On Wed, Aug 15, 2012 at 05:22:03PM +0300, Jussi Kivilinna wrote: > Patch replaces 'movb' instructions with 'movzbl' to break false > register dependencies and interleaves instructions better for > out-of-order scheduling. > > Also move common round code to separate function to reduce object > size. Ok, redid the first test $ modprobe twofish-avx-x86_64 $ modprobe tcrypt mode=504 sec=1 and from quickly juxtaposing the two results, I'd say the patch makes things slightly worse but you'd need to run your scripts on it to get the accurate results: [ 98.206067] testing speed of async ecb(twofish) encryption [ 98.214796] test 0 (128 bit key, 16 byte blocks): 4549296 operations in 1 seconds (72788736 bytes) [ 99.221569] test 1 (128 bit key, 64 byte blocks): 1995934 operations in 1 seconds (127739776 bytes) [ 100.228250] test 2 (128 bit key, 256 byte blocks): 535040 operations in 1 seconds (136970240 bytes) [ 101.234751] test 3 (128 bit key, 1024 byte blocks): 148602 operations in 1 seconds (152168448 bytes) [ 102.241345] test 4 (128 bit key, 8192 byte blocks): 19148 operations in 1 seconds (156860416 bytes) [ 103.247880] test 5 (192 bit key, 16 byte blocks): 4558391 operations in 1 seconds (72934256 bytes) [ 104.254547] test 6 (192 bit key, 64 byte blocks): 1997838 operations in 1 seconds (127861632 bytes) [ 105.261202] test 7 (192 bit key, 256 byte blocks): 534396 operations in 1 seconds (136805376 bytes) [ 106.267694] test 8 (192 bit key, 1024 byte blocks): 148199 operations in 1 seconds (151755776 bytes) [ 107.274296] test 9 (192 bit key, 8192 byte blocks): 18913 operations in 1 seconds (154935296 bytes) [ 108.280824] test 10 (256 bit key, 16 byte blocks): 4595524 operations in 1 seconds (73528384 bytes) [ 109.287496] test 11 (256 bit key, 64 byte blocks): 1997893 operations in 1 seconds (127865152 bytes) [ 110.294168] test 12 (256 bit key, 256 byte blocks): 533790 operations in 1 seconds (136650240 bytes) [ 111.300679] test 13 (256 bit key, 1024 byte blocks): 148787 operations in 1 seconds (152357888 bytes) [ 112.303561] test 14 (256 bit key, 8192 byte blocks): 19146 operations in 1 seconds (156844032 bytes) [ 113.310104] [ 113.310104] testing speed of async ecb(twofish) decryption [ 113.319419] test 0 (128 bit key, 16 byte blocks): 4754043 operations in 1 seconds (76064688 bytes) [ 114.324768] test 1 (128 bit key, 64 byte blocks): 1831420 operations in 1 seconds (117210880 bytes) [ 115.331441] test 2 (128 bit key, 256 byte blocks): 541170 operations in 1 seconds (138539520 bytes) [ 116.337957] test 3 (128 bit key, 1024 byte blocks): 150538 operations in 1 seconds (154150912 bytes) [ 117.344571] test 4 (128 bit key, 8192 byte blocks): 19397 operations in 1 seconds (158900224 bytes) [ 118.351122] test 5 (192 bit key, 16 byte blocks): 4753957 operations in 1 seconds (76063312 bytes) [ 119.357778] test 6 (192 bit key, 64 byte blocks): 1828676 operations in 1 seconds (117035264 bytes) [ 120.364459] test 7 (192 bit key, 256 byte blocks): 540331 operations in 1 seconds (138324736 bytes) [ 121.370969] test 8 (192 bit key, 1024 byte blocks): 150348 operations in 1 seconds (153956352 bytes) [ 122.377573] test 9 (192 bit key, 8192 byte blocks): 19196 operations in 1 seconds (157253632 bytes) [ 123.384080] test 10 (256 bit key, 16 byte blocks): 4664399 operations in 1 seconds (74630384 bytes) [ 124.390782] test 11 (256 bit key, 64 byte blocks): 1839324 operations in 1 seconds (117716736 bytes) [ 125.397463] test 12 (256 bit key, 256 byte blocks): 538735 operations in 1 seconds (137916160 bytes) [ 126.403962] test 13 (256 bit key, 1024 byte blocks): 150489 operations in 1 seconds (154100736 bytes) [ 127.410567] test 14 (256 bit key, 8192 byte blocks): 19397 operations in 1 seconds (158900224 bytes) [ 128.417091] [ 128.417091] testing speed of async cbc(twofish) encryption [ 128.431227] test 0 (128 bit key, 16 byte blocks): 4681239 operations in 1 seconds (74899824 bytes) [ 129.439466] test 1 (128 bit key, 64 byte blocks): 1836636 operations in 1 seconds (117544704 bytes) [ 130.446131] test 2 (128 bit key, 256 byte blocks): 536055 operations in 1 seconds (137230080 bytes) [ 131.452631] test 3 (128 bit key, 1024 byte blocks): 140955 operations in 1 seconds (144337920 bytes) [ 132.459243] test 4 (128 bit key, 8192 byte blocks): 17821 operations in 1 seconds (145989632 bytes) [ 133.466124] test 5 (192 bit key, 16 byte blocks): 4674373 operations in 1 seconds (74789968 bytes) [ 134.472728] test 6 (192 bit key, 64 byte blocks): 1835821 operations in 1 seconds (117492544 bytes) [ 135.479374] test 7 (192 bit key, 256 byte blocks): 535882 operations in 1 seconds (137185792 bytes) [ 136.485876] test 8 (192 bit key, 1024 byte blocks): 140917 operations in 1 seconds (144299008 bytes) [ 137.492470] test 9 (192 bit key, 8192 byte blocks): 17707 operations in 1 seconds (145055744 bytes) [ 138.498979] test 10 (256 bit key, 16 byte blocks): 4674648 operations in 1 seconds (74794368 bytes) [ 139.505660] test 11 (256 bit key, 64 byte blocks): 1828219 operations in 1 seconds (117006016 bytes) [ 140.512343] test 12 (256 bit key, 256 byte blocks): 535835 operations in 1 seconds (137173760 bytes) [ 141.518842] test 13 (256 bit key, 1024 byte blocks): 140884 operations in 1 seconds (144265216 bytes) [ 142.525447] test 14 (256 bit key, 8192 byte blocks): 17815 operations in 1 seconds (145940480 bytes) [ 143.531972] [ 143.531972] testing speed of async cbc(twofish) decryption [ 143.541345] test 0 (128 bit key, 16 byte blocks): 4461471 operations in 1 seconds (71383536 bytes) [ 144.546671] test 1 (128 bit key, 64 byte blocks): 1726158 operations in 1 seconds (110474112 bytes) [ 145.553334] test 2 (128 bit key, 256 byte blocks): 524618 operations in 1 seconds (134302208 bytes) [ 146.559862] test 3 (128 bit key, 1024 byte blocks): 145305 operations in 1 seconds (148792320 bytes) [ 147.566457] test 4 (128 bit key, 8192 byte blocks): 18667 operations in 1 seconds (152920064 bytes) [ 148.572965] test 5 (192 bit key, 16 byte blocks): 4458941 operations in 1 seconds (71343056 bytes) [ 149.579638] test 6 (192 bit key, 64 byte blocks): 1734677 operations in 1 seconds (111019328 bytes) [ 150.586303] test 7 (192 bit key, 256 byte blocks): 521797 operations in 1 seconds (133580032 bytes) [ 151.592811] test 8 (192 bit key, 1024 byte blocks): 144554 operations in 1 seconds (148023296 bytes) [ 152.599423] test 9 (192 bit key, 8192 byte blocks): 18461 operations in 1 seconds (151232512 bytes) [ 153.605932] test 10 (256 bit key, 16 byte blocks): 4454216 operations in 1 seconds (71267456 bytes) [ 154.612614] test 11 (256 bit key, 64 byte blocks): 1749350 operations in 1 seconds (111958400 bytes) [ 155.619270] test 12 (256 bit key, 256 byte blocks): 525143 operations in 1 seconds (134436608 bytes) [ 156.625778] test 13 (256 bit key, 1024 byte blocks): 145597 operations in 1 seconds (149091328 bytes) [ 157.632367] test 14 (256 bit key, 8192 byte blocks): 18667 operations in 1 seconds (152920064 bytes) [ 158.638911] [ 158.638911] testing speed of async ctr(twofish) encryption [ 158.652915] test 0 (128 bit key, 16 byte blocks): 4582013 operations in 1 seconds (73312208 bytes) [ 159.661274] test 1 (128 bit key, 64 byte blocks): 1949294 operations in 1 seconds (124754816 bytes) [ 160.667949] test 2 (128 bit key, 256 byte blocks): 519205 operations in 1 seconds (132916480 bytes) [ 161.674749] test 3 (128 bit key, 1024 byte blocks): 142060 operations in 1 seconds (145469440 bytes) [ 162.681372] test 4 (128 bit key, 8192 byte blocks): 18272 operations in 1 seconds (149684224 bytes) [ 163.687577] test 5 (192 bit key, 16 byte blocks): 4539161 operations in 1 seconds (72626576 bytes) [ 164.694561] test 6 (192 bit key, 64 byte blocks): 1935006 operations in 1 seconds (123840384 bytes) [ 165.701209] test 7 (192 bit key, 256 byte blocks): 517208 operations in 1 seconds (132405248 bytes) [ 166.707725] test 8 (192 bit key, 1024 byte blocks): 141790 operations in 1 seconds (145192960 bytes) [ 167.714338] test 9 (192 bit key, 8192 byte blocks): 18120 operations in 1 seconds (148439040 bytes) [ 168.720856] test 10 (256 bit key, 16 byte blocks): 4379275 operations in 1 seconds (70068400 bytes) [ 169.727530] test 11 (256 bit key, 64 byte blocks): 1957465 operations in 1 seconds (125277760 bytes) [ 170.734185] test 12 (256 bit key, 256 byte blocks): 519760 operations in 1 seconds (133058560 bytes) [ 171.740392] test 13 (256 bit key, 1024 byte blocks): 142374 operations in 1 seconds (145790976 bytes) [ 172.746986] test 14 (256 bit key, 8192 byte blocks): 18292 operations in 1 seconds (149848064 bytes) [ 173.753539] [ 173.753539] testing speed of async ctr(twofish) decryption [ 173.762929] test 0 (128 bit key, 16 byte blocks): 4465609 operations in 1 seconds (71449744 bytes) [ 174.768467] test 1 (128 bit key, 64 byte blocks): 1947565 operations in 1 seconds (124644160 bytes) [ 175.775139] test 2 (128 bit key, 256 byte blocks): 523259 operations in 1 seconds (133954304 bytes) [ 176.781352] test 3 (128 bit key, 1024 byte blocks): 141135 operations in 1 seconds (144522240 bytes) [ 177.787959] test 4 (128 bit key, 8192 byte blocks): 17984 operations in 1 seconds (147324928 bytes) [ 178.794512] test 5 (192 bit key, 16 byte blocks): 4541736 operations in 1 seconds (72667776 bytes) [ 179.801141] test 6 (192 bit key, 64 byte blocks): 1937279 operations in 1 seconds (123985856 bytes) [ 180.807805] test 7 (192 bit key, 256 byte blocks): 513856 operations in 1 seconds (131547136 bytes) [ 181.814331] test 8 (192 bit key, 1024 byte blocks): 141039 operations in 1 seconds (144423936 bytes) [ 182.820918] test 9 (192 bit key, 8192 byte blocks): 17825 operations in 1 seconds (146022400 bytes) [ 183.827461] test 10 (256 bit key, 16 byte blocks): 4380875 operations in 1 seconds (70094000 bytes) [ 184.834419] test 11 (256 bit key, 64 byte blocks): 1959937 operations in 1 seconds (125435968 bytes) [ 185.841075] test 12 (256 bit key, 256 byte blocks): 515782 operations in 1 seconds (132040192 bytes) [ 186.847585] test 13 (256 bit key, 1024 byte blocks): 142571 operations in 1 seconds (145992704 bytes) [ 187.854181] test 14 (256 bit key, 8192 byte blocks): 18105 operations in 1 seconds (148316160 bytes) [ 188.860717] [ 188.860717] testing speed of async lrw(twofish) encryption [ 188.875294] test 0 (256 bit key, 16 byte blocks): 3445285 operations in 1 seconds (55124560 bytes) [ 189.883381] test 1 (256 bit key, 64 byte blocks): 1585896 operations in 1 seconds (101497344 bytes) [ 190.890072] test 2 (256 bit key, 256 byte blocks): 449477 operations in 1 seconds (115066112 bytes) [ 191.896590] test 3 (256 bit key, 1024 byte blocks): 123541 operations in 1 seconds (126505984 bytes) [ 192.903174] test 4 (256 bit key, 8192 byte blocks): 15868 operations in 1 seconds (129990656 bytes) [ 193.909694] test 5 (320 bit key, 16 byte blocks): 3590396 operations in 1 seconds (57446336 bytes) [ 194.916355] test 6 (320 bit key, 64 byte blocks): 1579004 operations in 1 seconds (101056256 bytes) [ 195.923041] test 7 (320 bit key, 256 byte blocks): 449033 operations in 1 seconds (114952448 bytes) [ 196.929529] test 8 (320 bit key, 1024 byte blocks): 123347 operations in 1 seconds (126307328 bytes) [ 197.936142] test 9 (320 bit key, 8192 byte blocks): 15762 operations in 1 seconds (129122304 bytes) [ 198.942702] test 10 (384 bit key, 16 byte blocks): 3496049 operations in 1 seconds (55936784 bytes) [ 199.949333] test 11 (384 bit key, 64 byte blocks): 1589166 operations in 1 seconds (101706624 bytes) [ 200.955996] test 12 (384 bit key, 256 byte blocks): 449480 operations in 1 seconds (115066880 bytes) [ 201.962497] test 13 (384 bit key, 1024 byte blocks): 123767 operations in 1 seconds (126737408 bytes) [ 202.969101] test 14 (384 bit key, 8192 byte blocks): 15921 operations in 1 seconds (130424832 bytes) [ 203.971665] [ 203.971665] testing speed of async lrw(twofish) decryption [ 203.971755] test 0 (256 bit key, 16 byte blocks): 3558879 operations in 1 seconds (56942064 bytes) [ 204.974331] test 1 (256 bit key, 64 byte blocks): 1588116 operations in 1 seconds (101639424 bytes) [ 205.981001] test 2 (256 bit key, 256 byte blocks): 451198 operations in 1 seconds (115506688 bytes) [ 206.987510] test 3 (256 bit key, 1024 byte blocks): 124791 operations in 1 seconds (127785984 bytes) [ 207.994115] test 4 (256 bit key, 8192 byte blocks): 16087 operations in 1 seconds (131784704 bytes) [ 209.000650] test 5 (320 bit key, 16 byte blocks): 3559066 operations in 1 seconds (56945056 bytes) [ 210.007298] test 6 (320 bit key, 64 byte blocks): 1579234 operations in 1 seconds (101070976 bytes) [ 211.013960] test 7 (320 bit key, 256 byte blocks): 454953 operations in 1 seconds (116467968 bytes) [ 212.020469] test 8 (320 bit key, 1024 byte blocks): 124810 operations in 1 seconds (127805440 bytes) [ 213.027082] test 9 (320 bit key, 8192 byte blocks): 15887 operations in 1 seconds (130146304 bytes) [ 214.033610] test 10 (384 bit key, 16 byte blocks): 3554484 operations in 1 seconds (56871744 bytes) [ 215.040272] test 11 (384 bit key, 64 byte blocks): 1583334 operations in 1 seconds (101333376 bytes) [ 216.046937] test 12 (384 bit key, 256 byte blocks): 453554 operations in 1 seconds (116109824 bytes) [ 217.053436] test 13 (384 bit key, 1024 byte blocks): 124894 operations in 1 seconds (127891456 bytes) [ 218.060032] test 14 (384 bit key, 8192 byte blocks): 16080 operations in 1 seconds (131727360 bytes) [ 219.066597] [ 219.066597] testing speed of async xts(twofish) encryption [ 219.080737] test 0 (256 bit key, 16 byte blocks): 3105784 operations in 1 seconds (49692544 bytes) [ 220.089254] test 1 (256 bit key, 64 byte blocks): 1586587 operations in 1 seconds (101541568 bytes) [ 221.095918] test 2 (256 bit key, 256 byte blocks): 475166 operations in 1 seconds (121642496 bytes) [ 222.102427] test 3 (256 bit key, 1024 byte blocks): 133144 operations in 1 seconds (136339456 bytes) [ 223.109038] test 4 (256 bit key, 8192 byte blocks): 17219 operations in 1 seconds (141058048 bytes) [ 224.115549] test 5 (384 bit key, 16 byte blocks): 3097574 operations in 1 seconds (49561184 bytes) [ 225.122213] test 6 (384 bit key, 64 byte blocks): 1585836 operations in 1 seconds (101493504 bytes) [ 226.128885] test 7 (384 bit key, 256 byte blocks): 475173 operations in 1 seconds (121644288 bytes) [ 227.135398] test 8 (384 bit key, 1024 byte blocks): 133173 operations in 1 seconds (136369152 bytes) [ 228.138011] test 9 (384 bit key, 8192 byte blocks): 17254 operations in 1 seconds (141344768 bytes) [ 229.140563] test 10 (512 bit key, 16 byte blocks): 3171090 operations in 1 seconds (50737440 bytes) [ 230.147211] test 11 (512 bit key, 64 byte blocks): 1595445 operations in 1 seconds (102108480 bytes) [ 231.153866] test 12 (512 bit key, 256 byte blocks): 475161 operations in 1 seconds (121641216 bytes) [ 232.160384] test 13 (512 bit key, 1024 byte blocks): 133269 operations in 1 seconds (136467456 bytes) [ 233.166970] test 14 (512 bit key, 8192 byte blocks): 17225 operations in 1 seconds (141107200 bytes) [ 234.173501] [ 234.173501] testing speed of async xts(twofish) decryption [ 234.182898] test 0 (256 bit key, 16 byte blocks): 3095689 operations in 1 seconds (49531024 bytes) [ 235.188173] test 1 (256 bit key, 64 byte blocks): 1433025 operations in 1 seconds (91713600 bytes) [ 236.194753] test 2 (256 bit key, 256 byte blocks): 472038 operations in 1 seconds (120841728 bytes) [ 237.201347] test 3 (256 bit key, 1024 byte blocks): 134015 operations in 1 seconds (137231360 bytes) [ 238.207969] test 4 (256 bit key, 8192 byte blocks): 17446 operations in 1 seconds (142917632 bytes) [ 239.214478] test 5 (384 bit key, 16 byte blocks): 3099755 operations in 1 seconds (49596080 bytes) [ 240.221142] test 6 (384 bit key, 64 byte blocks): 1432335 operations in 1 seconds (91669440 bytes) [ 241.227711] test 7 (384 bit key, 256 byte blocks): 470340 operations in 1 seconds (120407040 bytes) [ 242.234314] test 8 (384 bit key, 1024 byte blocks): 133929 operations in 1 seconds (137143296 bytes) [ 243.240926] test 9 (384 bit key, 8192 byte blocks): 17442 operations in 1 seconds (142884864 bytes) [ 244.247453] test 10 (512 bit key, 16 byte blocks): 3193773 operations in 1 seconds (51100368 bytes) [ 245.254119] test 11 (512 bit key, 64 byte blocks): 1440631 operations in 1 seconds (92200384 bytes) [ 246.260689] test 12 (512 bit key, 256 byte blocks): 475293 operations in 1 seconds (121675008 bytes) [ 247.267283] test 13 (512 bit key, 1024 byte blocks): 134350 operations in 1 seconds (137574400 bytes) [ 248.273879] test 14 (512 bit key, 8192 byte blocks): 17441 operations in 1 seconds (142876672 bytes) -- Regards/Gruss, Boris. -- To unsubscribe from this list: send the line "unsubscribe linux-crypto" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
