton_mm_7 0.0375 ms 86.9% triton_mm_10 0.0376 ms 86.7% triton_mm_14 0.0377 ms 86.3% triton_mm_11 0.0386 ms 84.4% SingleProcess AUTOTUNE benchmarking takes 2.4239 seconds and 0.0132 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_35 0.0214 ms 100.0% triton_mm_27 0.0230 ms 93.0% triton_mm_31 0.0232 ms 92.3% triton_mm_20 0.0235 ms 90.9% triton_mm_23 0.0237 ms 90.0% triton_mm_30 0.0238 ms 89.9% triton_mm_26 0.0239 ms 89.4% bias_addmm 0.0241 ms 88.8% triton_mm_21 0.0244 ms 87.7% triton_mm_22 0.0257 ms 83.1% SingleProcess AUTOTUNE benchmarking takes 2.2050 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(6864x3264, 6864x1088, 1088x3264) triton_mm_52 0.2210 ms 100.0% triton_mm_53 0.2267 ms 97.5% triton_mm_54 0.2539 ms 87.0% triton_mm_47 0.2626 ms 84.1% triton_mm_49 0.2712 ms 81.5% triton_mm_45 0.2752 ms 80.3% bias_addmm 0.3002 ms 73.6% triton_mm_46 0.3053 ms 72.4% triton_mm_50 0.3088 ms 71.6% addmm 0.3230 ms 68.4% SingleProcess AUTOTUNE benchmarking takes 2.6487 seconds and 0.0139 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x1088, 1088x1088) triton_mm_72 0.0902 ms 100.0% triton_mm_71 0.0917 ms 98.4% bias_addmm 0.0972 ms 92.9% triton_mm_66 0.0981 ms 92.0% triton_mm_68 0.0982 ms 91.9% triton_mm_73 0.0994 ms 90.8% triton_mm_64 0.1015 ms 88.9% addmm 0.1101 ms 82.0% triton_mm_65 0.1116 ms 80.8% triton_mm_69 0.1122 ms 80.4% SingleProcess AUTOTUNE benchmarking takes 2.5137 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_90 0.2908 ms 100.0% triton_mm_91 0.2949 ms 98.6% bias_addmm 0.3255 ms 89.3% triton_mm_92 0.3340 ms 87.1% triton_mm_85 0.3430 ms 84.8% triton_mm_83 0.3614 ms 80.5% triton_mm_87 0.3647 ms 79.7% addmm 0.3682 ms 79.0% triton_mm_84 0.4000 ms 72.7% triton_mm_88 0.4116 ms 70.6% SingleProcess AUTOTUNE benchmarking takes 2.6512 seconds and 0.0096 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2917 ms 100.0% triton_mm_110 0.3078 ms 94.8% triton_mm_109 0.3163 ms 92.2% triton_mm_111 0.3264 ms 89.4% addmm 0.3276 ms 89.0% triton_mm_104 0.3559 ms 82.0% triton_mm_106 0.3596 ms 81.1% triton_mm_102 0.3695 ms 78.9% triton_mm_105 0.3709 ms 78.6% triton_mm_107 0.4092 ms 71.3% SingleProcess AUTOTUNE benchmarking takes 2.6552 seconds and 0.0008 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_183 0.2925 ms 100.0% triton_mm_184 0.2961 ms 98.8% bias_addmm 0.3252 ms 89.9% triton_mm_185 0.3356 ms 87.1% triton_mm_178 0.3441 ms 85.0% triton_mm_176 0.3621 ms 80.8% triton_mm_180 0.3635 ms 80.5% triton_mm_177 0.4013 ms 72.9% triton_mm_181 0.4121 ms 71.0% addmm 0.4173 ms 70.1% SingleProcess AUTOTUNE benchmarking takes 2.6479 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2930 ms 100.0% triton_mm_203 0.3082 ms 95.1% triton_mm_202 0.3167 ms 92.5% triton_mm_204 0.3263 ms 89.8% addmm 0.3295 ms 88.9% triton_mm_197 0.3574 ms 82.0% triton_mm_199 0.3593 ms 81.6% triton_mm_195 0.3689 ms 79.4% triton_mm_198 0.3714 ms 78.9% triton_mm_200 0.4095 ms 71.5% SingleProcess AUTOTUNE benchmarking takes 2.6434 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x1088, 1088x2176) triton_mm_1604 0.0145 ms 100.0% triton_mm_1612 0.0149 ms 97.0% triton_mm_1608 0.0151 ms 95.8% triton_mm_1616 0.0171 ms 84.4% triton_mm_1607 0.0175 ms 82.8% triton_mm_1602 0.0175 ms 82.5% triton_mm_1601 0.0178 ms 81.3% triton_mm_1603 0.0178 ms 81.1% mm 0.0181 ms 79.9% triton_mm_1611 0.0190 ms 76.0% SingleProcess AUTOTUNE benchmarking takes 7.4440 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x32768, 6864x1088, 1088x32768) triton_mm_1634 2.2287 ms 100.0% bias_addmm 2.2327 ms 99.8% triton_mm_1633 2.2444 ms 99.3% triton_mm_1635 2.5118 ms 88.7% triton_mm_1628 2.6554 ms 83.9% triton_mm_1630 2.7598 ms 80.8% triton_mm_1626 2.8526 ms 78.1% addmm 2.9711 ms 75.0% triton_mm_1627 3.1106 ms 71.7% triton_mm_1631 3.1190 ms 71.5% SingleProcess AUTOTUNE benchmarking takes 11.6023 seconds and 0.0000 seconds precompiling AUTOTUNE mm(32768x6864, 6864x1088) triton_mm_1671 2.0725 ms 100.0% triton_mm_1672 2.1486 ms 96.5% mm 2.2393 ms 92.6% triton_mm_1673 2.2509 ms 92.1% triton_mm_1668 2.6100 ms 79.4% triton_mm_1664 2.6739 ms 77.5% triton_mm_1666 2.6774 ms 77.4% triton_mm_1669 2.7849 ms 74.4% triton_mm_1667 2.8139 ms 73.7% triton_mm_1665 2.9262 ms 70.8% SingleProcess AUTOTUNE benchmarking takes 9.9401 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2910 ms 100.0% triton_mm_1744 0.2944 ms 98.8% triton_mm_1743 0.2962 ms 98.2% triton_mm_1745 0.2997 ms 97.1% triton_mm_1738 0.3510 ms 82.9% triton_mm_1736 0.3546 ms 82.1% triton_mm_1740 0.3624 ms 80.3% triton_mm_1739 0.3780 ms 77.0% triton_mm_1741 0.3975 ms 73.2% triton_mm_1737 0.4022 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 9.9351 seconds and 0.0000 seconds precompiling AUTOTUNE mm(2176x24, 24x1088) triton_mm_1696 0.0113 ms 100.0% triton_mm_1697 0.0116 ms 97.5% triton_mm_1698 0.0117 ms 95.9% triton_mm_1691 0.0122 ms 92.6% triton_mm_1699 0.0125 ms 90.3% triton_mm_1700 0.0125 ms 90.0% triton_mm_1702 0.0125 ms 90.0% triton_mm_1693 0.0125 ms 89.8% triton_mm_1705 0.0127 ms 88.9% triton_mm_1701 0.0127 ms 88.7% SingleProcess AUTOTUNE benchmarking takes 5.9980 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x2176, 2176x1088) triton_mm_1678 0.0157 ms 100.0% triton_mm_1682 0.0164 ms 95.3% mm 0.0173 ms 90.7% triton_mm_1686 0.0206 ms 76.0% triton_mm_1690 0.0224 ms 70.0% triton_mm_1677 0.0234 ms 67.1% triton_mm_1676 0.0238 ms 65.9% triton_mm_1681 0.0262 ms 59.9% triton_mm_1675 0.0263 ms 59.6% triton_mm_1685 0.0270 ms 58.1% SingleProcess AUTOTUNE benchmarking takes 7.6826 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x32768, 32768x1088) mm 2.0900 ms 100.0% triton_mm_1652 2.2950 ms 91.1% triton_mm_1653 2.3227 ms 90.0% triton_mm_1654 2.3739 ms 88.0% triton_mm_1645 2.7994 ms 74.7% triton_mm_1649 2.8409 ms 73.6% triton_mm_1647 2.8481 ms 73.4% triton_mm_1648 2.9165 ms 71.7% triton_mm_1646 2.9770 ms 70.2% triton_mm_1650 3.0166 ms 69.3% SingleProcess AUTOTUNE benchmarking takes 2.9390 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1724 0.2757 ms 100.0% mm 0.2808 ms 98.2% triton_mm_1725 0.2816 ms 97.9% triton_mm_1726 0.3168 ms 87.0% triton_mm_1717 0.3418 ms 80.6% triton_mm_1719 0.3437 ms 80.2% triton_mm_1721 0.3482 ms 79.2% triton_mm_1718 0.3660 ms 75.3% triton_mm_1722 0.3909 ms 70.5% triton_mm_1720 0.4268 ms 64.6% SingleProcess AUTOTUNE benchmarking takes 2.5267 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) triton_mm_1839 0.2231 ms 100.0% triton_mm_1838 0.2287 ms 97.5% triton_mm_1835 0.2562 ms 87.1% mm 0.2575 ms 86.6% triton_mm_1833 0.2601 ms 85.8% triton_mm_1831 0.2607 ms 85.5% triton_mm_1840 0.2820 ms 79.1% triton_mm_1834 0.2852 ms 78.2% triton_mm_1832 0.2888 ms 77.2% triton_mm_1836 0.2948 ms 75.7% SingleProcess AUTOTUNE benchmarking takes 9.8823 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1781 0.2912 ms 100.0% mm 0.2924 ms 99.6% triton_mm_1782 0.2970 ms 98.0% triton_mm_1783 0.3035 ms 96.0% triton_mm_1778 0.3493 ms 83.3% triton_mm_1774 0.3569 ms 81.6% triton_mm_1776 0.3602 ms 80.8% triton_mm_1777 0.3874 ms 75.2% triton_mm_1779 0.3952 ms 73.7% triton_mm_1775 0.4023 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5449 seconds and 0.0009 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2256 ms 100.0% triton_mm_1857 0.2794 ms 80.7% triton_mm_1858 0.2861 ms 78.8% triton_mm_1852 0.2863 ms 78.8% triton_mm_1854 0.2869 ms 78.6% triton_mm_1850 0.2891 ms 78.0% triton_mm_1859 0.2908 ms 77.6% triton_mm_1853 0.3148 ms 71.7% triton_mm_1855 0.3277 ms 68.8% triton_mm_1851 0.3323 ms 67.9% SingleProcess AUTOTUNE benchmarking takes 2.5178 seconds and 0.0009 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0978 ms 100.0% triton_mm_1821 0.0995 ms 98.3% triton_mm_1814 0.1044 ms 93.7% triton_mm_1820 0.1076 ms 90.9% triton_mm_1811 0.1189 ms 82.3% triton_mm_1816 0.1199 ms 81.6% triton_mm_1812 0.1199 ms 81.6% triton_mm_1819 0.1228 ms 79.6% triton_mm_1815 0.1273 ms 76.8% triton_mm_1817 0.1313 ms 74.5% SingleProcess AUTOTUNE benchmarking takes 9.7807 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3010 ms 100.0% triton_mm_1762 0.3053 ms 98.6% triton_mm_1763 0.3119 ms 96.5% triton_mm_1764 0.3225 ms 93.3% triton_mm_1759 0.3469 ms 86.8% triton_mm_1757 0.3528 ms 85.3% triton_mm_1755 0.3534 ms 85.2% triton_mm_1758 0.3682 ms 81.7% triton_mm_1756 0.3908 ms 77.0% triton_mm_1760 0.3979 ms 75.6% SingleProcess AUTOTUNE benchmarking takes 2.5355 seconds and 0.0011 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1801 0.0830 ms 100.0% mm 0.0839 ms 98.9% triton_mm_1800 0.0846 ms 98.1% triton_mm_1797 0.0927 ms 89.6% triton_mm_1793 0.0940 ms 88.3% triton_mm_1802 0.0946 ms 87.8% triton_mm_1795 0.0959 ms 86.6% triton_mm_1794 0.1028 ms 80.8% triton_mm_1798 0.1048 ms 79.2% triton_mm_1796 0.1231 ms 67.5% SingleProcess AUTOTUNE benchmarking takes 2.4177 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1910 0.2757 ms 100.0% mm 0.2766 ms 99.7% triton_mm_1911 0.2798 ms 98.5% triton_mm_1912 0.3163 ms 87.2% triton_mm_1903 0.3414 ms 80.8% triton_mm_1905 0.3433 ms 80.3% triton_mm_1907 0.3486 ms 79.1% triton_mm_1904 0.3656 ms 75.4% triton_mm_1908 0.3875 ms 71.2% triton_mm_1906 0.4263 ms 64.7% SingleProcess AUTOTUNE benchmarking takes 2.5238 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.2998 ms 100.0% triton_mm_1948 0.3057 ms 98.1% triton_mm_1949 0.3131 ms 95.8% triton_mm_1950 0.3234 ms 92.7% triton_mm_1945 0.3467 ms 86.5% triton_mm_1943 0.3522 ms 85.1% triton_mm_1941 0.3538 ms 84.7% triton_mm_1944 0.3704 ms 80.9% triton_mm_1942 0.3908 ms 76.7% triton_mm_1946 0.3984 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.5320 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_1883 0.0173 ms 100.0% triton_mm_1886 0.0173 ms 99.8% triton_mm_1885 0.0183 ms 94.7% triton_mm_1890 0.0183 ms 94.4% triton_mm_1889 0.0184 ms 94.3% triton_mm_1884 0.0184 ms 93.9% triton_mm_1892 0.0186 ms 93.3% triton_mm_1888 0.0186 ms 93.0% triton_mm_1882 0.0188 ms 92.2% triton_mm_1893 0.0193 ms 89.6% SingleProcess AUTOTUNE benchmarking takes 5.8895 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2916 ms 100.0% triton_mm_1930 0.2992 ms 97.5% triton_mm_1929 0.3002 ms 97.2% triton_mm_1931 0.3049 ms 95.7% triton_mm_1924 0.3562 ms 81.9% triton_mm_1922 0.3589 ms 81.3% triton_mm_1926 0.3676 ms 79.3% triton_mm_1925 0.3832 ms 76.1% triton_mm_1927 0.4046 ms 72.1% triton_mm_1923 0.4075 ms 71.6% SingleProcess AUTOTUNE benchmarking takes 2.5267 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1967 0.2930 ms 100.0% mm 0.2944 ms 99.5% triton_mm_1968 0.2982 ms 98.3% triton_mm_1969 0.3036 ms 96.5% triton_mm_1964 0.3510 ms 83.5% triton_mm_1960 0.3574 ms 82.0% triton_mm_1962 0.3619 ms 81.0% triton_mm_1963 0.3902 ms 75.1% triton_mm_1965 0.3960 ms 74.0% triton_mm_1961 0.4044 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5242 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0271 ms 100.0% triton_mm_1864 0.0280 ms 96.8% triton_mm_1868 0.0308 ms 87.9% triton_mm_1872 0.0400 ms 67.7% triton_mm_1876 0.0517 ms 52.4% triton_mm_1863 0.0525 ms 51.6% triton_mm_1862 0.0559 ms 48.4% triton_mm_1867 0.0609 ms 44.4% triton_mm_1871 0.0626 ms 43.3% triton_mm_1861 0.0655 ms 41.3% SingleProcess AUTOTUNE benchmarking takes 6.5927 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2252 ms 100.0% triton_mm_2043 0.2794 ms 80.6% triton_mm_2044 0.2862 ms 78.7% triton_mm_2038 0.2868 ms 78.5% triton_mm_2040 0.2881 ms 78.2% triton_mm_2036 0.2892 ms 77.9% triton_mm_2045 0.2902 ms 77.6% triton_mm_2039 0.3162 ms 71.2% triton_mm_2041 0.3274 ms 68.8% triton_mm_2037 0.3335 ms 67.5% SingleProcess AUTOTUNE benchmarking takes 2.5081 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0972 ms 100.0% triton_mm_2007 0.0993 ms 97.9% triton_mm_2000 0.1046 ms 93.0% triton_mm_2006 0.1076 ms 90.4% triton_mm_1997 0.1183 ms 82.2% triton_mm_1998 0.1198 ms 81.2% triton_mm_2002 0.1207 ms 80.6% triton_mm_2005 0.1218 ms 79.8% triton_mm_2001 0.1262 ms 77.0% triton_mm_2003 0.1301 ms 74.7% SingleProcess AUTOTUNE benchmarking takes 2.4052 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1987 0.0831 ms 100.0% mm 0.0841 ms 98.9% triton_mm_1986 0.0854 ms 97.3% triton_mm_1983 0.0918 ms 90.5% triton_mm_1979 0.0946 ms 87.9% triton_mm_1988 0.0947 ms 87.8% triton_mm_1981 0.0958 ms 86.8% triton_mm_1980 0.1035 ms 80.3% triton_mm_1984 0.1054 ms 78.9% triton_mm_1982 0.1221 ms 68.1% SingleProcess AUTOTUNE benchmarking takes 2.4134 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2263 ms 100.0% triton_mm_2025 0.2263 ms 100.0% triton_mm_2024 0.2320 ms 97.5% triton_mm_2026 0.2459 ms 92.0% triton_mm_2021 0.2609 ms 86.7% triton_mm_2017 0.2653 ms 85.3% triton_mm_2019 0.2653 ms 85.3% triton_mm_2020 0.2927 ms 77.3% triton_mm_2018 0.2947 ms 76.8% triton_mm_2022 0.3018 ms 75.0% SingleProcess AUTOTUNE benchmarking takes 2.5130 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2096 0.2764 ms 100.0% mm 0.2777 ms 99.6% triton_mm_2097 0.2814 ms 98.2% triton_mm_2098 0.3175 ms 87.1% triton_mm_2089 0.3435 ms 80.5% triton_mm_2091 0.3478 ms 79.5% triton_mm_2093 0.3501 ms 79.0% triton_mm_2090 0.3683 ms 75.1% triton_mm_2094 0.3882 ms 71.2% triton_mm_2092 0.4290 ms 64.4% SingleProcess AUTOTUNE benchmarking takes 2.5257 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3023 ms 100.0% triton_mm_2134 0.3060 ms 98.8% triton_mm_2135 0.3140 ms 96.3% triton_mm_2136 0.3236 ms 93.4% triton_mm_2131 0.3482 ms 86.8% triton_mm_2129 0.3518 ms 85.9% triton_mm_2127 0.3546 ms 85.2% triton_mm_2130 0.3697 ms 81.8% triton_mm_2128 0.3899 ms 77.5% triton_mm_2132 0.3984 ms 75.9% SingleProcess AUTOTUNE benchmarking takes 2.5360 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2075 0.0172 ms 100.0% triton_mm_2069 0.0173 ms 99.5% triton_mm_2072 0.0173 ms 99.5% triton_mm_2074 0.0173 ms 99.4% triton_mm_2073 0.0183 ms 94.1% triton_mm_2070 0.0184 ms 93.5% triton_mm_2071 0.0185 ms 93.1% triton_mm_2078 0.0186 ms 92.8% triton_mm_2068 0.0188 ms 91.7% triton_mm_2077 0.0193 ms 89.2% SingleProcess AUTOTUNE benchmarking takes 2.1285 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2937 ms 100.0% triton_mm_2116 0.3002 ms 97.8% triton_mm_2115 0.3009 ms 97.6% triton_mm_2117 0.3042 ms 96.6% triton_mm_2110 0.3569 ms 82.3% triton_mm_2108 0.3602 ms 81.5% triton_mm_2112 0.3692 ms 79.5% triton_mm_2111 0.3850 ms 76.3% triton_mm_2113 0.4052 ms 72.5% triton_mm_2109 0.4097 ms 71.7% SingleProcess AUTOTUNE benchmarking takes 2.5376 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_2153 0.2944 ms 100.0% mm 0.2951 ms 99.7% triton_mm_2154 0.2990 ms 98.5% triton_mm_2155 0.3057 ms 96.3% triton_mm_2150 0.3508 ms 83.9% triton_mm_2146 0.3602 ms 81.7% triton_mm_2148 0.3609 ms 81.6% triton_mm_2149 0.3897 ms 75.5% triton_mm_2151 0.3988 ms 73.8% triton_mm_2147 0.4060 ms 72.5% SingleProcess AUTOTUNE benchmarking takes 2.5386 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2050 0.0269 ms 100.0% mm 0.0271 ms 99.4% triton_mm_2054 0.0320 ms 84.0% triton_mm_2058 0.0406 ms 66.3% triton_mm_2062 0.0518 ms 52.0% triton_mm_2049 0.0525 ms 51.2% triton_mm_2048 0.0568 ms 47.4% triton_mm_2053 0.0611 ms 44.1% triton_mm_2057 0.0627 ms 42.9% triton_mm_2047 0.0664 ms 40.6% SingleProcess AUTOTUNE benchmarking takes 2.0769 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2610 ms 100.0% triton_mm_2229 0.2798 ms 93.3% triton_mm_2224 0.2859 ms 91.3% triton_mm_2230 0.2869 ms 91.0% triton_mm_2222 0.2885 ms 90.5% triton_mm_2226 0.2887 ms 90.4% triton_mm_2231 0.2918 ms 89.4% triton_mm_2225 0.3143 ms 83.0% triton_mm_2227 0.3291 ms 79.3% triton_mm_2223 0.3326 ms 78.5% SingleProcess AUTOTUNE benchmarking takes 2.5128 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0972 ms 100.0% triton_mm_2193 0.0984 ms 98.7% triton_mm_2186 0.1047 ms 92.8% triton_mm_2192 0.1076 ms 90.3% triton_mm_2183 0.1185 ms 82.0% triton_mm_2184 0.1199 ms 81.1% triton_mm_2188 0.1207 ms 80.5% triton_mm_2191 0.1218 ms 79.8% triton_mm_2187 0.1263 ms 76.9% triton_mm_2189 0.1302 ms 74.6% SingleProcess AUTOTUNE benchmarking takes 2.4239 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2173 0.0833 ms 100.0% mm 0.0844 ms 98.7% triton_mm_2172 0.0849 ms 98.1% triton_mm_2169 0.0932 ms 89.4% triton_mm_2174 0.0943 ms 88.3% triton_mm_2167 0.0952 ms 87.5% triton_mm_2165 0.0955 ms 87.2% triton_mm_2166 0.1028 ms 81.0% triton_mm_2170 0.1051 ms 79.3% triton_mm_2168 0.1221 ms 68.2% SingleProcess AUTOTUNE benchmarking takes 2.4239 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2258 ms 100.0% triton_mm_2211 0.2276 ms 99.2% triton_mm_2210 0.2333 ms 96.8% triton_mm_2212 0.2454 ms 92.0% triton_mm_2207 0.2592 ms 87.1% triton_mm_2205 0.2665 ms 84.7% triton_mm_2203 0.2667 ms 84.7% triton_mm_2204 0.2938 ms 76.8% triton_mm_2206 0.2940 ms 76.8% triton_mm_2208 0.3002 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.5068 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0270 ms 100.0% triton_mm_2236 0.0281 ms 95.9% triton_mm_2240 0.0308 ms 87.5% triton_mm_2244 0.0401 ms 67.3% triton_mm_2235 0.0517 ms 52.2% triton_mm_2248 0.0522 ms 51.7% triton_mm_2234 0.0569 ms 47.4% triton_mm_2239 0.0609 ms 44.3% triton_mm_2243 0.0626 ms 43.1% triton_mm_2233 0.0664 ms 40.6% SingleProcess AUTOTUNE benchmarking takes 2.0757 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0976 ms 100.0% triton_mm_2379 0.0984 ms 99.2% triton_mm_2372 0.1046 ms 93.3% triton_mm_2378 0.1076 ms 90.7% triton_mm_2369 0.1181 ms 82.6% triton_mm_2370 0.1190 ms 82.0% triton_mm_2374 0.1207 ms 80.9% triton_mm_2377 0.1218 ms 80.1% triton_mm_2373 0.1262 ms 77.4% triton_mm_2375 0.1301 ms 75.0% SingleProcess AUTOTUNE benchmarking takes 2.4232 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2359 0.0832 ms 100.0% mm 0.0847 ms 98.2% triton_mm_2358 0.0847 ms 98.1% triton_mm_2355 0.0926 ms 89.8% triton_mm_2360 0.0943 ms 88.2% triton_mm_2351 0.0951 ms 87.4% triton_mm_2353 0.0951 ms 87.4% triton_mm_2352 0.1027 ms 81.0% triton_mm_2356 0.1049 ms 79.3% triton_mm_2354 0.1221 ms 68.1% SingleProcess AUTOTUNE benchmarking takes 2.4264 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2253 ms 100.0% triton_mm_2397 0.2264 ms 99.5% triton_mm_2396 0.2332 ms 96.6% triton_mm_2398 0.2451 ms 91.9% triton_mm_2393 0.2596 ms 86.8% triton_mm_2391 0.2637 ms 85.4% triton_mm_2389 0.2681 ms 84.0% triton_mm_2392 0.2911 ms 77.4% triton_mm_2390 0.2933 ms 76.8% triton_mm_2394 0.3001 ms 75.1% SingleProcess AUTOTUNE benchmarking takes 2.5195 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) mm 0.2763 ms 100.0% triton_mm_2468 0.2768 ms 99.8% triton_mm_2469 0.2811 ms 98.3% triton_mm_2470 0.3171 ms 87.1% triton_mm_2461 0.3430 ms 80.5% triton_mm_2463 0.3452 ms 80.0% triton_mm_2465 0.3506 ms 78.8% triton_mm_2462 0.3660 ms 75.5% triton_mm_2466 0.3908 ms 70.7% triton_mm_2464 0.4264 ms 64.8% SingleProcess AUTOTUNE benchmarking takes 2.5287 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3032 ms 100.0% triton_mm_2506 0.3054 ms 99.3% triton_mm_2507 0.3127 ms 97.0% triton_mm_2508 0.3236 ms 93.7% triton_mm_2503 0.3470 ms 87.4% triton_mm_2501 0.3531 ms 85.9% triton_mm_2499 0.3553 ms 85.3% triton_mm_2502 0.3694 ms 82.1% triton_mm_2500 0.3908 ms 77.6% triton_mm_2504 0.3994 ms 75.9% SingleProcess AUTOTUNE benchmarking takes 2.5342 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2446 0.0173 ms 100.0% triton_mm_2443 0.0173 ms 99.8% triton_mm_2441 0.0180 ms 95.9% triton_mm_2445 0.0183 ms 94.6% triton_mm_2442 0.0184 ms 93.9% triton_mm_2447 0.0185 ms 93.6% triton_mm_2440 0.0187 ms 92.8% triton_mm_2444 0.0188 ms 92.0% triton_mm_2448 0.0193 ms 89.9% triton_mm_2451 0.0193 ms 89.6% SingleProcess AUTOTUNE benchmarking takes 2.1396 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2923 ms 100.0% triton_mm_2488 0.2994 ms 97.6% triton_mm_2487 0.3006 ms 97.2% triton_mm_2489 0.3050 ms 95.8% triton_mm_2482 0.3586 ms 81.5% triton_mm_2480 0.3594 ms 81.3% triton_mm_2484 0.3688 ms 79.2% triton_mm_2483 0.3854 ms 75.8% triton_mm_2485 0.4036 ms 72.4% triton_mm_2481 0.4092 ms 71.4% SingleProcess AUTOTUNE benchmarking takes 2.5427 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_2525 0.2929 ms 100.0% mm 0.2954 ms 99.1% triton_mm_2526 0.2983 ms 98.2% triton_mm_2527 0.3027 ms 96.7% triton_mm_2522 0.3510 ms 83.4% triton_mm_2518 0.3580 ms 81.8% triton_mm_2520 0.3628 ms 80.7% triton_mm_2521 0.3900 ms 75.1% triton_mm_2523 0.3971 ms 73.8% triton_mm_2519 0.4058 ms 72.2% SingleProcess AUTOTUNE benchmarking takes 2.5354 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2422 0.0269 ms 100.0% mm 0.0273 ms 98.5% triton_mm_2426 0.0322 ms 83.6% triton_mm_2430 0.0414 ms 65.0% triton_mm_2434 0.0508 ms 53.0% triton_mm_2421 0.0516 ms 52.1% triton_mm_2420 0.0570 ms 47.2% triton_mm_2425 0.0601 ms 44.8% triton_mm_2429 0.0626 ms 43.0% triton_mm_2419 0.0665 ms 40.5% SingleProcess AUTOTUNE benchmarking takes 2.0887 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2627 0.0173 ms 100.0% triton_mm_2629 0.0173 ms 100.0% triton_mm_2632 0.0173 ms 99.8% triton_mm_2633 0.0180 ms 96.3% triton_mm_2628 0.0184 ms 93.9% triton_mm_2630 0.0186 ms 93.0% triton_mm_2626 0.0187 ms 92.8% triton_mm_2634 0.0191 ms 90.6% triton_mm_2631 0.0195 ms 89.0% triton_mm_2636 0.0195 ms 88.8% SingleProcess AUTOTUNE benchmarking takes 2.1450 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2910 ms 100.0% triton_mm_2674 0.3002 ms 96.9% triton_mm_2673 0.3008 ms 96.7% triton_mm_2675 0.3038 ms 95.8% triton_mm_2668 0.3564 ms 81.6% triton_mm_2666 0.3593 ms 81.0% triton_mm_2670 0.3669 ms 79.3% triton_mm_2669 0.3837 ms 75.8% triton_mm_2671 0.4042 ms 72.0% triton_mm_2667 0.4092 ms 71.1% SingleProcess AUTOTUNE benchmarking takes 2.5408 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2939 ms 100.0% triton_mm_2711 0.2939 ms 100.0% triton_mm_2712 0.2985 ms 98.5% triton_mm_2713 0.3037 ms 96.8% triton_mm_2708 0.3494 ms 84.1% triton_mm_2704 0.3572 ms 82.3% triton_mm_2706 0.3627 ms 81.0% triton_mm_2707 0.3880 ms 75.8% triton_mm_2709 0.3971 ms 74.0% triton_mm_2705 0.4071 ms 72.2% SingleProcess AUTOTUNE benchmarking takes 2.5389 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0272 ms 100.0% triton_mm_2608 0.0280 ms 97.1% triton_mm_2612 0.0320 ms 84.9% triton_mm_2616 0.0412 ms 65.9% triton_mm_2620 0.0508 ms 53.5% triton_mm_2607 0.0524 ms 51.8% triton_mm_2606 0.0568 ms 47.9% triton_mm_2611 0.0609 ms 44.6% triton_mm_2615 0.0626 ms 43.4% triton_mm_2605 0.0663 ms 41.0% SingleProcess AUTOTUNE benchmarking takes 2.0886 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2818 0.0173 ms 100.0% triton_mm_2815 0.0173 ms 99.8% triton_mm_2812 0.0178 ms 97.1% triton_mm_2820 0.0184 ms 94.1% triton_mm_2822 0.0186 ms 93.3% triton_mm_2813 0.0186 ms 93.0% triton_mm_2819 0.0186 ms 93.0% triton_mm_2816 0.0188 ms 92.0% triton_mm_2823 0.0193 ms 89.6% triton_mm_2817 0.0196 ms 88.3% SingleProcess AUTOTUNE benchmarking takes 2.1404 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2916 ms 100.0% triton_mm_2860 0.3006 ms 97.0% triton_mm_2859 0.3012 ms 96.8% triton_mm_2861 0.3041 ms 95.9% triton_mm_2854 0.3561 ms 81.9% triton_mm_2852 0.3589 ms 81.3% triton_mm_2856 0.3670 ms 79.5% triton_mm_2855 0.3834 ms 76.1% triton_mm_2857 0.4032 ms 72.3% triton_mm_2853 0.4074 ms 71.6% SingleProcess AUTOTUNE benchmarking takes 2.5389 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2934 ms 100.0% triton_mm_2897 0.2936 ms 99.9% triton_mm_2898 0.2990 ms 98.1% triton_mm_2899 0.3037 ms 96.6% triton_mm_2894 0.3494 ms 84.0% triton_mm_2890 0.3566 ms 82.3% triton_mm_2892 0.3612 ms 81.2% triton_mm_2893 0.3872 ms 75.8% triton_mm_2895 0.3963 ms 74.0% triton_mm_2891 0.4053 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5362 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0271 ms 100.0% triton_mm_2794 0.0281 ms 96.5% triton_mm_2798 0.0322 ms 84.4% triton_mm_2802 0.0413 ms 65.7% triton_mm_2806 0.0508 ms 53.4% triton_mm_2793 0.0516 ms 52.5% triton_mm_2792 0.0559 ms 48.5% triton_mm_2797 0.0611 ms 44.4% triton_mm_2801 0.0626 ms 43.4% triton_mm_2791 0.0656 ms 41.3% SingleProcess AUTOTUNE benchmarking takes 2.0960 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_3083 0.2916 ms 100.0% mm 0.2916 ms 100.0% triton_mm_3084 0.2971 ms 98.1% triton_mm_3085 0.3031 ms 96.2% triton_mm_3080 0.3503 ms 83.2% triton_mm_3076 0.3572 ms 81.6% triton_mm_3078 0.3604 ms 80.9% triton_mm_3079 0.3874 ms 75.3% triton_mm_3081 0.3962 ms 73.6% triton_mm_3077 0.4031 ms 72.3% SingleProcess AUTOTUNE benchmarking takes 2.5380 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2980 0.0269 ms 100.0% mm 0.0272 ms 98.8% triton_mm_2984 0.0308 ms 87.3% triton_mm_2988 0.0412 ms 65.3% triton_mm_2992 0.0508 ms 52.9% triton_mm_2979 0.0524 ms 51.3% triton_mm_2978 0.0568 ms 47.4% triton_mm_2983 0.0609 ms 44.2% triton_mm_2987 0.0634 ms 42.4% triton_mm_2977 0.0655 ms 41.1% SingleProcess AUTOTUNE benchmarking takes 2.0956 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6840, 6840x64) mm 0.0295 ms 100.0% triton_mm_4874 0.0453 ms 65.2% triton_mm_4882 0.0483 ms 61.2% triton_mm_4878 0.0489 ms 60.4% triton_mm_4872 0.0591 ms 49.9% triton_mm_4873 0.0645 ms 45.8% triton_mm_4881 0.0676 ms 43.7% triton_mm_4871 0.0715 ms 41.3% triton_mm_4877 0.0723 ms 40.9% triton_mm_4880 0.0794 ms 37.2% SingleProcess AUTOTUNE benchmarking takes 9.5016 seconds and 0.0000 seconds precompiling /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): AUTOTUNE addmm(6840x1088, 6840x64, 64x1088) triton_mm_9 0.0322 ms 100.0% triton_mm_13 0.0329 ms 98.1% triton_mm_16 0.0346 ms 93.3% triton_mm_8 0.0348 ms 92.7% triton_mm_6 0.0349 ms 92.4% triton_mm_5 0.0353 ms 91.3% triton_mm_10 0.0374 ms 86.3% triton_mm_14 0.0376 ms 85.7% triton_mm_7 0.0379 ms 85.0% triton_mm_11 0.0384 ms 83.9% SingleProcess AUTOTUNE benchmarking takes 2.4376 seconds and 0.0064 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_35 0.0213 ms 100.0% triton_mm_31 0.0216 ms 98.5% triton_mm_26 0.0227 ms 93.9% triton_mm_27 0.0231 ms 92.2% triton_mm_30 0.0237 ms 89.8% triton_mm_23 0.0238 ms 89.6% bias_addmm 0.0241 ms 88.6% triton_mm_22 0.0245 ms 87.1% triton_mm_20 0.0246 ms 86.8% triton_mm_21 0.0252 ms 84.5% SingleProcess AUTOTUNE benchmarking takes 2.1945 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(6864x3264, 6864x1088, 1088x3264) triton_mm_52 0.2216 ms 100.0% triton_mm_53 0.2276 ms 97.4% triton_mm_54 0.2540 ms 87.3% triton_mm_47 0.2628 ms 84.3% triton_mm_49 0.2740 ms 80.9% triton_mm_45 0.2754 ms 80.5% bias_addmm 0.3010 ms 73.6% triton_mm_46 0.3063 ms 72.4% triton_mm_50 0.3113 ms 71.2% addmm 0.3243 ms 68.3% SingleProcess AUTOTUNE benchmarking takes 2.5728 seconds and 0.0330 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x1088, 1088x1088) triton_mm_72 0.0902 ms 100.0% triton_mm_71 0.0912 ms 98.9% bias_addmm 0.0970 ms 93.0% triton_mm_68 0.0974 ms 92.6% triton_mm_66 0.0975 ms 92.5% triton_mm_73 0.1000 ms 90.2% triton_mm_64 0.1011 ms 89.2% addmm 0.1102 ms 81.9% triton_mm_65 0.1108 ms 81.4% triton_mm_69 0.1130 ms 79.8% SingleProcess AUTOTUNE benchmarking takes 2.5041 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_90 0.2917 ms 100.0% triton_mm_91 0.2960 ms 98.6% bias_addmm 0.3279 ms 89.0% triton_mm_92 0.3331 ms 87.6% triton_mm_85 0.3436 ms 84.9% triton_mm_83 0.3623 ms 80.5% triton_mm_87 0.3645 ms 80.0% triton_mm_84 0.4016 ms 72.7% triton_mm_88 0.4099 ms 71.2% addmm 0.4170 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 2.6195 seconds and 0.0050 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2926 ms 100.0% triton_mm_110 0.3077 ms 95.1% triton_mm_109 0.3166 ms 92.4% addmm 0.3258 ms 89.8% triton_mm_111 0.3265 ms 89.6% triton_mm_106 0.3591 ms 81.5% triton_mm_104 0.3605 ms 81.2% triton_mm_105 0.3719 ms 78.7% triton_mm_102 0.3733 ms 78.4% triton_mm_107 0.4089 ms 71.6% SingleProcess AUTOTUNE benchmarking takes 2.6559 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_124 0.0216 ms 100.0% triton_mm_128 0.0224 ms 96.6% triton_mm_120 0.0229 ms 94.5% triton_mm_116 0.0229 ms 94.4% triton_mm_113 0.0235 ms 92.0% triton_mm_123 0.0237 ms 91.1% triton_mm_119 0.0238 ms 91.0% triton_mm_114 0.0242 ms 89.3% bias_addmm 0.0251 ms 86.1% triton_mm_115 0.0255 ms 84.8% SingleProcess AUTOTUNE benchmarking takes 2.1876 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_183 0.2919 ms 100.0% triton_mm_184 0.2964 ms 98.5% bias_addmm 0.3275 ms 89.1% triton_mm_185 0.3340 ms 87.4% triton_mm_178 0.3435 ms 85.0% triton_mm_176 0.3629 ms 80.4% triton_mm_180 0.3638 ms 80.2% triton_mm_177 0.4021 ms 72.6% triton_mm_181 0.4124 ms 70.8% addmm 0.4166 ms 70.1% SingleProcess AUTOTUNE benchmarking takes 2.6431 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2933 ms 100.0% triton_mm_203 0.3077 ms 95.3% triton_mm_202 0.3170 ms 92.5% triton_mm_204 0.3270 ms 89.7% addmm 0.3284 ms 89.3% triton_mm_199 0.3592 ms 81.6% triton_mm_197 0.3615 ms 81.1% triton_mm_198 0.3729 ms 78.7% triton_mm_195 0.3729 ms 78.6% triton_mm_200 0.4092 ms 71.7% SingleProcess AUTOTUNE benchmarking takes 2.6536 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_221 0.0213 ms 100.0% triton_mm_217 0.0216 ms 98.7% triton_mm_213 0.0231 ms 92.4% triton_mm_206 0.0236 ms 90.6% triton_mm_209 0.0239 ms 89.4% triton_mm_212 0.0239 ms 89.4% triton_mm_208 0.0245 ms 87.2% triton_mm_216 0.0249 ms 85.6% triton_mm_207 0.0254 ms 84.1% bias_addmm 0.0273 ms 78.2% SingleProcess AUTOTUNE benchmarking takes 2.1973 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_276 0.2920 ms 100.0% triton_mm_277 0.2962 ms 98.6% bias_addmm 0.3270 ms 89.3% triton_mm_278 0.3340 ms 87.4% triton_mm_271 0.3439 ms 84.9% triton_mm_269 0.3625 ms 80.6% triton_mm_273 0.3643 ms 80.2% triton_mm_270 0.4016 ms 72.7% triton_mm_274 0.4162 ms 70.2% addmm 0.4170 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 2.6444 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2942 ms 100.0% triton_mm_296 0.3082 ms 95.4% triton_mm_295 0.3172 ms 92.8% triton_mm_297 0.3268 ms 90.0% addmm 0.3280 ms 89.7% triton_mm_292 0.3589 ms 82.0% triton_mm_290 0.3609 ms 81.5% triton_mm_291 0.3722 ms 79.0% triton_mm_288 0.3752 ms 78.4% triton_mm_293 0.4095 ms 71.8% SingleProcess AUTOTUNE benchmarking takes 2.6511 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_310 0.0217 ms 100.0% triton_mm_314 0.0224 ms 96.9% triton_mm_305 0.0227 ms 95.3% triton_mm_306 0.0231 ms 93.6% triton_mm_309 0.0237 ms 91.2% triton_mm_302 0.0238 ms 91.1% triton_mm_300 0.0242 ms 89.4% triton_mm_301 0.0244 ms 88.6% triton_mm_299 0.0245 ms 88.5% bias_addmm 0.0271 ms 80.0% SingleProcess AUTOTUNE benchmarking takes 2.1908 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_369 0.2928 ms 100.0% triton_mm_370 0.2971 ms 98.6% bias_addmm 0.3258 ms 89.8% triton_mm_371 0.3335 ms 87.8% triton_mm_364 0.3439 ms 85.1% triton_mm_362 0.3621 ms 80.8% triton_mm_366 0.3633 ms 80.6% triton_mm_363 0.4014 ms 72.9% triton_mm_367 0.4125 ms 71.0% addmm 0.4169 ms 70.2% SingleProcess AUTOTUNE benchmarking takes 2.6482 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2931 ms 100.0% triton_mm_389 0.3076 ms 95.3% triton_mm_388 0.3168 ms 92.5% triton_mm_390 0.3266 ms 89.7% addmm 0.3274 ms 89.5% triton_mm_385 0.3584 ms 81.8% triton_mm_383 0.3608 ms 81.2% triton_mm_384 0.3725 ms 78.7% triton_mm_381 0.3735 ms 78.5% triton_mm_386 0.4093 ms 71.6% SingleProcess AUTOTUNE benchmarking takes 2.6588 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_482 0.3065 ms 100.0% triton_mm_481 0.3169 ms 96.7% triton_mm_483 0.3270 ms 93.7% bias_addmm 0.3307 ms 92.7% triton_mm_478 0.3574 ms 85.8% triton_mm_476 0.3587 ms 85.4% addmm 0.3664 ms 83.6% triton_mm_477 0.3719 ms 82.4% triton_mm_474 0.3739 ms 82.0% triton_mm_479 0.4079 ms 75.1% SingleProcess AUTOTUNE benchmarking takes 2.6471 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_575 0.3066 ms 100.0% triton_mm_574 0.3173 ms 96.6% triton_mm_576 0.3264 ms 93.9% addmm 0.3265 ms 93.9% bias_addmm 0.3307 ms 92.7% triton_mm_571 0.3572 ms 85.8% triton_mm_569 0.3590 ms 85.4% triton_mm_570 0.3724 ms 82.3% triton_mm_567 0.3726 ms 82.3% triton_mm_572 0.4073 ms 75.3% SingleProcess AUTOTUNE benchmarking takes 2.6633 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_668 0.3067 ms 100.0% triton_mm_667 0.3159 ms 97.1% triton_mm_669 0.3256 ms 94.2% addmm 0.3257 ms 94.2% bias_addmm 0.3317 ms 92.5% triton_mm_664 0.3574 ms 85.8% triton_mm_662 0.3595 ms 85.3% triton_mm_663 0.3710 ms 82.7% triton_mm_660 0.3720 ms 82.5% triton_mm_665 0.4076 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.6562 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_761 0.3063 ms 100.0% triton_mm_760 0.3172 ms 96.6% addmm 0.3256 ms 94.1% triton_mm_762 0.3268 ms 93.7% bias_addmm 0.3307 ms 92.6% triton_mm_757 0.3571 ms 85.8% triton_mm_755 0.3594 ms 85.2% triton_mm_756 0.3722 ms 82.3% triton_mm_753 0.3727 ms 82.2% triton_mm_758 0.4073 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.6506 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_854 0.3069 ms 100.0% triton_mm_853 0.3159 ms 97.2% triton_mm_855 0.3258 ms 94.2% bias_addmm 0.3317 ms 92.5% triton_mm_850 0.3576 ms 85.8% triton_mm_848 0.3593 ms 85.4% addmm 0.3664 ms 83.8% triton_mm_846 0.3687 ms 83.2% triton_mm_849 0.3709 ms 82.7% triton_mm_851 0.4079 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.6620 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_947 0.3057 ms 100.0% triton_mm_946 0.3172 ms 96.4% triton_mm_948 0.3260 ms 93.7% addmm 0.3265 ms 93.6% bias_addmm 0.3308 ms 92.4% triton_mm_943 0.3564 ms 85.8% triton_mm_941 0.3591 ms 85.1% triton_mm_942 0.3724 ms 82.1% triton_mm_939 0.3726 ms 82.0% triton_mm_944 0.4069 ms 75.1% SingleProcess AUTOTUNE benchmarking takes 2.6656 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x1088, 1088x2176) triton_mm_1604 0.0133 ms 100.0% triton_mm_1608 0.0139 ms 96.1% triton_mm_1612 0.0149 ms 89.3% mm 0.0167 ms 79.9% triton_mm_1616 0.0172 ms 77.4% triton_mm_1603 0.0175 ms 76.1% triton_mm_1607 0.0184 ms 72.5% triton_mm_1602 0.0186 ms 71.8% triton_mm_1601 0.0188 ms 70.9% triton_mm_1614 0.0196 ms 68.2% SingleProcess AUTOTUNE benchmarking takes 7.6230 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x32768, 6864x1088, 1088x32768) bias_addmm 2.1808 ms 100.0% triton_mm_1633 2.2212 ms 98.2% triton_mm_1634 2.2306 ms 97.8% triton_mm_1635 2.5131 ms 86.8% triton_mm_1628 2.6265 ms 83.0% triton_mm_1630 2.7599 ms 79.0% triton_mm_1626 2.8493 ms 76.5% addmm 2.9334 ms 74.3% triton_mm_1631 3.1185 ms 69.9% triton_mm_1627 3.1455 ms 69.3% SingleProcess AUTOTUNE benchmarking takes 12.0565 seconds and 0.0000 seconds precompiling AUTOTUNE mm(32768x6864, 6864x1088) triton_mm_1671 2.0713 ms 100.0% triton_mm_1672 2.1447 ms 96.6% mm 2.2452 ms 92.3% triton_mm_1673 2.2486 ms 92.1% triton_mm_1668 2.6188 ms 79.1% triton_mm_1666 2.6769 ms 77.4% triton_mm_1664 2.7070 ms 76.5% triton_mm_1669 2.7699 ms 74.8% triton_mm_1667 2.8420 ms 72.9% triton_mm_1665 2.9605 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 10.1418 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2907 ms 100.0% triton_mm_1744 0.2942 ms 98.8% triton_mm_1743 0.2970 ms 97.9% triton_mm_1745 0.2998 ms 96.9% triton_mm_1738 0.3513 ms 82.7% triton_mm_1736 0.3543 ms 82.0% triton_mm_1740 0.3624 ms 80.2% triton_mm_1739 0.3786 ms 76.8% triton_mm_1741 0.3967 ms 73.3% triton_mm_1737 0.4023 ms 72.2% SingleProcess AUTOTUNE benchmarking takes 10.0478 seconds and 0.0000 seconds precompiling AUTOTUNE mm(2176x24, 24x1088) triton_mm_1700 0.0113 ms 100.0% triton_mm_1702 0.0113 ms 100.0% triton_mm_1697 0.0114 ms 99.7% triton_mm_1699 0.0114 ms 99.4% triton_mm_1696 0.0118 ms 96.2% triton_mm_1701 0.0119 ms 94.9% triton_mm_1691 0.0122 ms 93.2% triton_mm_1695 0.0124 ms 91.5% triton_mm_1705 0.0127 ms 89.4% triton_mm_1707 0.0127 ms 89.4% SingleProcess AUTOTUNE benchmarking takes 5.9800 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x2176, 2176x1088) triton_mm_1678 0.0155 ms 100.0% triton_mm_1682 0.0162 ms 95.3% mm 0.0178 ms 86.9% triton_mm_1686 0.0201 ms 77.0% triton_mm_1677 0.0222 ms 69.6% triton_mm_1690 0.0224 ms 68.9% triton_mm_1676 0.0235 ms 65.9% triton_mm_1681 0.0260 ms 59.6% triton_mm_1675 0.0260 ms 59.5% triton_mm_1685 0.0267 ms 57.8% SingleProcess AUTOTUNE benchmarking takes 7.7125 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x32768, 32768x1088) mm 2.0902 ms 100.0% triton_mm_1652 2.2771 ms 91.8% triton_mm_1653 2.3024 ms 90.8% triton_mm_1654 2.3515 ms 88.9% triton_mm_1645 2.7436 ms 76.2% triton_mm_1647 2.7835 ms 75.1% triton_mm_1649 2.7865 ms 75.0% triton_mm_1648 2.8357 ms 73.7% triton_mm_1646 2.9594 ms 70.6% triton_mm_1650 2.9667 ms 70.5% SingleProcess AUTOTUNE benchmarking takes 10.2501 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1724 0.2755 ms 100.0% triton_mm_1725 0.2796 ms 98.5% mm 0.2798 ms 98.4% triton_mm_1726 0.3166 ms 87.0% triton_mm_1717 0.3426 ms 80.4% triton_mm_1719 0.3444 ms 80.0% triton_mm_1721 0.3469 ms 79.4% triton_mm_1718 0.3644 ms 75.6% triton_mm_1722 0.3863 ms 71.3% triton_mm_1720 0.4225 ms 65.2% SingleProcess AUTOTUNE benchmarking takes 2.5234 seconds and 0.0011 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) triton_mm_1839 0.2234 ms 100.0% triton_mm_1838 0.2286 ms 97.7% triton_mm_1840 0.2401 ms 93.0% triton_mm_1835 0.2557 ms 87.4% mm 0.2580 ms 86.6% triton_mm_1833 0.2621 ms 85.3% triton_mm_1831 0.2630 ms 85.0% triton_mm_1834 0.2870 ms 77.8% triton_mm_1832 0.2892 ms 77.3% triton_mm_1836 0.2961 ms 75.5% SingleProcess AUTOTUNE benchmarking takes 9.7710 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1781 0.2898 ms 100.0% mm 0.2920 ms 99.2% triton_mm_1782 0.2930 ms 98.9% triton_mm_1783 0.2995 ms 96.8% triton_mm_1778 0.3461 ms 83.7% triton_mm_1774 0.3540 ms 81.9% triton_mm_1776 0.3561 ms 81.4% triton_mm_1777 0.3806 ms 76.1% triton_mm_1779 0.3905 ms 74.2% triton_mm_1775 0.3988 ms 72.7% SingleProcess AUTOTUNE benchmarking takes 10.0936 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2244 ms 100.0% triton_mm_1857 0.2761 ms 81.3% triton_mm_1852 0.2828 ms 79.4% triton_mm_1858 0.2832 ms 79.3% triton_mm_1854 0.2841 ms 79.0% triton_mm_1850 0.2851 ms 78.7% triton_mm_1859 0.2876 ms 78.0% triton_mm_1853 0.3095 ms 72.5% triton_mm_1855 0.3231 ms 69.5% triton_mm_1851 0.3282 ms 68.4% SingleProcess AUTOTUNE benchmarking takes 9.9169 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.1010 ms 100.0% triton_mm_1814 0.1043 ms 96.8% triton_mm_1820 0.1078 ms 93.7% triton_mm_1811 0.1186 ms 85.1% triton_mm_1812 0.1196 ms 84.4% triton_mm_1816 0.1196 ms 84.4% triton_mm_1821 0.1208 ms 83.6% triton_mm_1819 0.1220 ms 82.8% triton_mm_1815 0.1270 ms 79.5% triton_mm_1817 0.1311 ms 77.0% SingleProcess AUTOTUNE benchmarking takes 9.7592 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3008 ms 100.0% triton_mm_1762 0.3027 ms 99.4% triton_mm_1763 0.3075 ms 97.8% triton_mm_1764 0.3179 ms 94.6% triton_mm_1759 0.3438 ms 87.5% triton_mm_1757 0.3479 ms 86.5% triton_mm_1755 0.3484 ms 86.3% triton_mm_1758 0.3618 ms 83.1% triton_mm_1756 0.3841 ms 78.3% triton_mm_1760 0.3932 ms 76.5% SingleProcess AUTOTUNE benchmarking takes 10.0326 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1801 0.0836 ms 100.0% mm 0.0840 ms 99.6% triton_mm_1800 0.0850 ms 98.4% triton_mm_1797 0.0916 ms 91.4% triton_mm_1802 0.0940 ms 89.0% triton_mm_1793 0.0950 ms 88.1% triton_mm_1795 0.0958 ms 87.3% triton_mm_1794 0.1015 ms 82.4% triton_mm_1798 0.1045 ms 80.0% triton_mm_1796 0.1227 ms 68.2% SingleProcess AUTOTUNE benchmarking takes 10.2404 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1910 0.2764 ms 100.0% mm 0.2777 ms 99.6% triton_mm_1911 0.2794 ms 99.0% triton_mm_1912 0.3156 ms 87.6% triton_mm_1903 0.3428 ms 80.6% triton_mm_1905 0.3455 ms 80.0% triton_mm_1907 0.3474 ms 79.6% triton_mm_1904 0.3659 ms 75.6% triton_mm_1908 0.3873 ms 71.4% triton_mm_1906 0.4234 ms 65.3% SingleProcess AUTOTUNE benchmarking takes 2.5282 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3004 ms 100.0% triton_mm_1948 0.3053 ms 98.4% triton_mm_1949 0.3134 ms 95.8% triton_mm_1950 0.3221 ms 93.3% triton_mm_1945 0.3470 ms 86.6% triton_mm_1943 0.3511 ms 85.5% triton_mm_1941 0.3524 ms 85.2% triton_mm_1944 0.3675 ms 81.7% triton_mm_1942 0.3879 ms 77.4% triton_mm_1946 0.3974 ms 75.6% SingleProcess AUTOTUNE benchmarking takes 2.5348 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_1888 0.0174 ms 100.0% triton_mm_1883 0.0176 ms 99.1% triton_mm_1885 0.0176 ms 98.9% triton_mm_1882 0.0178 ms 97.8% triton_mm_1889 0.0183 ms 95.4% triton_mm_1892 0.0184 ms 94.8% triton_mm_1886 0.0185 ms 94.1% triton_mm_1887 0.0186 ms 94.0% triton_mm_1890 0.0192 ms 91.0% triton_mm_1891 0.0192 ms 90.8% SingleProcess AUTOTUNE benchmarking takes 5.8540 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2915 ms 100.0% triton_mm_1930 0.2986 ms 97.6% triton_mm_1929 0.2995 ms 97.3% triton_mm_1931 0.3035 ms 96.0% triton_mm_1924 0.3566 ms 81.7% triton_mm_1922 0.3588 ms 81.2% triton_mm_1926 0.3670 ms 79.4% triton_mm_1925 0.3843 ms 75.9% triton_mm_1927 0.4019 ms 72.5% triton_mm_1923 0.4077 ms 71.5% SingleProcess AUTOTUNE benchmarking takes 2.5345 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2937 ms 100.0% triton_mm_1967 0.2940 ms 99.9% triton_mm_1968 0.2971 ms 98.8% triton_mm_1969 0.3042 ms 96.5% triton_mm_1964 0.3499 ms 83.9% triton_mm_1960 0.3580 ms 82.0% triton_mm_1962 0.3619 ms 81.2% triton_mm_1963 0.3869 ms 75.9% triton_mm_1965 0.3965 ms 74.1% triton_mm_1961 0.4049 ms 72.5% SingleProcess AUTOTUNE benchmarking takes 2.5304 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_1864 0.0267 ms 100.0% mm 0.0268 ms 99.9% triton_mm_1868 0.0307 ms 87.0% triton_mm_1872 0.0399 ms 67.0% triton_mm_1863 0.0514 ms 52.0% triton_mm_1876 0.0515 ms 51.9% triton_mm_1862 0.0561 ms 47.6% triton_mm_1867 0.0595 ms 44.9% triton_mm_1871 0.0628 ms 42.5% triton_mm_1861 0.0656 ms 40.8% SingleProcess AUTOTUNE benchmarking takes 6.7355 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0979 ms 100.0% triton_mm_2007 0.0995 ms 98.5% triton_mm_2000 0.1052 ms 93.0% triton_mm_2006 0.1086 ms 90.2% triton_mm_1997 0.1180 ms 83.0% triton_mm_1998 0.1197 ms 81.8% triton_mm_2002 0.1197 ms 81.8% triton_mm_2005 0.1222 ms 80.1% triton_mm_2001 0.1272 ms 77.0% triton_mm_2003 0.1309 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.4168 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1987 0.0829 ms 100.0% mm 0.0832 ms 99.6% triton_mm_1986 0.0844 ms 98.2% triton_mm_1983 0.0925 ms 89.6% triton_mm_1988 0.0943 ms 87.9% triton_mm_1979 0.0950 ms 87.2% triton_mm_1981 0.0950 ms 87.2% triton_mm_1980 0.1018 ms 81.5% triton_mm_1984 0.1048 ms 79.1% triton_mm_1982 0.1235 ms 67.1% SingleProcess AUTOTUNE benchmarking takes 2.4093 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2252 ms 100.0% triton_mm_2025 0.2268 ms 99.3% triton_mm_2024 0.2312 ms 97.4% triton_mm_2026 0.2442 ms 92.3% triton_mm_2021 0.2589 ms 87.0% triton_mm_2019 0.2649 ms 85.0% triton_mm_2017 0.2657 ms 84.8% triton_mm_2020 0.2905 ms 77.5% triton_mm_2018 0.2937 ms 76.7% triton_mm_2022 0.2990 ms 75.3% SingleProcess AUTOTUNE benchmarking takes 2.5114 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2096 0.2771 ms 100.0% mm 0.2771 ms 100.0% triton_mm_2097 0.2801 ms 98.9% triton_mm_2098 0.3172 ms 87.3% triton_mm_2089 0.3433 ms 80.7% triton_mm_2091 0.3461 ms 80.1% triton_mm_2093 0.3482 ms 79.6% triton_mm_2090 0.3666 ms 75.6% triton_mm_2094 0.3878 ms 71.4% triton_mm_2092 0.4234 ms 65.4% SingleProcess AUTOTUNE benchmarking takes 2.5299 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3037 ms 100.0% triton_mm_2134 0.3052 ms 99.5% triton_mm_2135 0.3125 ms 97.2% triton_mm_2136 0.3231 ms 94.0% triton_mm_2131 0.3480 ms 87.3% triton_mm_2129 0.3521 ms 86.2% triton_mm_2127 0.3548 ms 85.6% triton_mm_2130 0.3676 ms 82.6% triton_mm_2128 0.3899 ms 77.9% triton_mm_2132 0.3982 ms 76.3% SingleProcess AUTOTUNE benchmarking takes 2.5286 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2075 0.0172 ms 100.0% triton_mm_2074 0.0174 ms 98.5% triton_mm_2068 0.0178 ms 96.6% triton_mm_2076 0.0182 ms 94.0% triton_mm_2078 0.0183 ms 93.5% triton_mm_2069 0.0184 ms 93.2% triton_mm_2071 0.0184 ms 93.2% triton_mm_2072 0.0185 ms 92.7% triton_mm_2070 0.0185 ms 92.6% triton_mm_2073 0.0185 ms 92.6% SingleProcess AUTOTUNE benchmarking takes 2.1273 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0979 ms 100.0% triton_mm_2193 0.0992 ms 98.7% triton_mm_2186 0.1053 ms 93.0% triton_mm_2192 0.1087 ms 90.1% triton_mm_2183 0.1189 ms 82.3% triton_mm_2184 0.1189 ms 82.3% triton_mm_2188 0.1197 ms 81.8% triton_mm_2191 0.1221 ms 80.2% triton_mm_2187 0.1272 ms 77.0% triton_mm_2189 0.1309 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.4212 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) mm 0.0832 ms 100.0% triton_mm_2173 0.0838 ms 99.3% triton_mm_2172 0.0851 ms 97.8% triton_mm_2169 0.0917 ms 90.7% triton_mm_2174 0.0932 ms 89.3% triton_mm_2165 0.0943 ms 88.2% triton_mm_2167 0.0959 ms 86.8% triton_mm_2166 0.1031 ms 80.8% triton_mm_2170 0.1056 ms 78.8% triton_mm_2168 0.1227 ms 67.8% SingleProcess AUTOTUNE benchmarking takes 2.4108 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2247 ms 100.0% triton_mm_2211 0.2256 ms 99.6% triton_mm_2210 0.2314 ms 97.1% triton_mm_2212 0.2445 ms 91.9% triton_mm_2207 0.2600 ms 86.4% triton_mm_2205 0.2660 ms 84.5% triton_mm_2203 0.2665 ms 84.3% triton_mm_2206 0.2897 ms 77.6% triton_mm_2204 0.2926 ms 76.8% triton_mm_2208 0.3005 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.5099 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) triton_mm_2379 0.0992 ms 100.0% triton_mm_2372 0.1054 ms 94.1% triton_mm_2378 0.1085 ms 91.4% mm 0.1173 ms 84.6% triton_mm_2369 0.1188 ms 83.5% triton_mm_2370 0.1190 ms 83.4% triton_mm_2374 0.1197 ms 82.9% triton_mm_2377 0.1221 ms 81.3% triton_mm_2373 0.1271 ms 78.1% triton_mm_2375 0.1310 ms 75.7% SingleProcess AUTOTUNE benchmarking takes 2.4232 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2359 0.0829 ms 100.0% mm 0.0832 ms 99.7% triton_mm_2358 0.0844 ms 98.2% triton_mm_2355 0.0926 ms 89.5% triton_mm_2360 0.0942 ms 88.0% triton_mm_2353 0.0950 ms 87.3% triton_mm_2351 0.0952 ms 87.1% triton_mm_2352 0.1018 ms 81.5% triton_mm_2356 0.1047 ms 79.2% triton_mm_2354 0.1236 ms 67.1% SingleProcess AUTOTUNE benchmarking takes 2.4205 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2255 ms 100.0% triton_mm_2397 0.2264 ms 99.6% triton_mm_2396 0.2311 ms 97.6% triton_mm_2398 0.2438 ms 92.5% triton_mm_2393 0.2593 ms 87.0% triton_mm_2391 0.2648 ms 85.2% triton_mm_2389 0.2660 ms 84.8% triton_mm_2392 0.2908 ms 77.6% triton_mm_2390 0.2925 ms 77.1% triton_mm_2394 0.2990 ms 75.4% SingleProcess AUTOTUNE benchmarking takes 2.5142 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0980 ms 100.0% triton_mm_2565 0.0996 ms 98.4% triton_mm_2558 0.1053 ms 93.0% triton_mm_2564 0.1088 ms 90.1% triton_mm_2555 0.1180 ms 83.1% triton_mm_2560 0.1197 ms 81.9% triton_mm_2556 0.1197 ms 81.9% triton_mm_2563 0.1221 ms 80.2% triton_mm_2559 0.1271 ms 77.1% triton_mm_2561 0.1310 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.4203 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2545 0.0829 ms 100.0% mm 0.0832 ms 99.7% triton_mm_2544 0.0844 ms 98.2% triton_mm_2541 0.0917 ms 90.4% triton_mm_2546 0.0939 ms 88.3% triton_mm_2539 0.0950 ms 87.3% triton_mm_2537 0.0951 ms 87.2% triton_mm_2538 0.1017 ms 81.5% triton_mm_2542 0.1055 ms 78.6% triton_mm_2540 0.1235 ms 67.1% SingleProcess AUTOTUNE benchmarking takes 2.4256 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2243 ms 100.0% triton_mm_2583 0.2272 ms 98.7% triton_mm_2582 0.2314 ms 96.9% triton_mm_2584 0.2437 ms 92.0% triton_mm_2579 0.2589 ms 86.6% triton_mm_2577 0.2657 ms 84.4% triton_mm_2575 0.2671 ms 83.9% triton_mm_2578 0.2897 ms 77.4% triton_mm_2576 0.2931 ms 76.5% triton_mm_2580 0.2991 ms 75.0% SingleProcess AUTOTUNE benchmarking takes 2.5160 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0971 ms 100.0% triton_mm_2751 0.0994 ms 97.7% triton_mm_2744 0.1044 ms 93.0% triton_mm_2750 0.1086 ms 89.4% triton_mm_2741 0.1189 ms 81.7% triton_mm_2742 0.1189 ms 81.7% triton_mm_2746 0.1197 ms 81.1% triton_mm_2749 0.1221 ms 79.5% triton_mm_2745 0.1260 ms 77.1% triton_mm_2747 0.1308 ms 74.2% SingleProcess AUTOTUNE benchmarking takes 2.4193 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2731 0.0829 ms 100.0% mm 0.0832 ms 99.6% triton_mm_2730 0.0844 ms 98.2% triton_mm_2727 0.0916 ms 90.5% triton_mm_2732 0.0943 ms 87.9% triton_mm_2725 0.0948 ms 87.4% triton_mm_2723 0.0950 ms 87.2% triton_mm_2724 0.1015 ms 81.7% triton_mm_2728 0.1046 ms 79.3% triton_mm_2726 0.1224 ms 67.7% SingleProcess AUTOTUNE benchmarking takes 2.4026 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2244 ms 100.0% triton_mm_2769 0.2268 ms 98.9% triton_mm_2768 0.2314 ms 97.0% triton_mm_2770 0.2442 ms 91.9% triton_mm_2765 0.2588 ms 86.7% triton_mm_2763 0.2649 ms 84.7% triton_mm_2761 0.2657 ms 84.5% triton_mm_2764 0.2905 ms 77.2% triton_mm_2762 0.2939 ms 76.4% triton_mm_2766 0.2988 ms 75.1% SingleProcess AUTOTUNE benchmarking takes 2.5095 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6840, 6840x64) mm 0.0295 ms 100.0% triton_mm_4874 0.0451 ms 65.3% triton_mm_4882 0.0479 ms 61.4% triton_mm_4878 0.0492 ms 59.9% triton_mm_4872 0.0590 ms 49.9% triton_mm_4873 0.0643 ms 45.8% triton_mm_4881 0.0675 ms 43.6% triton_mm_4871 0.0711 ms 41.4% triton_mm_4877 0.0731 ms 40.3% triton_mm_4880 0.0791 ms 37.2% SingleProcess AUTOTUNE benchmarking takes 8.2142 seconds and 0.0000 seconds precompiling /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): AUTOTUNE addmm(6840x1088, 6840x64, 64x1088) triton_mm_9 0.0324 ms 100.0% triton_mm_13 0.0328 ms 98.5% triton_mm_8 0.0340 ms 95.1% triton_mm_16 0.0349 ms 92.8% triton_mm_6 0.0358 ms 90.3% triton_mm_5 0.0361 ms 89.5% triton_mm_7 0.0372 ms 86.9% triton_mm_11 0.0377 ms 85.8% triton_mm_14 0.0380 ms 85.0% triton_mm_10 0.0386 ms 83.8% SingleProcess AUTOTUNE benchmarking takes 2.4236 seconds and 0.0249 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_35 0.0212 ms 100.0% triton_mm_26 0.0225 ms 94.0% triton_mm_31 0.0229 ms 92.7% triton_mm_23 0.0229 ms 92.6% triton_mm_27 0.0231 ms 91.7% triton_mm_20 0.0235 ms 90.3% triton_mm_30 0.0235 ms 90.2% triton_mm_21 0.0240 ms 88.1% bias_addmm 0.0241 ms 88.0% triton_mm_22 0.0245 ms 86.5% SingleProcess AUTOTUNE benchmarking takes 2.1986 seconds and 0.0007 seconds precompiling AUTOTUNE addmm(6864x3264, 6864x1088, 1088x3264) triton_mm_52 0.2258 ms 100.0% triton_mm_53 0.2316 ms 97.5% triton_mm_54 0.2607 ms 86.6% triton_mm_47 0.2716 ms 83.2% triton_mm_49 0.2798 ms 80.7% triton_mm_45 0.2837 ms 79.6% bias_addmm 0.2993 ms 75.4% triton_mm_46 0.3146 ms 71.8% triton_mm_50 0.3175 ms 71.1% addmm 0.3226 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 2.5971 seconds and 0.0197 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x1088, 1088x1088) triton_mm_72 0.0912 ms 100.0% triton_mm_71 0.0914 ms 99.8% bias_addmm 0.0985 ms 92.6% triton_mm_68 0.0994 ms 91.8% triton_mm_73 0.1004 ms 90.9% triton_mm_66 0.1012 ms 90.1% triton_mm_64 0.1033 ms 88.3% addmm 0.1122 ms 81.3% triton_mm_65 0.1125 ms 81.1% triton_mm_69 0.1140 ms 80.0% SingleProcess AUTOTUNE benchmarking takes 2.5107 seconds and 0.0010 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_90 0.2954 ms 100.0% triton_mm_91 0.2977 ms 99.2% bias_addmm 0.3259 ms 90.6% triton_mm_92 0.3383 ms 87.3% triton_mm_85 0.3489 ms 84.7% triton_mm_83 0.3665 ms 80.6% triton_mm_87 0.3675 ms 80.4% triton_mm_84 0.4085 ms 72.3% triton_mm_88 0.4157 ms 71.1% addmm 0.4175 ms 70.8% SingleProcess AUTOTUNE benchmarking takes 2.6165 seconds and 0.0372 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2930 ms 100.0% triton_mm_110 0.3113 ms 94.1% triton_mm_109 0.3194 ms 91.7% addmm 0.3302 ms 88.7% triton_mm_111 0.3308 ms 88.6% triton_mm_106 0.3610 ms 81.2% triton_mm_104 0.3644 ms 80.4% triton_mm_105 0.3756 ms 78.0% triton_mm_102 0.3772 ms 77.7% triton_mm_107 0.4136 ms 70.8% SingleProcess AUTOTUNE benchmarking takes 2.6586 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_128 0.0212 ms 100.0% triton_mm_124 0.0217 ms 97.6% triton_mm_116 0.0226 ms 93.8% triton_mm_120 0.0229 ms 92.7% triton_mm_119 0.0235 ms 90.1% triton_mm_114 0.0240 ms 88.4% triton_mm_113 0.0244 ms 87.1% triton_mm_123 0.0244 ms 86.9% triton_mm_115 0.0245 ms 86.7% bias_addmm 0.0250 ms 84.9% SingleProcess AUTOTUNE benchmarking takes 2.2060 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_203 0.3088 ms 100.0% triton_mm_202 0.3179 ms 97.1% bias_addmm 0.3277 ms 94.2% triton_mm_204 0.3285 ms 94.0% addmm 0.3287 ms 93.9% triton_mm_199 0.3593 ms 85.9% triton_mm_197 0.3618 ms 85.3% triton_mm_198 0.3756 ms 82.2% triton_mm_195 0.3763 ms 82.1% triton_mm_200 0.4121 ms 74.9% SingleProcess AUTOTUNE benchmarking takes 2.6403 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_296 0.3081 ms 100.0% triton_mm_295 0.3178 ms 96.9% bias_addmm 0.3285 ms 93.8% triton_mm_297 0.3286 ms 93.8% triton_mm_292 0.3577 ms 86.2% triton_mm_290 0.3611 ms 85.3% addmm 0.3675 ms 83.8% triton_mm_288 0.3732 ms 82.6% triton_mm_291 0.3733 ms 82.5% triton_mm_293 0.4087 ms 75.4% SingleProcess AUTOTUNE benchmarking takes 2.6423 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_389 0.3092 ms 100.0% triton_mm_388 0.3172 ms 97.5% bias_addmm 0.3280 ms 94.3% triton_mm_390 0.3289 ms 94.0% triton_mm_385 0.3584 ms 86.3% triton_mm_383 0.3618 ms 85.5% addmm 0.3677 ms 84.1% triton_mm_384 0.3727 ms 83.0% triton_mm_381 0.3745 ms 82.6% triton_mm_386 0.4095 ms 75.5% SingleProcess AUTOTUNE benchmarking takes 2.6543 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) triton_mm_482 0.3074 ms 100.0% triton_mm_481 0.3178 ms 96.7% bias_addmm 0.3277 ms 93.8% triton_mm_483 0.3280 ms 93.7% addmm 0.3295 ms 93.3% triton_mm_478 0.3580 ms 85.9% triton_mm_476 0.3599 ms 85.4% triton_mm_477 0.3733 ms 82.3% triton_mm_474 0.3746 ms 82.1% triton_mm_479 0.4088 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.6606 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x1088, 1088x2176) triton_mm_1604 0.0133 ms 100.0% triton_mm_1608 0.0139 ms 95.6% triton_mm_1612 0.0159 ms 83.7% triton_mm_1616 0.0159 ms 83.7% mm 0.0166 ms 80.0% triton_mm_1603 0.0166 ms 80.0% triton_mm_1607 0.0174 ms 76.6% triton_mm_1602 0.0183 ms 72.6% triton_mm_1601 0.0188 ms 71.0% triton_mm_1611 0.0199 ms 67.0% SingleProcess AUTOTUNE benchmarking takes 7.6293 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x32768, 6864x1088, 1088x32768) bias_addmm 2.2057 ms 100.0% triton_mm_1633 2.2452 ms 98.2% triton_mm_1634 2.2555 ms 97.8% triton_mm_1635 2.5116 ms 87.8% triton_mm_1628 2.6881 ms 82.1% triton_mm_1630 2.7895 ms 79.1% triton_mm_1626 2.8800 ms 76.6% addmm 2.9780 ms 74.1% triton_mm_1627 3.1433 ms 70.2% triton_mm_1631 3.1531 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 11.9184 seconds and 0.0000 seconds precompiling AUTOTUNE mm(32768x6864, 6864x1088) triton_mm_1671 2.0920 ms 100.0% triton_mm_1672 2.1700 ms 96.4% mm 2.2301 ms 93.8% triton_mm_1673 2.2748 ms 92.0% triton_mm_1668 2.6494 ms 79.0% triton_mm_1666 2.6834 ms 78.0% triton_mm_1664 2.7115 ms 77.2% triton_mm_1669 2.8057 ms 74.6% triton_mm_1667 2.8686 ms 72.9% triton_mm_1665 2.9595 ms 70.7% SingleProcess AUTOTUNE benchmarking takes 10.1533 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2924 ms 100.0% triton_mm_1744 0.2946 ms 99.3% triton_mm_1743 0.2962 ms 98.7% triton_mm_1745 0.2996 ms 97.6% triton_mm_1738 0.3523 ms 83.0% triton_mm_1736 0.3546 ms 82.5% triton_mm_1740 0.3621 ms 80.8% triton_mm_1739 0.3772 ms 77.5% triton_mm_1741 0.3972 ms 73.6% triton_mm_1737 0.4042 ms 72.3% SingleProcess AUTOTUNE benchmarking takes 9.9981 seconds and 0.0000 seconds precompiling AUTOTUNE mm(2176x24, 24x1088) triton_mm_1696 0.0112 ms 100.0% triton_mm_1700 0.0113 ms 99.4% triton_mm_1702 0.0113 ms 99.4% triton_mm_1699 0.0115 ms 97.5% triton_mm_1706 0.0119 ms 94.1% triton_mm_1698 0.0120 ms 93.3% triton_mm_1691 0.0120 ms 93.1% triton_mm_1705 0.0124 ms 90.7% triton_mm_1707 0.0124 ms 90.7% triton_mm_1693 0.0124 ms 90.4% SingleProcess AUTOTUNE benchmarking takes 6.1763 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x2176, 2176x1088) triton_mm_1682 0.0163 ms 100.0% triton_mm_1678 0.0167 ms 97.5% mm 0.0170 ms 95.7% triton_mm_1686 0.0201 ms 80.8% triton_mm_1677 0.0222 ms 73.2% triton_mm_1690 0.0224 ms 72.5% triton_mm_1676 0.0246 ms 66.0% triton_mm_1685 0.0254 ms 64.0% triton_mm_1681 0.0260 ms 62.5% triton_mm_1675 0.0270 ms 60.3% SingleProcess AUTOTUNE benchmarking takes 7.8090 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x32768, 32768x1088) mm 2.1137 ms 100.0% triton_mm_1652 2.3067 ms 91.6% triton_mm_1653 2.3414 ms 90.3% triton_mm_1654 2.3768 ms 88.9% triton_mm_1645 2.8430 ms 74.3% triton_mm_1647 2.8509 ms 74.1% triton_mm_1649 2.8647 ms 73.8% triton_mm_1648 2.9504 ms 71.6% triton_mm_1646 3.0095 ms 70.2% triton_mm_1650 3.0176 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 2.9320 seconds and 0.0009 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1724 0.2742 ms 100.0% triton_mm_1725 0.2760 ms 99.3% mm 0.2809 ms 97.6% triton_mm_1726 0.3121 ms 87.9% triton_mm_1717 0.3371 ms 81.3% triton_mm_1719 0.3398 ms 80.7% triton_mm_1721 0.3439 ms 79.7% triton_mm_1718 0.3622 ms 75.7% triton_mm_1722 0.3833 ms 71.5% triton_mm_1720 0.4188 ms 65.5% SingleProcess AUTOTUNE benchmarking takes 10.3370 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) triton_mm_1839 0.2268 ms 100.0% triton_mm_1838 0.2318 ms 97.9% triton_mm_1840 0.2440 ms 93.0% mm 0.2564 ms 88.5% triton_mm_1835 0.2599 ms 87.3% triton_mm_1831 0.2645 ms 85.8% triton_mm_1833 0.2652 ms 85.6% triton_mm_1834 0.2924 ms 77.6% triton_mm_1832 0.2941 ms 77.1% triton_mm_1836 0.2997 ms 75.7% SingleProcess AUTOTUNE benchmarking takes 2.4998 seconds and 0.0009 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1781 0.2896 ms 100.0% mm 0.2915 ms 99.3% triton_mm_1782 0.2933 ms 98.7% triton_mm_1783 0.2982 ms 97.1% triton_mm_1778 0.3460 ms 83.7% triton_mm_1774 0.3519 ms 82.3% triton_mm_1776 0.3556 ms 81.4% triton_mm_1777 0.3815 ms 75.9% triton_mm_1779 0.3918 ms 73.9% triton_mm_1775 0.3990 ms 72.6% SingleProcess AUTOTUNE benchmarking takes 9.9384 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2254 ms 100.0% triton_mm_1857 0.2773 ms 81.3% triton_mm_1854 0.2833 ms 79.6% triton_mm_1852 0.2838 ms 79.4% triton_mm_1858 0.2841 ms 79.3% triton_mm_1850 0.2856 ms 78.9% triton_mm_1859 0.2871 ms 78.5% triton_mm_1853 0.3106 ms 72.6% triton_mm_1855 0.3238 ms 69.6% triton_mm_1851 0.3287 ms 68.6% SingleProcess AUTOTUNE benchmarking takes 9.9259 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) triton_mm_1821 0.0992 ms 100.0% mm 0.1001 ms 99.1% triton_mm_1814 0.1043 ms 95.2% triton_mm_1820 0.1074 ms 92.4% triton_mm_1811 0.1179 ms 84.2% triton_mm_1812 0.1188 ms 83.5% triton_mm_1816 0.1204 ms 82.4% triton_mm_1819 0.1210 ms 82.0% triton_mm_1815 0.1270 ms 78.1% triton_mm_1817 0.1299 ms 76.4% SingleProcess AUTOTUNE benchmarking takes 9.9094 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3016 ms 100.0% triton_mm_1762 0.3030 ms 99.5% triton_mm_1763 0.3078 ms 98.0% triton_mm_1764 0.3183 ms 94.8% triton_mm_1759 0.3421 ms 88.2% triton_mm_1757 0.3487 ms 86.5% triton_mm_1755 0.3493 ms 86.3% triton_mm_1758 0.3633 ms 83.0% triton_mm_1756 0.3849 ms 78.4% triton_mm_1760 0.3933 ms 76.7% SingleProcess AUTOTUNE benchmarking takes 9.7705 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) mm 0.0839 ms 100.0% triton_mm_1801 0.0840 ms 99.8% triton_mm_1800 0.0858 ms 97.8% triton_mm_1797 0.0931 ms 90.1% triton_mm_1802 0.0936 ms 89.6% triton_mm_1793 0.0953 ms 88.0% triton_mm_1795 0.0959 ms 87.5% triton_mm_1794 0.1027 ms 81.7% triton_mm_1798 0.1049 ms 80.0% triton_mm_1796 0.1220 ms 68.7% SingleProcess AUTOTUNE benchmarking takes 2.4161 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) mm 0.2770 ms 100.0% triton_mm_1910 0.2776 ms 99.8% triton_mm_1911 0.2820 ms 98.2% triton_mm_1912 0.3194 ms 86.7% triton_mm_1903 0.3467 ms 79.9% triton_mm_1905 0.3467 ms 79.9% triton_mm_1907 0.3504 ms 79.0% triton_mm_1904 0.3676 ms 75.3% triton_mm_1908 0.3910 ms 70.8% triton_mm_1906 0.4284 ms 64.7% SingleProcess AUTOTUNE benchmarking takes 2.5279 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3017 ms 100.0% triton_mm_1948 0.3056 ms 98.7% triton_mm_1949 0.3135 ms 96.2% triton_mm_1950 0.3249 ms 92.9% triton_mm_1945 0.3469 ms 87.0% triton_mm_1941 0.3550 ms 85.0% triton_mm_1943 0.3555 ms 84.9% triton_mm_1944 0.3724 ms 81.0% triton_mm_1942 0.3910 ms 77.2% triton_mm_1946 0.4008 ms 75.3% SingleProcess AUTOTUNE benchmarking takes 2.5366 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_1885 0.0172 ms 100.0% triton_mm_1888 0.0173 ms 99.3% triton_mm_1882 0.0177 ms 96.8% triton_mm_1889 0.0182 ms 94.2% triton_mm_1883 0.0182 ms 94.0% triton_mm_1892 0.0183 ms 93.9% triton_mm_1884 0.0183 ms 93.5% triton_mm_1886 0.0184 ms 93.4% triton_mm_1891 0.0191 ms 89.6% triton_mm_1893 0.0192 ms 89.5% SingleProcess AUTOTUNE benchmarking takes 5.8688 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2915 ms 100.0% triton_mm_1929 0.3011 ms 96.8% triton_mm_1930 0.3012 ms 96.8% triton_mm_1931 0.3046 ms 95.7% triton_mm_1924 0.3580 ms 81.4% triton_mm_1922 0.3608 ms 80.8% triton_mm_1926 0.3702 ms 78.7% triton_mm_1925 0.3866 ms 75.4% triton_mm_1927 0.4060 ms 71.8% triton_mm_1923 0.4095 ms 71.2% SingleProcess AUTOTUNE benchmarking takes 2.5375 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1967 0.2927 ms 100.0% mm 0.2957 ms 99.0% triton_mm_1968 0.2981 ms 98.2% triton_mm_1969 0.3044 ms 96.1% triton_mm_1964 0.3538 ms 82.7% triton_mm_1960 0.3608 ms 81.1% triton_mm_1962 0.3634 ms 80.5% triton_mm_1963 0.3904 ms 75.0% triton_mm_1965 0.3985 ms 73.4% triton_mm_1961 0.4073 ms 71.9% SingleProcess AUTOTUNE benchmarking takes 2.5320 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0267 ms 100.0% triton_mm_1864 0.0276 ms 96.5% triton_mm_1868 0.0307 ms 87.1% triton_mm_1872 0.0399 ms 66.9% triton_mm_1876 0.0508 ms 52.6% triton_mm_1863 0.0523 ms 51.0% triton_mm_1862 0.0559 ms 47.8% triton_mm_1867 0.0601 ms 44.4% triton_mm_1871 0.0616 ms 43.3% triton_mm_1861 0.0659 ms 40.5% SingleProcess AUTOTUNE benchmarking takes 6.6771 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2253 ms 100.0% triton_mm_2043 0.2797 ms 80.6% triton_mm_2038 0.2869 ms 78.5% triton_mm_2044 0.2871 ms 78.5% triton_mm_2040 0.2879 ms 78.3% triton_mm_2036 0.2901 ms 77.7% triton_mm_2045 0.2912 ms 77.4% triton_mm_2039 0.3150 ms 71.5% triton_mm_2041 0.3276 ms 68.8% triton_mm_2037 0.3329 ms 67.7% SingleProcess AUTOTUNE benchmarking takes 2.5073 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1987 0.0832 ms 100.0% mm 0.0836 ms 99.5% triton_mm_1986 0.0848 ms 98.2% triton_mm_1983 0.0930 ms 89.5% triton_mm_1988 0.0943 ms 88.3% triton_mm_1979 0.0952 ms 87.4% triton_mm_1981 0.0959 ms 86.8% triton_mm_1980 0.1027 ms 81.0% triton_mm_1984 0.1049 ms 79.3% triton_mm_1982 0.1219 ms 68.3% SingleProcess AUTOTUNE benchmarking takes 2.4134 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2265 ms 100.0% triton_mm_2025 0.2266 ms 100.0% triton_mm_2024 0.2318 ms 97.7% triton_mm_2026 0.2453 ms 92.3% triton_mm_2021 0.2615 ms 86.6% triton_mm_2019 0.2661 ms 85.1% triton_mm_2017 0.2674 ms 84.7% triton_mm_2020 0.2929 ms 77.3% triton_mm_2018 0.2933 ms 77.2% triton_mm_2022 0.3009 ms 75.3% SingleProcess AUTOTUNE benchmarking takes 2.5132 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2096 0.2763 ms 100.0% mm 0.2772 ms 99.7% triton_mm_2097 0.2800 ms 98.7% triton_mm_2098 0.3191 ms 86.6% triton_mm_2089 0.3425 ms 80.7% triton_mm_2091 0.3461 ms 79.8% triton_mm_2093 0.3504 ms 78.8% triton_mm_2090 0.3682 ms 75.0% triton_mm_2094 0.3906 ms 70.7% triton_mm_2092 0.4270 ms 64.7% SingleProcess AUTOTUNE benchmarking takes 2.5194 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.3022 ms 100.0% triton_mm_2134 0.3063 ms 98.7% triton_mm_2135 0.3145 ms 96.1% triton_mm_2136 0.3235 ms 93.4% triton_mm_2131 0.3473 ms 87.0% triton_mm_2127 0.3540 ms 85.4% triton_mm_2129 0.3564 ms 84.8% triton_mm_2130 0.3704 ms 81.6% triton_mm_2128 0.3905 ms 77.4% triton_mm_2132 0.3989 ms 75.8% SingleProcess AUTOTUNE benchmarking takes 2.5288 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2075 0.0171 ms 100.0% triton_mm_2069 0.0171 ms 99.8% triton_mm_2072 0.0173 ms 98.9% triton_mm_2076 0.0182 ms 94.0% triton_mm_2078 0.0183 ms 93.5% triton_mm_2070 0.0183 ms 93.2% triton_mm_2071 0.0185 ms 92.5% triton_mm_2074 0.0185 ms 92.2% triton_mm_2068 0.0186 ms 91.9% triton_mm_2079 0.0192 ms 89.1% SingleProcess AUTOTUNE benchmarking takes 2.1131 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2923 ms 100.0% triton_mm_2116 0.3002 ms 97.4% triton_mm_2115 0.3006 ms 97.3% triton_mm_2117 0.3065 ms 95.4% triton_mm_2110 0.3580 ms 81.6% triton_mm_2108 0.3606 ms 81.1% triton_mm_2112 0.3695 ms 79.1% triton_mm_2111 0.3850 ms 75.9% triton_mm_2113 0.4038 ms 72.4% triton_mm_2109 0.4108 ms 71.2% SingleProcess AUTOTUNE benchmarking takes 2.5342 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_2153 0.2930 ms 100.0% mm 0.2947 ms 99.4% triton_mm_2154 0.2985 ms 98.2% triton_mm_2155 0.3049 ms 96.1% triton_mm_2150 0.3513 ms 83.4% triton_mm_2146 0.3604 ms 81.3% triton_mm_2148 0.3617 ms 81.0% triton_mm_2149 0.3899 ms 75.2% triton_mm_2151 0.3980 ms 73.6% triton_mm_2147 0.4049 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5245 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0267 ms 100.0% triton_mm_2050 0.0278 ms 96.0% triton_mm_2054 0.0320 ms 83.4% triton_mm_2058 0.0399 ms 66.8% triton_mm_2062 0.0508 ms 52.5% triton_mm_2049 0.0524 ms 50.8% triton_mm_2048 0.0560 ms 47.6% triton_mm_2053 0.0593 ms 44.9% triton_mm_2057 0.0624 ms 42.7% triton_mm_2047 0.0653 ms 40.8% SingleProcess AUTOTUNE benchmarking takes 2.0743 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2251 ms 100.0% triton_mm_2229 0.2788 ms 80.7% triton_mm_2224 0.2858 ms 78.8% triton_mm_2230 0.2868 ms 78.5% triton_mm_2226 0.2886 ms 78.0% triton_mm_2222 0.2905 ms 77.5% triton_mm_2231 0.2911 ms 77.3% triton_mm_2225 0.3147 ms 71.5% triton_mm_2227 0.3274 ms 68.7% triton_mm_2223 0.3329 ms 67.6% SingleProcess AUTOTUNE benchmarking takes 2.5135 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) mm 0.0836 ms 100.0% triton_mm_2173 0.0839 ms 99.6% triton_mm_2172 0.0847 ms 98.7% triton_mm_2169 0.0929 ms 90.0% triton_mm_2174 0.0936 ms 89.3% triton_mm_2165 0.0947 ms 88.2% triton_mm_2167 0.0961 ms 87.0% triton_mm_2166 0.1029 ms 81.2% triton_mm_2170 0.1048 ms 79.7% triton_mm_2168 0.1227 ms 68.1% SingleProcess AUTOTUNE benchmarking takes 2.4070 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2251 ms 100.0% triton_mm_2211 0.2270 ms 99.1% triton_mm_2210 0.2319 ms 97.1% triton_mm_2212 0.2445 ms 92.1% triton_mm_2207 0.2612 ms 86.2% triton_mm_2205 0.2651 ms 84.9% triton_mm_2203 0.2657 ms 84.7% triton_mm_2206 0.2927 ms 76.9% triton_mm_2204 0.2946 ms 76.4% triton_mm_2208 0.3006 ms 74.9% SingleProcess AUTOTUNE benchmarking takes 2.4960 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2902 ms 100.0% triton_mm_2302 0.2999 ms 96.8% triton_mm_2301 0.3009 ms 96.4% triton_mm_2303 0.3034 ms 95.7% triton_mm_2296 0.3572 ms 81.3% triton_mm_2294 0.3585 ms 81.0% triton_mm_2298 0.3676 ms 78.9% triton_mm_2297 0.3845 ms 75.5% triton_mm_2299 0.4035 ms 71.9% triton_mm_2295 0.4082 ms 71.1% SingleProcess AUTOTUNE benchmarking takes 2.5361 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_2339 0.2930 ms 100.0% mm 0.2935 ms 99.8% triton_mm_2340 0.2983 ms 98.2% triton_mm_2341 0.3033 ms 96.6% triton_mm_2336 0.3511 ms 83.5% triton_mm_2332 0.3588 ms 81.7% triton_mm_2334 0.3620 ms 81.0% triton_mm_2335 0.3883 ms 75.5% triton_mm_2337 0.3961 ms 74.0% triton_mm_2333 0.4057 ms 72.2% SingleProcess AUTOTUNE benchmarking takes 2.5371 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0267 ms 100.0% triton_mm_2236 0.0278 ms 96.0% triton_mm_2240 0.0320 ms 83.6% triton_mm_2244 0.0399 ms 67.0% triton_mm_2248 0.0520 ms 51.4% triton_mm_2235 0.0524 ms 51.0% triton_mm_2234 0.0561 ms 47.7% triton_mm_2239 0.0600 ms 44.5% triton_mm_2243 0.0624 ms 42.8% triton_mm_2233 0.0651 ms 41.0% SingleProcess AUTOTUNE benchmarking takes 2.0899 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2359 0.0839 ms 100.0% mm 0.0843 ms 99.5% triton_mm_2358 0.0854 ms 98.2% triton_mm_2355 0.0920 ms 91.3% triton_mm_2360 0.0935 ms 89.7% triton_mm_2351 0.0945 ms 88.8% triton_mm_2353 0.0960 ms 87.5% triton_mm_2352 0.1028 ms 81.6% triton_mm_2356 0.1057 ms 79.4% triton_mm_2354 0.1228 ms 68.4% SingleProcess AUTOTUNE benchmarking takes 2.4200 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2246 ms 100.0% triton_mm_2397 0.2265 ms 99.2% triton_mm_2396 0.2316 ms 97.0% triton_mm_2398 0.2448 ms 91.8% triton_mm_2393 0.2593 ms 86.6% triton_mm_2391 0.2644 ms 84.9% triton_mm_2389 0.2664 ms 84.3% triton_mm_2392 0.2909 ms 77.2% triton_mm_2390 0.2938 ms 76.4% triton_mm_2394 0.3003 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.5172 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2901 ms 100.0% triton_mm_2487 0.3005 ms 96.5% triton_mm_2488 0.3006 ms 96.5% triton_mm_2489 0.3038 ms 95.5% triton_mm_2482 0.3569 ms 81.3% triton_mm_2480 0.3590 ms 80.8% triton_mm_2484 0.3683 ms 78.8% triton_mm_2483 0.3919 ms 74.0% triton_mm_2485 0.4032 ms 72.0% triton_mm_2481 0.4082 ms 71.1% SingleProcess AUTOTUNE benchmarking takes 2.5464 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_2525 0.2933 ms 100.0% mm 0.2942 ms 99.7% triton_mm_2526 0.2984 ms 98.3% triton_mm_2527 0.3036 ms 96.6% triton_mm_2522 0.3510 ms 83.6% triton_mm_2518 0.3581 ms 81.9% triton_mm_2520 0.3637 ms 80.7% triton_mm_2521 0.3902 ms 75.2% triton_mm_2523 0.3961 ms 74.1% triton_mm_2519 0.4038 ms 72.6% SingleProcess AUTOTUNE benchmarking takes 2.5383 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0265 ms 100.0% triton_mm_2422 0.0266 ms 99.5% triton_mm_2426 0.0318 ms 83.4% triton_mm_2430 0.0399 ms 66.4% triton_mm_2434 0.0518 ms 51.1% triton_mm_2421 0.0524 ms 50.6% triton_mm_2420 0.0559 ms 47.4% triton_mm_2425 0.0593 ms 44.7% triton_mm_2429 0.0623 ms 42.5% triton_mm_2419 0.0651 ms 40.7% SingleProcess AUTOTUNE benchmarking takes 2.0974 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) mm 0.0838 ms 100.0% triton_mm_2545 0.0840 ms 99.7% triton_mm_2544 0.0846 ms 99.0% triton_mm_2541 0.0920 ms 91.1% triton_mm_2546 0.0944 ms 88.8% triton_mm_2537 0.0954 ms 87.8% triton_mm_2539 0.0960 ms 87.3% triton_mm_2538 0.1028 ms 81.5% triton_mm_2542 0.1057 ms 79.3% triton_mm_2540 0.1227 ms 68.3% SingleProcess AUTOTUNE benchmarking takes 2.4271 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2254 ms 100.0% triton_mm_2583 0.2267 ms 99.4% triton_mm_2582 0.2312 ms 97.5% triton_mm_2584 0.2441 ms 92.3% triton_mm_2579 0.2598 ms 86.8% triton_mm_2577 0.2652 ms 85.0% triton_mm_2575 0.2665 ms 84.6% triton_mm_2578 0.2913 ms 77.4% triton_mm_2576 0.2940 ms 76.7% triton_mm_2580 0.2997 ms 75.2% SingleProcess AUTOTUNE benchmarking takes 2.5168 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0268 ms 100.0% triton_mm_2608 0.0276 ms 97.2% triton_mm_2612 0.0318 ms 84.4% triton_mm_2616 0.0399 ms 67.2% triton_mm_2607 0.0516 ms 52.0% triton_mm_2620 0.0518 ms 51.7% triton_mm_2606 0.0558 ms 48.0% triton_mm_2611 0.0601 ms 44.6% triton_mm_2615 0.0623 ms 43.1% triton_mm_2605 0.0651 ms 41.2% SingleProcess AUTOTUNE benchmarking takes 2.0994 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_2731 0.0840 ms 100.0% mm 0.0841 ms 99.8% triton_mm_2730 0.0847 ms 99.1% triton_mm_2727 0.0929 ms 90.4% triton_mm_2732 0.0936 ms 89.7% triton_mm_2723 0.0946 ms 88.8% triton_mm_2725 0.0962 ms 87.2% triton_mm_2724 0.1032 ms 81.3% triton_mm_2728 0.1049 ms 80.1% triton_mm_2726 0.1227 ms 68.5% SingleProcess AUTOTUNE benchmarking takes 2.4162 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2256 ms 100.0% triton_mm_2769 0.2271 ms 99.3% triton_mm_2768 0.2325 ms 97.1% triton_mm_2770 0.2440 ms 92.5% triton_mm_2765 0.2599 ms 86.8% triton_mm_2761 0.2659 ms 84.8% triton_mm_2763 0.2660 ms 84.8% triton_mm_2764 0.2924 ms 77.2% triton_mm_2762 0.2944 ms 76.6% triton_mm_2766 0.2998 ms 75.3% SingleProcess AUTOTUNE benchmarking takes 2.5092 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2794 0.0266 ms 100.0% mm 0.0268 ms 99.3% triton_mm_2798 0.0307 ms 86.8% triton_mm_2802 0.0399 ms 66.8% triton_mm_2806 0.0518 ms 51.4% triton_mm_2793 0.0523 ms 50.9% triton_mm_2792 0.0558 ms 47.7% triton_mm_2797 0.0593 ms 44.9% triton_mm_2801 0.0624 ms 42.7% triton_mm_2791 0.0658 ms 40.4% SingleProcess AUTOTUNE benchmarking takes 2.0847 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2230 ms 100.0% triton_mm_2955 0.2270 ms 98.3% triton_mm_2954 0.2314 ms 96.4% triton_mm_2956 0.2442 ms 91.3% triton_mm_2951 0.2595 ms 85.9% triton_mm_2949 0.2660 ms 83.8% triton_mm_2947 0.2665 ms 83.7% triton_mm_2948 0.2925 ms 76.2% triton_mm_2950 0.2927 ms 76.2% triton_mm_2952 0.2993 ms 74.5% SingleProcess AUTOTUNE benchmarking takes 2.5090 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2234 ms 100.0% triton_mm_3141 0.2261 ms 98.8% triton_mm_3140 0.2322 ms 96.2% triton_mm_3142 0.2444 ms 91.4% triton_mm_3137 0.2601 ms 85.9% triton_mm_3135 0.2647 ms 84.4% triton_mm_3133 0.2655 ms 84.1% triton_mm_3136 0.2902 ms 77.0% triton_mm_3134 0.2932 ms 76.2% triton_mm_3138 0.2996 ms 74.5% SingleProcess AUTOTUNE benchmarking takes 2.5026 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2222 ms 100.0% triton_mm_3327 0.2263 ms 98.2% triton_mm_3326 0.2308 ms 96.3% triton_mm_3328 0.2432 ms 91.4% triton_mm_3323 0.2595 ms 85.6% triton_mm_3321 0.2646 ms 84.0% triton_mm_3319 0.2665 ms 83.4% triton_mm_3322 0.2919 ms 76.1% triton_mm_3320 0.2923 ms 76.0% triton_mm_3324 0.2992 ms 74.3% SingleProcess AUTOTUNE benchmarking takes 2.5054 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6840, 6840x64) mm 0.0291 ms 100.0% triton_mm_4874 0.0442 ms 65.8% triton_mm_4878 0.0478 ms 60.8% triton_mm_4882 0.0488 ms 59.6% triton_mm_4872 0.0588 ms 49.5% triton_mm_4873 0.0633 ms 45.9% triton_mm_4881 0.0683 ms 42.6% triton_mm_4871 0.0706 ms 41.2% triton_mm_4877 0.0729 ms 39.9% triton_mm_4880 0.0781 ms 37.2% SingleProcess AUTOTUNE benchmarking takes 7.7809 seconds and 0.0000 seconds precompiling /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): AUTOTUNE addmm(6840x1088, 6840x64, 64x1088) triton_mm_13 0.0320 ms 100.0% triton_mm_9 0.0336 ms 95.4% triton_mm_16 0.0337 ms 95.1% triton_mm_8 0.0342 ms 93.7% triton_mm_6 0.0359 ms 89.3% triton_mm_5 0.0364 ms 88.0% triton_mm_7 0.0372 ms 86.2% triton_mm_12 0.0376 ms 85.3% triton_mm_11 0.0383 ms 83.6% triton_mm_10 0.0387 ms 82.7% SingleProcess AUTOTUNE benchmarking takes 2.5364 seconds and 0.0168 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_35 0.0214 ms 100.0% triton_mm_26 0.0227 ms 94.2% triton_mm_27 0.0229 ms 93.6% triton_mm_23 0.0230 ms 92.9% triton_mm_31 0.0230 ms 92.9% triton_mm_30 0.0237 ms 90.2% bias_addmm 0.0241 ms 89.0% triton_mm_21 0.0241 ms 88.8% triton_mm_22 0.0245 ms 87.2% triton_mm_20 0.0246 ms 86.9% SingleProcess AUTOTUNE benchmarking takes 2.1990 seconds and 0.0008 seconds precompiling AUTOTUNE addmm(6864x3264, 6864x1088, 1088x3264) triton_mm_52 0.2206 ms 100.0% triton_mm_53 0.2263 ms 97.5% triton_mm_54 0.2544 ms 86.7% triton_mm_47 0.2625 ms 84.0% triton_mm_49 0.2708 ms 81.5% triton_mm_45 0.2733 ms 80.7% bias_addmm 0.2999 ms 73.6% triton_mm_46 0.3053 ms 72.3% triton_mm_50 0.3085 ms 71.5% addmm 0.3233 ms 68.2% SingleProcess AUTOTUNE benchmarking takes 2.5484 seconds and 0.0217 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x1088, 1088x1088) triton_mm_72 0.0903 ms 100.0% triton_mm_71 0.0907 ms 99.6% bias_addmm 0.0977 ms 92.4% triton_mm_68 0.0981 ms 92.0% triton_mm_66 0.0991 ms 91.1% triton_mm_73 0.0992 ms 91.0% triton_mm_64 0.1023 ms 88.3% addmm 0.1109 ms 81.4% triton_mm_65 0.1110 ms 81.3% triton_mm_69 0.1132 ms 79.8% SingleProcess AUTOTUNE benchmarking takes 2.5081 seconds and 0.0010 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_90 0.2898 ms 100.0% triton_mm_91 0.2941 ms 98.6% bias_addmm 0.3258 ms 89.0% triton_mm_92 0.3330 ms 87.0% triton_mm_85 0.3389 ms 85.5% triton_mm_83 0.3588 ms 80.8% triton_mm_87 0.3612 ms 80.2% addmm 0.3647 ms 79.5% triton_mm_84 0.3998 ms 72.5% triton_mm_88 0.4097 ms 70.7% SingleProcess AUTOTUNE benchmarking takes 2.7066 seconds and 0.0434 seconds precompiling AUTOTUNE addmm(6864x1088, 6864x4352, 4352x1088) bias_addmm 0.2880 ms 100.0% triton_mm_110 0.3048 ms 94.5% triton_mm_109 0.3149 ms 91.5% addmm 0.3243 ms 88.8% triton_mm_111 0.3260 ms 88.4% triton_mm_106 0.3537 ms 81.4% triton_mm_104 0.3549 ms 81.2% triton_mm_102 0.3705 ms 77.7% triton_mm_105 0.3706 ms 77.7% triton_mm_107 0.4049 ms 71.1% SingleProcess AUTOTUNE benchmarking takes 2.6513 seconds and 0.0009 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_128 0.0214 ms 100.0% triton_mm_119 0.0227 ms 94.4% triton_mm_116 0.0228 ms 94.0% triton_mm_120 0.0228 ms 93.8% triton_mm_124 0.0231 ms 92.5% bias_addmm 0.0241 ms 88.7% triton_mm_114 0.0241 ms 88.7% triton_mm_115 0.0245 ms 87.1% triton_mm_113 0.0246 ms 87.0% triton_mm_123 0.0249 ms 85.9% SingleProcess AUTOTUNE benchmarking takes 2.1958 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_183 0.2906 ms 100.0% triton_mm_184 0.2937 ms 99.0% bias_addmm 0.3250 ms 89.4% triton_mm_185 0.3317 ms 87.6% triton_mm_178 0.3398 ms 85.5% triton_mm_176 0.3589 ms 81.0% triton_mm_180 0.3608 ms 80.5% addmm 0.3799 ms 76.5% triton_mm_177 0.3978 ms 73.1% triton_mm_181 0.4084 ms 71.2% SingleProcess AUTOTUNE benchmarking takes 2.6484 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(24x6528, 24x1088, 1088x6528) triton_mm_221 0.0214 ms 100.0% triton_mm_213 0.0228 ms 93.8% triton_mm_209 0.0228 ms 93.7% triton_mm_217 0.0230 ms 93.2% triton_mm_212 0.0237 ms 90.3% triton_mm_216 0.0238 ms 90.0% triton_mm_206 0.0244 ms 87.7% triton_mm_207 0.0252 ms 85.1% triton_mm_208 0.0256 ms 83.6% bias_addmm 0.0260 ms 82.2% SingleProcess AUTOTUNE benchmarking takes 2.1991 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_276 0.2905 ms 100.0% triton_mm_277 0.2932 ms 99.1% bias_addmm 0.3253 ms 89.3% triton_mm_278 0.3314 ms 87.7% triton_mm_271 0.3387 ms 85.8% triton_mm_269 0.3574 ms 81.3% triton_mm_273 0.3605 ms 80.6% addmm 0.3659 ms 79.4% triton_mm_270 0.3977 ms 73.1% triton_mm_274 0.4069 ms 71.4% SingleProcess AUTOTUNE benchmarking takes 2.6450 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_369 0.2900 ms 100.0% triton_mm_370 0.2947 ms 98.4% bias_addmm 0.3253 ms 89.2% triton_mm_371 0.3345 ms 86.7% triton_mm_364 0.3395 ms 85.4% triton_mm_362 0.3574 ms 81.1% triton_mm_366 0.3617 ms 80.2% addmm 0.3643 ms 79.6% triton_mm_363 0.3952 ms 73.4% triton_mm_367 0.4084 ms 71.0% SingleProcess AUTOTUNE benchmarking takes 2.6469 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x4352, 6864x1088, 1088x4352) triton_mm_462 0.2895 ms 100.0% triton_mm_463 0.2941 ms 98.4% bias_addmm 0.3262 ms 88.8% triton_mm_464 0.3314 ms 87.4% triton_mm_457 0.3389 ms 85.4% triton_mm_455 0.3576 ms 80.9% triton_mm_459 0.3612 ms 80.2% triton_mm_456 0.3979 ms 72.8% triton_mm_460 0.4097 ms 70.7% addmm 0.4180 ms 69.2% SingleProcess AUTOTUNE benchmarking takes 2.6537 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x1088, 1088x2176) triton_mm_1604 0.0134 ms 100.0% triton_mm_1608 0.0141 ms 95.0% triton_mm_1612 0.0164 ms 81.5% triton_mm_1616 0.0166 ms 80.7% mm 0.0168 ms 79.5% triton_mm_1602 0.0175 ms 76.3% triton_mm_1603 0.0178 ms 75.0% triton_mm_1607 0.0184 ms 72.6% triton_mm_1601 0.0186 ms 71.8% triton_mm_1611 0.0199 ms 67.3% SingleProcess AUTOTUNE benchmarking takes 7.6984 seconds and 0.0000 seconds precompiling AUTOTUNE addmm(6864x32768, 6864x1088, 1088x32768) bias_addmm 2.1867 ms 100.0% triton_mm_1634 2.2019 ms 99.3% triton_mm_1633 2.2285 ms 98.1% triton_mm_1635 2.4913 ms 87.8% triton_mm_1628 2.6270 ms 83.2% triton_mm_1630 2.7377 ms 79.9% triton_mm_1626 2.8268 ms 77.4% addmm 2.9103 ms 75.1% triton_mm_1627 3.0875 ms 70.8% triton_mm_1631 3.0937 ms 70.7% SingleProcess AUTOTUNE benchmarking takes 11.7217 seconds and 0.0000 seconds precompiling AUTOTUNE mm(32768x6864, 6864x1088) triton_mm_1671 2.0466 ms 100.0% triton_mm_1672 2.1252 ms 96.3% mm 2.2375 ms 91.5% triton_mm_1673 2.2545 ms 90.8% triton_mm_1668 2.5760 ms 79.4% triton_mm_1664 2.6382 ms 77.6% triton_mm_1666 2.6482 ms 77.3% triton_mm_1669 2.7332 ms 74.9% triton_mm_1667 2.8164 ms 72.7% triton_mm_1665 2.9252 ms 70.0% SingleProcess AUTOTUNE benchmarking takes 10.0483 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2882 ms 100.0% triton_mm_1744 0.2915 ms 98.9% triton_mm_1743 0.2929 ms 98.4% triton_mm_1745 0.2953 ms 97.6% triton_mm_1738 0.3461 ms 83.3% triton_mm_1736 0.3503 ms 82.3% triton_mm_1740 0.3568 ms 80.8% triton_mm_1739 0.3723 ms 77.4% triton_mm_1741 0.3930 ms 73.3% triton_mm_1737 0.3984 ms 72.3% SingleProcess AUTOTUNE benchmarking takes 9.8694 seconds and 0.0000 seconds precompiling AUTOTUNE mm(2176x24, 24x1088) triton_mm_1696 0.0113 ms 100.0% triton_mm_1701 0.0115 ms 98.1% triton_mm_1699 0.0116 ms 97.8% triton_mm_1703 0.0116 ms 97.2% triton_mm_1700 0.0119 ms 95.1% triton_mm_1706 0.0120 ms 94.1% triton_mm_1704 0.0121 ms 93.4% triton_mm_1698 0.0122 ms 92.8% triton_mm_1691 0.0122 ms 92.7% triton_mm_1702 0.0122 ms 92.5% SingleProcess AUTOTUNE benchmarking takes 6.0052 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x2176, 2176x1088) triton_mm_1678 0.0155 ms 100.0% mm 0.0173 ms 89.3% triton_mm_1682 0.0178 ms 87.2% triton_mm_1686 0.0215 ms 71.9% triton_mm_1677 0.0223 ms 69.4% triton_mm_1690 0.0238 ms 65.0% triton_mm_1676 0.0241 ms 64.2% triton_mm_1685 0.0256 ms 60.4% triton_mm_1675 0.0263 ms 59.0% triton_mm_1681 0.0263 ms 58.9% SingleProcess AUTOTUNE benchmarking takes 7.6580 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x32768, 32768x1088) mm 2.0708 ms 100.0% triton_mm_1652 2.2486 ms 92.1% triton_mm_1653 2.2630 ms 91.5% triton_mm_1654 2.3265 ms 89.0% triton_mm_1645 2.7183 ms 76.2% triton_mm_1647 2.7477 ms 75.4% triton_mm_1649 2.7648 ms 74.9% triton_mm_1648 2.8028 ms 73.9% triton_mm_1646 2.8764 ms 72.0% triton_mm_1650 2.9194 ms 70.9% SingleProcess AUTOTUNE benchmarking takes 10.1712 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1724 0.2733 ms 100.0% mm 0.2743 ms 99.6% triton_mm_1725 0.2768 ms 98.7% triton_mm_1726 0.3134 ms 87.2% triton_mm_1717 0.3353 ms 81.5% triton_mm_1719 0.3407 ms 80.2% triton_mm_1721 0.3453 ms 79.1% triton_mm_1718 0.3632 ms 75.3% triton_mm_1722 0.3852 ms 71.0% triton_mm_1720 0.4246 ms 64.4% SingleProcess AUTOTUNE benchmarking takes 2.5278 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) triton_mm_1839 0.2197 ms 100.0% triton_mm_1838 0.2269 ms 96.8% triton_mm_1840 0.2375 ms 92.5% triton_mm_1835 0.2521 ms 87.2% triton_mm_1833 0.2557 ms 85.9% mm 0.2569 ms 85.5% triton_mm_1831 0.2575 ms 85.3% triton_mm_1834 0.2822 ms 77.9% triton_mm_1832 0.2850 ms 77.1% triton_mm_1836 0.2908 ms 75.6% SingleProcess AUTOTUNE benchmarking takes 9.8628 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2882 ms 100.0% triton_mm_1781 0.2886 ms 99.9% triton_mm_1782 0.2950 ms 97.7% triton_mm_1783 0.3008 ms 95.8% triton_mm_1778 0.3444 ms 83.7% triton_mm_1774 0.3514 ms 82.0% triton_mm_1776 0.3554 ms 81.1% triton_mm_1777 0.3815 ms 75.6% triton_mm_1779 0.3908 ms 73.7% triton_mm_1775 0.3991 ms 72.2% SingleProcess AUTOTUNE benchmarking takes 2.5415 seconds and 0.0012 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2236 ms 100.0% triton_mm_1857 0.2752 ms 81.2% triton_mm_1852 0.2776 ms 80.5% triton_mm_1854 0.2793 ms 80.1% triton_mm_1850 0.2811 ms 79.5% triton_mm_1858 0.2817 ms 79.4% triton_mm_1859 0.2857 ms 78.3% triton_mm_1853 0.3071 ms 72.8% triton_mm_1855 0.3202 ms 69.8% triton_mm_1851 0.3259 ms 68.6% SingleProcess AUTOTUNE benchmarking takes 9.9161 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0968 ms 100.0% triton_mm_1821 0.0997 ms 97.1% triton_mm_1814 0.1046 ms 92.5% triton_mm_1820 0.1088 ms 88.9% triton_mm_1811 0.1186 ms 81.6% triton_mm_1812 0.1194 ms 81.1% triton_mm_1816 0.1199 ms 80.7% triton_mm_1819 0.1225 ms 79.0% triton_mm_1815 0.1260 ms 76.8% triton_mm_1817 0.1314 ms 73.6% SingleProcess AUTOTUNE benchmarking takes 2.4202 seconds and 0.0009 seconds precompiling AUTOTUNE mm(6864x4352, 4352x1088) mm 0.2964 ms 100.0% triton_mm_1762 0.3018 ms 98.2% triton_mm_1763 0.3094 ms 95.8% triton_mm_1764 0.3196 ms 92.7% triton_mm_1759 0.3409 ms 87.0% triton_mm_1757 0.3484 ms 85.1% triton_mm_1755 0.3490 ms 84.9% triton_mm_1758 0.3665 ms 80.9% triton_mm_1756 0.3841 ms 77.2% triton_mm_1760 0.3947 ms 75.1% SingleProcess AUTOTUNE benchmarking takes 2.5292 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x1088, 1088x1088) triton_mm_1801 0.0843 ms 100.0% mm 0.0851 ms 99.1% triton_mm_1800 0.0858 ms 98.2% triton_mm_1797 0.0935 ms 90.2% triton_mm_1802 0.0948 ms 88.9% triton_mm_1793 0.0954 ms 88.4% triton_mm_1795 0.0965 ms 87.4% triton_mm_1794 0.1031 ms 81.8% triton_mm_1798 0.1052 ms 80.2% triton_mm_1796 0.1233 ms 68.4% SingleProcess AUTOTUNE benchmarking takes 2.4099 seconds and 0.0010 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_1910 0.2721 ms 100.0% mm 0.2740 ms 99.3% triton_mm_1911 0.2765 ms 98.4% triton_mm_1912 0.3146 ms 86.5% triton_mm_1903 0.3371 ms 80.7% triton_mm_1905 0.3405 ms 79.9% triton_mm_1907 0.3440 ms 79.1% triton_mm_1904 0.3608 ms 75.4% triton_mm_1908 0.3848 ms 70.7% triton_mm_1906 0.4242 ms 64.1% SingleProcess AUTOTUNE benchmarking takes 2.5231 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_1889 0.0173 ms 100.0% triton_mm_1888 0.0174 ms 99.3% triton_mm_1886 0.0175 ms 99.1% triton_mm_1882 0.0177 ms 97.8% triton_mm_1885 0.0183 ms 94.7% triton_mm_1887 0.0184 ms 94.3% triton_mm_1892 0.0184 ms 94.3% triton_mm_1883 0.0186 ms 93.1% triton_mm_1884 0.0192 ms 90.3% triton_mm_1893 0.0193 ms 89.7% SingleProcess AUTOTUNE benchmarking takes 5.9205 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x4352) mm 0.2875 ms 100.0% triton_mm_1930 0.2958 ms 97.2% triton_mm_1929 0.2980 ms 96.5% triton_mm_1931 0.3006 ms 95.6% triton_mm_1924 0.3524 ms 81.6% triton_mm_1922 0.3543 ms 81.1% triton_mm_1926 0.3617 ms 79.5% triton_mm_1925 0.3792 ms 75.8% triton_mm_1927 0.4002 ms 71.8% triton_mm_1923 0.4045 ms 71.1% SingleProcess AUTOTUNE benchmarking takes 2.5367 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) triton_mm_1967 0.2894 ms 100.0% mm 0.2908 ms 99.5% triton_mm_1968 0.2948 ms 98.2% triton_mm_1969 0.3007 ms 96.2% triton_mm_1964 0.3451 ms 83.9% triton_mm_1960 0.3545 ms 81.6% triton_mm_1962 0.3562 ms 81.2% triton_mm_1963 0.3820 ms 75.7% triton_mm_1965 0.3919 ms 73.8% triton_mm_1961 0.3994 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5414 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0268 ms 100.0% triton_mm_1864 0.0278 ms 96.5% triton_mm_1868 0.0308 ms 87.2% triton_mm_1872 0.0409 ms 65.6% triton_mm_1876 0.0519 ms 51.7% triton_mm_1863 0.0525 ms 51.1% triton_mm_1862 0.0564 ms 47.5% triton_mm_1867 0.0606 ms 44.2% triton_mm_1871 0.0628 ms 42.7% triton_mm_1861 0.0662 ms 40.5% SingleProcess AUTOTUNE benchmarking takes 6.6142 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2222 ms 100.0% triton_mm_2043 0.2771 ms 80.2% triton_mm_2038 0.2836 ms 78.4% triton_mm_2044 0.2840 ms 78.3% triton_mm_2040 0.2840 ms 78.2% triton_mm_2036 0.2847 ms 78.1% triton_mm_2045 0.2891 ms 76.9% triton_mm_2039 0.3104 ms 71.6% triton_mm_2041 0.3248 ms 68.4% triton_mm_2037 0.3288 ms 67.6% SingleProcess AUTOTUNE benchmarking takes 2.5075 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0968 ms 100.0% triton_mm_2007 0.0991 ms 97.7% triton_mm_2000 0.1058 ms 91.5% triton_mm_2006 0.1088 ms 89.0% triton_mm_1997 0.1179 ms 82.1% triton_mm_2002 0.1202 ms 80.5% triton_mm_1998 0.1205 ms 80.3% triton_mm_2005 0.1223 ms 79.1% triton_mm_2001 0.1274 ms 76.0% triton_mm_2003 0.1316 ms 73.5% SingleProcess AUTOTUNE benchmarking takes 2.4132 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2218 ms 100.0% triton_mm_2025 0.2251 ms 98.5% triton_mm_2024 0.2296 ms 96.6% triton_mm_2026 0.2419 ms 91.7% triton_mm_2021 0.2564 ms 86.5% triton_mm_2019 0.2616 ms 84.8% triton_mm_2017 0.2630 ms 84.3% triton_mm_2020 0.2896 ms 76.6% triton_mm_2018 0.2902 ms 76.4% triton_mm_2022 0.2961 ms 74.9% SingleProcess AUTOTUNE benchmarking takes 2.5121 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2096 0.2734 ms 100.0% mm 0.2744 ms 99.6% triton_mm_2097 0.2777 ms 98.5% triton_mm_2098 0.3161 ms 86.5% triton_mm_2089 0.3383 ms 80.8% triton_mm_2091 0.3421 ms 79.9% triton_mm_2093 0.3463 ms 79.0% triton_mm_2090 0.3643 ms 75.0% triton_mm_2094 0.3860 ms 70.8% triton_mm_2092 0.4247 ms 64.4% SingleProcess AUTOTUNE benchmarking takes 2.5241 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2069 0.0172 ms 100.0% triton_mm_2072 0.0175 ms 98.7% triton_mm_2074 0.0175 ms 98.7% triton_mm_2076 0.0183 ms 94.1% triton_mm_2070 0.0185 ms 93.4% triton_mm_2071 0.0185 ms 93.4% triton_mm_2075 0.0185 ms 93.3% triton_mm_2068 0.0187 ms 92.4% triton_mm_2073 0.0192 ms 89.8% triton_mm_2079 0.0193 ms 89.5% SingleProcess AUTOTUNE benchmarking takes 2.1346 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2907 ms 100.0% triton_mm_2153 0.2914 ms 99.8% triton_mm_2154 0.2956 ms 98.3% triton_mm_2155 0.3012 ms 96.5% triton_mm_2150 0.3461 ms 84.0% triton_mm_2146 0.3541 ms 82.1% triton_mm_2148 0.3564 ms 81.6% triton_mm_2149 0.3829 ms 75.9% triton_mm_2151 0.3936 ms 73.9% triton_mm_2147 0.4019 ms 72.3% SingleProcess AUTOTUNE benchmarking takes 2.5322 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2050 0.0265 ms 100.0% mm 0.0270 ms 98.1% triton_mm_2054 0.0320 ms 82.7% triton_mm_2058 0.0400 ms 66.2% triton_mm_2062 0.0509 ms 52.0% triton_mm_2049 0.0516 ms 51.3% triton_mm_2048 0.0555 ms 47.7% triton_mm_2053 0.0605 ms 43.8% triton_mm_2057 0.0628 ms 42.2% triton_mm_2047 0.0660 ms 40.2% SingleProcess AUTOTUNE benchmarking takes 2.0700 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2232 ms 100.0% triton_mm_2229 0.2775 ms 80.4% triton_mm_2224 0.2834 ms 78.8% triton_mm_2226 0.2840 ms 78.6% triton_mm_2230 0.2855 ms 78.2% triton_mm_2222 0.2885 ms 77.4% triton_mm_2231 0.2897 ms 77.1% triton_mm_2225 0.3127 ms 71.4% triton_mm_2227 0.3252 ms 68.7% triton_mm_2223 0.3316 ms 67.3% SingleProcess AUTOTUNE benchmarking takes 2.5081 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0969 ms 100.0% triton_mm_2193 0.0986 ms 98.3% triton_mm_2186 0.1058 ms 91.6% triton_mm_2192 0.1080 ms 89.7% triton_mm_2183 0.1187 ms 81.6% triton_mm_2188 0.1202 ms 80.6% triton_mm_2184 0.1203 ms 80.5% triton_mm_2191 0.1216 ms 79.7% triton_mm_2187 0.1274 ms 76.1% triton_mm_2189 0.1315 ms 73.7% SingleProcess AUTOTUNE benchmarking takes 3.3956 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2227 ms 100.0% triton_mm_2211 0.2264 ms 98.4% triton_mm_2210 0.2310 ms 96.4% triton_mm_2212 0.2430 ms 91.6% triton_mm_2207 0.2570 ms 86.7% triton_mm_2205 0.2622 ms 84.9% triton_mm_2203 0.2642 ms 84.3% triton_mm_2206 0.2895 ms 76.9% triton_mm_2204 0.2912 ms 76.5% triton_mm_2208 0.2977 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.5109 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2282 0.2751 ms 100.0% mm 0.2756 ms 99.8% triton_mm_2283 0.2791 ms 98.6% triton_mm_2284 0.3153 ms 87.3% triton_mm_2275 0.3382 ms 81.3% triton_mm_2277 0.3419 ms 80.5% triton_mm_2279 0.3460 ms 79.5% triton_mm_2276 0.3635 ms 75.7% triton_mm_2280 0.3864 ms 71.2% triton_mm_2278 0.4232 ms 65.0% SingleProcess AUTOTUNE benchmarking takes 2.5276 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6528x24, 24x1088) triton_mm_2261 0.0173 ms 100.0% triton_mm_2255 0.0182 ms 95.2% triton_mm_2257 0.0185 ms 93.4% triton_mm_2260 0.0186 ms 92.7% triton_mm_2254 0.0187 ms 92.6% triton_mm_2258 0.0187 ms 92.6% triton_mm_2259 0.0190 ms 90.9% triton_mm_2262 0.0192 ms 90.1% triton_mm_2265 0.0193 ms 89.7% triton_mm_2264 0.0193 ms 89.6% SingleProcess AUTOTUNE benchmarking takes 2.1510 seconds and 0.0000 seconds precompiling AUTOTUNE mm(4352x6864, 6864x1088) mm 0.2914 ms 100.0% triton_mm_2339 0.2918 ms 99.9% triton_mm_2340 0.2958 ms 98.5% triton_mm_2341 0.3018 ms 96.6% triton_mm_2336 0.3473 ms 83.9% triton_mm_2332 0.3553 ms 82.0% triton_mm_2334 0.3582 ms 81.4% triton_mm_2335 0.3843 ms 75.8% triton_mm_2337 0.3945 ms 73.9% triton_mm_2333 0.4027 ms 72.4% SingleProcess AUTOTUNE benchmarking takes 2.5424 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2236 0.0265 ms 100.0% mm 0.0269 ms 98.5% triton_mm_2240 0.0308 ms 85.9% triton_mm_2244 0.0400 ms 66.2% triton_mm_2248 0.0514 ms 51.6% triton_mm_2235 0.0516 ms 51.3% triton_mm_2234 0.0563 ms 47.1% triton_mm_2239 0.0605 ms 43.8% triton_mm_2243 0.0621 ms 42.7% triton_mm_2233 0.0661 ms 40.1% SingleProcess AUTOTUNE benchmarking takes 2.0993 seconds and 0.0000 seconds precompiling AUTOTUNE mm(3264x6864, 6864x1088) mm 0.2225 ms 100.0% triton_mm_2415 0.2793 ms 79.7% triton_mm_2412 0.2837 ms 78.4% triton_mm_2410 0.2838 ms 78.4% triton_mm_2416 0.2863 ms 77.7% triton_mm_2408 0.2873 ms 77.4% triton_mm_2417 0.2909 ms 76.5% triton_mm_2411 0.3127 ms 71.2% triton_mm_2413 0.3259 ms 68.3% triton_mm_2409 0.3309 ms 67.2% SingleProcess AUTOTUNE benchmarking takes 2.5157 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6864, 6864x1088) mm 0.0977 ms 100.0% triton_mm_2379 0.0992 ms 98.5% triton_mm_2372 0.1050 ms 93.1% triton_mm_2378 0.1088 ms 89.8% triton_mm_2369 0.1186 ms 82.4% triton_mm_2370 0.1196 ms 81.7% triton_mm_2374 0.1210 ms 80.8% triton_mm_2377 0.1223 ms 79.9% triton_mm_2373 0.1265 ms 77.3% triton_mm_2375 0.1306 ms 74.8% SingleProcess AUTOTUNE benchmarking takes 2.4276 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x3264, 3264x1088) mm 0.2234 ms 100.0% triton_mm_2397 0.2255 ms 99.1% triton_mm_2396 0.2317 ms 96.4% triton_mm_2398 0.2439 ms 91.6% triton_mm_2393 0.2588 ms 86.3% triton_mm_2389 0.2629 ms 85.0% triton_mm_2391 0.2632 ms 84.9% triton_mm_2392 0.2896 ms 77.1% triton_mm_2390 0.2926 ms 76.3% triton_mm_2394 0.2992 ms 74.7% SingleProcess AUTOTUNE benchmarking takes 2.5174 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) mm 0.2749 ms 100.0% triton_mm_2468 0.2760 ms 99.6% triton_mm_2469 0.2780 ms 98.9% triton_mm_2470 0.3147 ms 87.4% triton_mm_2461 0.3392 ms 81.0% triton_mm_2463 0.3432 ms 80.1% triton_mm_2465 0.3440 ms 79.9% triton_mm_2462 0.3650 ms 75.3% triton_mm_2466 0.3862 ms 71.2% triton_mm_2464 0.4245 ms 64.8% SingleProcess AUTOTUNE benchmarking takes 2.5363 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2422 0.0265 ms 100.0% mm 0.0268 ms 98.7% triton_mm_2426 0.0321 ms 82.5% triton_mm_2430 0.0414 ms 64.0% triton_mm_2434 0.0510 ms 52.0% triton_mm_2421 0.0525 ms 50.5% triton_mm_2420 0.0555 ms 47.7% triton_mm_2425 0.0597 ms 44.4% triton_mm_2429 0.0630 ms 42.1% triton_mm_2419 0.0653 ms 40.6% SingleProcess AUTOTUNE benchmarking takes 2.0999 seconds and 0.0000 seconds precompiling AUTOTUNE mm(6864x1088, 1088x4352) triton_mm_2654 0.2736 ms 100.0% mm 0.2744 ms 99.7% triton_mm_2655 0.2769 ms 98.8% triton_mm_2656 0.3134 ms 87.3% triton_mm_2647 0.3385 ms 80.8% triton_mm_2649 0.3415 ms 80.1% triton_mm_2651 0.3443 ms 79.5% triton_mm_2648 0.3635 ms 75.3% triton_mm_2652 0.3845 ms 71.2% triton_mm_2650 0.4258 ms 64.3% SingleProcess AUTOTUNE benchmarking takes 2.5365 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) mm 0.0268 ms 100.0% triton_mm_2608 0.0276 ms 97.3% triton_mm_2612 0.0308 ms 87.0% triton_mm_2616 0.0409 ms 65.7% triton_mm_2620 0.0514 ms 52.2% triton_mm_2607 0.0525 ms 51.2% triton_mm_2606 0.0563 ms 47.7% triton_mm_2611 0.0606 ms 44.3% triton_mm_2615 0.0621 ms 43.2% triton_mm_2605 0.0660 ms 40.7% SingleProcess AUTOTUNE benchmarking takes 2.0865 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2794 0.0265 ms 100.0% mm 0.0271 ms 97.7% triton_mm_2798 0.0308 ms 86.0% triton_mm_2802 0.0413 ms 64.1% triton_mm_2806 0.0509 ms 52.1% triton_mm_2793 0.0516 ms 51.3% triton_mm_2792 0.0563 ms 47.0% triton_mm_2797 0.0597 ms 44.4% triton_mm_2801 0.0630 ms 42.1% triton_mm_2791 0.0653 ms 40.6% SingleProcess AUTOTUNE benchmarking takes 2.0958 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_2980 0.0265 ms 100.0% mm 0.0272 ms 97.4% triton_mm_2984 0.0321 ms 82.4% triton_mm_2988 0.0405 ms 65.4% triton_mm_2992 0.0509 ms 52.0% triton_mm_2979 0.0526 ms 50.3% triton_mm_2978 0.0555 ms 47.7% triton_mm_2983 0.0597 ms 44.3% triton_mm_2987 0.0621 ms 42.6% triton_mm_2977 0.0661 ms 40.0% SingleProcess AUTOTUNE benchmarking takes 2.0913 seconds and 0.0000 seconds precompiling AUTOTUNE mm(24x6528, 6528x1088) triton_mm_3166 0.0265 ms 100.0% mm 0.0270 ms 98.1% triton_mm_3170 0.0319 ms 83.0% triton_mm_3174 0.0400 ms 66.2% triton_mm_3178 0.0520 ms 50.9% triton_mm_3165 0.0524 ms 50.6% triton_mm_3164 0.0563 ms 47.0% triton_mm_3169 0.0605 ms 43.8% triton_mm_3173 0.0628 ms 42.2% triton_mm_3163 0.0661 ms 40.1% SingleProcess AUTOTUNE benchmarking takes 2.1015 seconds and 0.0000 seconds precompiling AUTOTUNE mm(1088x6840, 6840x64) mm 0.0294 ms 100.0% triton_mm_4874 0.0438 ms 67.2% triton_mm_4882 0.0478 ms 61.5% triton_mm_4878 0.0479 ms 61.5% triton_mm_4872 0.0591 ms 49.8% triton_mm_4873 0.0644 ms 45.7% triton_mm_4881 0.0676 ms 43.5% triton_mm_4871 0.0704 ms 41.8% triton_mm_4877 0.0723 ms 40.7% triton_mm_4880 0.0793 ms 37.1% SingleProcess AUTOTUNE benchmarking takes 7.8694 seconds and 0.0000 seconds precompiling /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): /home/user/VAR/tokenizer/tokenizer_image/dino_enc/dinov2.py:122: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead. with torch.cuda.amp.autocast(enabled=False): Validation Progress: 0%| | 0/44 [00:00