================================================================================ ANALYSIS SUMMARY ================================================================================ Overall: 693 cases, 90 faster >1%, 276 slower >1%, 327 wash (±1%), avg +0.98% Per-benchmark averages: bench_gemm -2.48% (23/25 faster) ← clear winner bench_gemm_double +0.20% (wash) bench_gemv +1.01% (mostly regression) bench_vecadd +0.32% (wash) bench_reductions +0.62% (mixed) bench_fixed_size +0.98% (slight regression) bench_fixed_size_double +1.18% (slight regression) bench_trsm +2.33% (worst avg — TRSV especially bad) bench_trmv +1.75% (float Upper consistently bad) bench_triangular_product +1.03% (regression) bench_selfadjoint_product +0.76% (wash) bench_dot +1.27% (noisy, outliers both ways) bench_block_ops +1.74% (mostly regression) Notable wins: BM_EigenGemm/192/192/192 -8.53% BM_EigenGemm/160/160/160 -8.45% TRMV_float_Lower/512 -8.46% VectorNorm_double/65536 -14.01% VectorMinCoeff_float/4096 -7.59% VectorMinCoeff_float/262144 -7.43% VectorProd_float/262144 -7.48% VectorSum_float/262144 -6.57% Notable regressions: BlockRead_float/512/64 +78.29% (likely noise/cache artifact) VectorLpNormInf_double/1048576 +20.62% TRMV_float_UnitUpper/1024 +18.89% TRSV_float_Lower/512 +18.78% TRMV_float_Upper/1024 +18.51% TRSV_float_Lower/128 +17.61% Dot_double/65536 +15.78% Dot_cfloat/1048576 +15.76% TRMV_double_UnitUpper/2048 +15.63% VectorLpNormInf_double/65536 +11.47% TRSV_float_Lower/32 +11.15% TRMV_float_UnitUpper/64 +11.20% Pattern: __builtin_assume helps float GEMM consistently (compiler better exploits alignment/size hints for the GEBP kernel) but hurts triangular solvers (TRSV/TRMV Upper) and large-vector reductions significantly. The hints appear to misguide vectorization decisions for upper-triangular traversal patterns and large strided reductions where the compiler's default heuristics were better. benchmark case baseline assume delta ------------------------------------------------------------------------------------------------------------------------ bench_gemm BM_EigenGemm/1024/1024/1024 15370160.7 14966207.5 -2.63% ▼ bench_gemm BM_EigenGemm/1024/256/256 943602.7 933646.5 -1.06% ▼ bench_gemm BM_EigenGemm/1024/64/64 62870.2 60982.8 -3.00% ▼ bench_gemm BM_EigenGemm/128/128/128 30253.1 28302.0 -6.45% ▼ bench_gemm BM_EigenGemm/1536/1536/1536 51310418.0 50463862.3 -1.65% ▼ bench_gemm BM_EigenGemm/16/16/16 98.3 96.0 -2.31% ▼ bench_gemm BM_EigenGemm/160/160/160 58022.4 53118.2 -8.45% ▼ bench_gemm BM_EigenGemm/192/192/192 99923.4 91399.8 -8.53% ▼ bench_gemm BM_EigenGemm/2048/2048/2048 121240155.3 119787774.3 -1.20% ▼ bench_gemm BM_EigenGemm/224/224/224 159143.2 150377.5 -5.51% ▼ bench_gemm BM_EigenGemm/256/256/1024 944048.7 935531.9 -0.90% bench_gemm BM_EigenGemm/256/256/256 238890.5 233717.3 -2.17% ▼ bench_gemm BM_EigenGemm/288/288/288 338872.8 332307.9 -1.94% ▼ bench_gemm BM_EigenGemm/32/32/32 531.5 524.1 -1.38% ▼ bench_gemm BM_EigenGemm/320/320/320 462592.8 457127.2 -1.18% ▼ bench_gemm BM_EigenGemm/384/384/384 802774.4 791469.7 -1.41% ▼ bench_gemm BM_EigenGemm/4096/4096/4096 964079853.0 954828771.0 -0.96% bench_gemm BM_EigenGemm/448/448/448 1271951.9 1257827.3 -1.11% ▼ bench_gemm BM_EigenGemm/512/512/512 1887325.0 1861254.6 -1.38% ▼ bench_gemm BM_EigenGemm/64/1024/64 60957.3 60337.3 -1.02% ▼ bench_gemm BM_EigenGemm/64/64/1024 63606.6 62078.3 -2.40% ▼ bench_gemm BM_EigenGemm/64/64/64 3974.7 3926.0 -1.23% ▼ bench_gemm BM_EigenGemm/768/768/768 6377113.5 6300491.5 -1.20% ▼ bench_gemm BM_EigenGemm/8/8/8 48.8 47.9 -1.83% ▼ bench_gemm BM_EigenGemm/96/96/96 13027.9 12897.1 -1.00% ▼ bench_gemm_double BM_EigenGemm/1024/1024/1024 31160952.5 31594124.2 +1.39% ▲ bench_gemm_double BM_EigenGemm/1024/256/256 1917285.4 1931923.3 +0.76% bench_gemm_double BM_EigenGemm/1024/64/64 130185.2 133689.8 +2.69% ▲ bench_gemm_double BM_EigenGemm/128/128/128 60833.6 60247.6 -0.96% bench_gemm_double BM_EigenGemm/1536/1536/1536 104175692.4 104992046.3 +0.78% bench_gemm_double BM_EigenGemm/16/16/16 168.4 165.7 -1.61% ▼ bench_gemm_double BM_EigenGemm/160/160/160 117817.1 117629.2 -0.16% bench_gemm_double BM_EigenGemm/192/192/192 202454.0 202528.1 +0.04% bench_gemm_double BM_EigenGemm/2048/2048/2048 246366396.7 246627620.0 +0.11% bench_gemm_double BM_EigenGemm/224/224/224 322340.4 324635.7 +0.71% bench_gemm_double BM_EigenGemm/256/256/1024 1939623.9 1967729.9 +1.45% ▲ bench_gemm_double BM_EigenGemm/256/256/256 482907.5 483541.8 +0.13% bench_gemm_double BM_EigenGemm/288/288/288 680497.1 689752.6 +1.36% ▲ bench_gemm_double BM_EigenGemm/32/32/32 1054.8 1041.7 -1.24% ▼ bench_gemm_double BM_EigenGemm/320/320/320 936305.6 947626.1 +1.21% ▲ bench_gemm_double BM_EigenGemm/384/384/384 1602909.7 1663692.2 +3.79% ▲ bench_gemm_double BM_EigenGemm/4096/4096/4096 1955234582.0 1963792216.0 +0.44% bench_gemm_double BM_EigenGemm/448/448/448 2543251.8 2594724.6 +2.02% ▲ bench_gemm_double BM_EigenGemm/512/512/512 3835843.0 3955670.0 +3.12% ▲ bench_gemm_double BM_EigenGemm/64/1024/64 125294.6 127319.6 +1.62% ▲ bench_gemm_double BM_EigenGemm/64/64/1024 129812.4 131591.2 +1.37% ▲ bench_gemm_double BM_EigenGemm/64/64/64 8163.1 8128.1 -0.43% bench_gemm_double BM_EigenGemm/768/768/768 12979498.7 13138476.8 +1.22% ▲ bench_gemm_double BM_EigenGemm/8/8/8 53.8 46.1 -14.40% ▼ bench_gemm_double BM_EigenGemm/96/96/96 26514.5 26425.6 -0.34% bench_gemv GemvAdj_cfloat/1/1024 1585.3 1593.1 +0.49% bench_gemv GemvAdj_cfloat/1/256 394.8 402.1 +1.85% ▲ bench_gemv GemvAdj_cfloat/1024/1 92.5 93.1 +0.65% bench_gemv GemvAdj_cfloat/1024/1024 85317.9 85671.6 +0.41% bench_gemv GemvAdj_cfloat/1024/16 1190.1 1202.9 +1.08% ▲ bench_gemv GemvAdj_cfloat/128/128 1338.8 1372.6 +2.53% ▲ bench_gemv GemvAdj_cfloat/16/1024 2585.3 2612.9 +1.07% ▲ bench_gemv GemvAdj_cfloat/16/256 640.3 641.9 +0.25% bench_gemv GemvAdj_cfloat/256/1 29.3 29.5 +0.76% bench_gemv GemvAdj_cfloat/256/16 311.5 317.8 +2.04% ▲ bench_gemv GemvAdj_cfloat/32/32 116.0 118.5 +2.16% ▲ bench_gemv GemvAdj_cfloat/512/512 22727.0 23489.0 +3.35% ▲ bench_gemv GemvAdj_cfloat/8/8 16.6 17.1 +2.80% ▲ bench_gemv GemvConj_cfloat/1/1024 886.0 891.9 +0.67% bench_gemv GemvConj_cfloat/1/256 226.8 228.2 +0.61% bench_gemv GemvConj_cfloat/1024/1 151.1 150.7 -0.28% bench_gemv GemvConj_cfloat/1024/1024 88128.7 92006.2 +4.40% ▲ bench_gemv GemvConj_cfloat/1024/16 1238.1 1245.5 +0.59% bench_gemv GemvConj_cfloat/128/128 1242.8 1250.4 +0.61% bench_gemv GemvConj_cfloat/16/1024 1350.1 1381.2 +2.31% ▲ bench_gemv GemvConj_cfloat/16/256 338.3 340.0 +0.49% bench_gemv GemvConj_cfloat/256/1 38.5 38.7 +0.66% bench_gemv GemvConj_cfloat/256/16 311.8 310.1 -0.53% bench_gemv GemvConj_cfloat/32/32 78.2 78.8 +0.71% bench_gemv GemvConj_cfloat/512/512 24048.9 25293.8 +5.18% ▲ bench_gemv GemvConj_cfloat/8/8 9.7 9.8 +0.64% bench_gemv GemvTrans_cfloat/1/1024 1639.5 1650.4 +0.67% bench_gemv GemvTrans_cfloat/1/256 410.3 412.0 +0.44% bench_gemv GemvTrans_cfloat/1024/1 92.3 92.8 +0.51% bench_gemv GemvTrans_cfloat/1024/1024 85387.3 87333.1 +2.28% ▲ bench_gemv GemvTrans_cfloat/1024/16 1186.1 1212.8 +2.25% ▲ bench_gemv GemvTrans_cfloat/128/128 1337.2 1343.5 +0.47% bench_gemv GemvTrans_cfloat/16/1024 2557.9 2574.9 +0.66% bench_gemv GemvTrans_cfloat/16/256 637.0 640.1 +0.48% bench_gemv GemvTrans_cfloat/256/1 29.1 29.3 +0.61% bench_gemv GemvTrans_cfloat/256/16 310.8 317.3 +2.12% ▲ bench_gemv GemvTrans_cfloat/32/32 116.0 116.7 +0.62% bench_gemv GemvTrans_cfloat/512/512 22764.8 22965.7 +0.88% bench_gemv GemvTrans_cfloat/8/8 16.6 16.7 +0.68% bench_gemv GemvTrans_double/1/1024 754.3 759.2 +0.65% bench_gemv GemvTrans_double/1/256 188.1 191.7 +1.96% ▲ bench_gemv GemvTrans_double/1024/1 55.4 55.7 +0.53% bench_gemv GemvTrans_double/1024/1024 67695.4 68496.1 +1.18% ▲ bench_gemv GemvTrans_double/1024/16 741.4 745.3 +0.52% bench_gemv GemvTrans_double/128/128 889.0 900.8 +1.33% ▲ bench_gemv GemvTrans_double/16/1024 976.8 983.6 +0.69% bench_gemv GemvTrans_double/16/256 238.3 240.3 +0.85% bench_gemv GemvTrans_double/256/1 14.2 14.3 +0.40% bench_gemv GemvTrans_double/256/16 132.7 134.8 +1.59% ▲ bench_gemv GemvTrans_double/32/32 42.8 42.1 -1.66% ▼ bench_gemv GemvTrans_double/512/512 18210.9 18182.0 -0.16% bench_gemv GemvTrans_double/8/8 9.3 9.2 -1.16% ▼ bench_gemv GemvTrans_float/1/1024 899.6 898.1 -0.17% bench_gemv GemvTrans_float/1/256 224.2 224.5 +0.12% bench_gemv GemvTrans_float/1024/1 27.9 27.9 -0.00% bench_gemv GemvTrans_float/1024/1024 32686.0 33144.8 +1.40% ▲ bench_gemv GemvTrans_float/1024/16 319.2 319.4 +0.08% bench_gemv GemvTrans_float/128/128 448.6 452.1 +0.78% bench_gemv GemvTrans_float/16/1024 950.2 949.6 -0.07% bench_gemv GemvTrans_float/16/256 235.3 235.0 -0.11% bench_gemv GemvTrans_float/256/1 7.4 7.4 -0.02% bench_gemv GemvTrans_float/256/16 70.5 71.1 +0.91% bench_gemv GemvTrans_float/32/32 38.2 37.6 -1.53% ▼ bench_gemv GemvTrans_float/512/512 7348.0 7566.8 +2.98% ▲ bench_gemv GemvTrans_float/8/8 9.6 9.9 +2.57% ▲ bench_gemv Gemv_cfloat/1/1024 695.1 706.1 +1.59% ▲ bench_gemv Gemv_cfloat/1/256 170.3 173.3 +1.78% ▲ bench_gemv Gemv_cfloat/1024/1 148.8 149.8 +0.66% bench_gemv Gemv_cfloat/1024/1024 90079.4 90206.9 +0.14% bench_gemv Gemv_cfloat/1024/16 1234.6 1258.9 +1.97% ▲ bench_gemv Gemv_cfloat/128/128 1238.0 1266.1 +2.27% ▲ bench_gemv Gemv_cfloat/16/1024 1350.9 1372.6 +1.60% ▲ bench_gemv Gemv_cfloat/16/256 338.2 345.3 +2.09% ▲ bench_gemv Gemv_cfloat/256/1 38.4 38.5 +0.25% bench_gemv Gemv_cfloat/256/16 307.5 309.3 +0.59% bench_gemv Gemv_cfloat/32/32 77.9 79.7 +2.29% ▲ bench_gemv Gemv_cfloat/512/512 24002.7 24899.1 +3.73% ▲ bench_gemv Gemv_cfloat/8/8 9.6 9.7 +1.14% ▲ bench_gemv Gemv_double/1/1024 815.5 816.8 +0.15% bench_gemv Gemv_double/1/256 162.0 162.3 +0.17% bench_gemv Gemv_double/1024/1 62.9 63.5 +0.96% bench_gemv Gemv_double/1024/1024 92398.6 95268.6 +3.11% ▲ bench_gemv Gemv_double/1024/16 1005.7 1008.7 +0.30% bench_gemv Gemv_double/128/128 888.4 899.8 +1.28% ▲ bench_gemv Gemv_double/16/1024 704.6 706.4 +0.26% bench_gemv Gemv_double/16/256 172.5 171.4 -0.65% bench_gemv Gemv_double/256/1 18.5 18.8 +1.89% ▲ bench_gemv Gemv_double/256/16 127.6 129.2 +1.24% ▲ bench_gemv Gemv_double/32/32 33.9 34.0 +0.47% bench_gemv Gemv_double/512/512 20986.6 21975.7 +4.71% ▲ bench_gemv Gemv_double/8/8 6.7 6.6 -1.64% ▼ bench_gemv Gemv_float/1/1024 817.7 816.4 -0.16% bench_gemv Gemv_float/1/256 163.2 163.3 +0.03% bench_gemv Gemv_float/1024/1 32.9 33.2 +0.95% bench_gemv Gemv_float/1024/1024 39439.3 40003.8 +1.43% ▲ bench_gemv Gemv_float/1024/16 404.5 407.6 +0.77% bench_gemv Gemv_float/128/128 327.6 334.4 +2.06% ▲ bench_gemv Gemv_float/16/1024 488.2 487.2 -0.22% bench_gemv Gemv_float/16/256 118.5 118.5 -0.04% bench_gemv Gemv_float/256/1 11.0 11.0 +0.25% bench_gemv Gemv_float/256/16 65.3 65.0 -0.52% bench_gemv Gemv_float/32/32 21.8 22.5 +3.32% ▲ bench_gemv Gemv_float/512/512 7762.7 7992.6 +2.96% ▲ bench_gemv Gemv_float/8/8 6.2 6.2 +0.19% bench_vecadd BM_MatAdd/128 627.3 628.9 +0.25% bench_vecadd BM_MatAdd/16 6.9 7.0 +0.56% bench_vecadd BM_MatAdd/256 2138.8 2119.6 -0.89% bench_vecadd BM_MatAdd/32 27.7 28.2 +1.60% ▲ bench_vecadd BM_MatAdd/512 16517.0 16543.6 +0.16% bench_vecadd BM_MatAdd/64 111.2 111.3 +0.08% bench_vecadd BM_MatAdd/8 3.4 3.5 +1.47% ▲ bench_vecadd BM_VecAdd/1024 28.5 28.8 +0.80% bench_vecadd BM_VecAdd/1048576 63872.9 64543.7 +1.05% ▲ bench_vecadd BM_VecAdd/16384 670.0 671.7 +0.26% bench_vecadd BM_VecAdd/256 8.5 8.4 -0.57% bench_vecadd BM_VecAdd/262144 16722.1 16601.3 -0.72% bench_vecadd BM_VecAdd/4096 112.2 113.7 +1.27% ▲ bench_vecadd BM_VecAdd/64 3.3 3.3 -0.92% bench_vecadd BM_VecAdd/65536 2280.9 2291.4 +0.46% bench_reductions MatrixNorm_double/1024 67735.2 70479.9 +4.05% ▲ bench_reductions MatrixNorm_double/128 725.1 727.9 +0.38% bench_reductions MatrixNorm_double/256 2783.3 2792.4 +0.33% bench_reductions MatrixNorm_double/32 42.8 43.2 +0.97% bench_reductions MatrixNorm_double/512 16645.8 17006.1 +2.16% ▲ bench_reductions MatrixNorm_double/64 171.1 172.3 +0.68% bench_reductions MatrixNorm_double/8 3.1 3.1 +0.27% bench_reductions MatrixNorm_float/1024 36758.9 37450.7 +1.88% ▲ bench_reductions MatrixNorm_float/128 364.4 367.4 +0.83% bench_reductions MatrixNorm_float/256 1455.9 1475.0 +1.32% ▲ bench_reductions MatrixNorm_float/32 22.0 21.0 -4.59% ▼ bench_reductions MatrixNorm_float/512 6893.7 7424.3 +7.70% ▲ bench_reductions MatrixNorm_float/64 86.4 88.2 +2.04% ▲ bench_reductions MatrixNorm_float/8 2.0 2.0 -2.65% ▼ bench_reductions MatrixSum_double/1024 64291.9 69068.9 +7.43% ▲ bench_reductions MatrixSum_double/128 651.6 685.3 +5.18% ▲ bench_reductions MatrixSum_double/256 2623.1 2714.8 +3.50% ▲ bench_reductions MatrixSum_double/32 31.8 32.3 +1.61% ▲ bench_reductions MatrixSum_double/512 16318.9 16718.2 +2.45% ▲ bench_reductions MatrixSum_double/64 151.6 152.7 +0.69% bench_reductions MatrixSum_double/8 1.9 2.0 +1.45% ▲ bench_reductions MatrixSum_float/1024 37214.6 37025.9 -0.51% bench_reductions MatrixSum_float/128 337.8 341.1 +0.99% bench_reductions MatrixSum_float/256 1381.5 1386.4 +0.35% bench_reductions MatrixSum_float/32 14.9 15.1 +0.95% bench_reductions MatrixSum_float/512 6906.3 6701.1 -2.97% ▼ bench_reductions MatrixSum_float/64 70.5 71.6 +1.47% ▲ bench_reductions MatrixSum_float/8 1.2 1.3 +1.44% ▲ bench_reductions VectorLpNorm1_double/1024 37.1 36.7 -1.04% ▼ bench_reductions VectorLpNorm1_double/1048576 63875.1 67973.2 +6.42% ▲ bench_reductions VectorLpNorm1_double/16384 723.0 723.4 +0.05% bench_reductions VectorLpNorm1_double/256 8.2 8.3 +0.50% bench_reductions VectorLpNorm1_double/262144 16405.3 16641.1 +1.44% ▲ bench_reductions VectorLpNorm1_double/4096 163.7 161.9 -1.07% ▼ bench_reductions VectorLpNorm1_double/64 2.2 2.2 +0.44% bench_reductions VectorLpNorm1_double/65536 2844.0 2824.5 -0.68% bench_reductions VectorLpNorm1_float/1024 18.0 18.1 +0.67% bench_reductions VectorLpNorm1_float/1048576 37134.4 37573.2 +1.18% ▲ bench_reductions VectorLpNorm1_float/16384 374.5 376.3 +0.47% bench_reductions VectorLpNorm1_float/256 4.4 4.4 +0.70% bench_reductions VectorLpNorm1_float/262144 6997.6 6929.1 -0.98% bench_reductions VectorLpNorm1_float/4096 78.1 78.7 +0.75% bench_reductions VectorLpNorm1_float/64 1.5 1.5 +1.69% ▲ bench_reductions VectorLpNorm1_float/65536 1451.3 1481.2 +2.06% ▲ bench_reductions VectorLpNormInf_double/1024 30.9 31.7 +2.41% ▲ bench_reductions VectorLpNormInf_double/1048576 63801.8 76956.0 +20.62% ▲ bench_reductions VectorLpNormInf_double/16384 645.0 667.2 +3.44% ▲ bench_reductions VectorLpNormInf_double/256 7.8 7.9 +0.91% bench_reductions VectorLpNormInf_double/262144 16253.4 17368.8 +6.86% ▲ bench_reductions VectorLpNormInf_double/4096 143.6 125.1 -12.90% ▼ bench_reductions VectorLpNormInf_double/64 2.2 2.2 +0.66% bench_reductions VectorLpNormInf_double/65536 2343.7 2612.6 +11.47% ▲ bench_reductions VectorLpNormInf_float/1024 16.1 16.0 -0.82% bench_reductions VectorLpNormInf_float/1048576 36338.5 37060.4 +1.99% ▲ bench_reductions VectorLpNormInf_float/16384 321.9 322.3 +0.12% bench_reductions VectorLpNormInf_float/256 4.3 4.2 -0.88% bench_reductions VectorLpNormInf_float/262144 6860.2 6380.0 -7.00% ▼ bench_reductions VectorLpNormInf_float/4096 62.6 67.9 +8.41% ▲ bench_reductions VectorLpNormInf_float/64 1.5 1.5 +0.77% bench_reductions VectorLpNormInf_float/65536 1297.3 1304.9 +0.58% bench_reductions VectorMaxCoeff_double/1024 32.1 32.1 -0.25% bench_reductions VectorMaxCoeff_double/1048576 66729.5 68938.2 +3.31% ▲ bench_reductions VectorMaxCoeff_double/16384 635.6 634.3 -0.20% bench_reductions VectorMaxCoeff_double/256 7.6 7.6 +0.59% bench_reductions VectorMaxCoeff_double/262144 16539.0 16565.6 +0.16% bench_reductions VectorMaxCoeff_double/4096 141.8 142.5 +0.48% bench_reductions VectorMaxCoeff_double/64 2.1 2.1 +0.57% bench_reductions VectorMaxCoeff_double/65536 2214.0 2328.5 +5.17% ▲ bench_reductions VectorMaxCoeff_float/1024 16.0 16.2 +0.83% bench_reductions VectorMaxCoeff_float/1048576 36554.6 37328.2 +2.12% ▲ bench_reductions VectorMaxCoeff_float/16384 316.0 324.2 +2.60% ▲ bench_reductions VectorMaxCoeff_float/256 4.1 4.1 +0.21% bench_reductions VectorMaxCoeff_float/262144 6835.6 6504.2 -4.85% ▼ bench_reductions VectorMaxCoeff_float/4096 69.0 71.4 +3.46% ▲ bench_reductions VectorMaxCoeff_float/64 1.3 1.3 -0.11% bench_reductions VectorMaxCoeff_float/65536 1251.9 1291.2 +3.14% ▲ bench_reductions VectorMean_double/1024 32.5 33.6 +3.32% ▲ bench_reductions VectorMean_double/1048576 64198.6 69515.6 +8.28% ▲ bench_reductions VectorMean_double/16384 822.6 721.9 -12.25% ▼ bench_reductions VectorMean_double/256 7.7 7.6 -0.08% bench_reductions VectorMean_double/262144 16330.8 16702.5 +2.28% ▲ bench_reductions VectorMean_double/4096 153.4 152.7 -0.51% bench_reductions VectorMean_double/64 2.4 2.3 -0.59% bench_reductions VectorMean_double/65536 2884.4 2848.8 -1.23% ▼ bench_reductions VectorMean_float/1024 15.3 15.7 +2.11% ▲ bench_reductions VectorMean_float/1048576 37580.7 37464.7 -0.31% bench_reductions VectorMean_float/16384 377.4 383.0 +1.49% ▲ bench_reductions VectorMean_float/256 4.2 4.2 +1.81% ▲ bench_reductions VectorMean_float/262144 6896.7 6784.1 -1.63% ▼ bench_reductions VectorMean_float/4096 72.2 72.5 +0.45% bench_reductions VectorMean_float/64 1.6 1.6 +1.93% ▲ bench_reductions VectorMean_float/65536 1524.5 1549.3 +1.62% ▲ bench_reductions VectorMinCoeff_double/1024 32.1 32.2 +0.13% bench_reductions VectorMinCoeff_double/1048576 64193.1 67803.0 +5.62% ▲ bench_reductions VectorMinCoeff_double/16384 636.1 627.6 -1.35% ▼ bench_reductions VectorMinCoeff_double/256 7.7 7.7 -0.49% bench_reductions VectorMinCoeff_double/262144 16256.2 16575.8 +1.97% ▲ bench_reductions VectorMinCoeff_double/4096 138.7 139.3 +0.46% bench_reductions VectorMinCoeff_double/64 2.1 2.1 -1.14% ▼ bench_reductions VectorMinCoeff_double/65536 2196.2 2327.2 +5.97% ▲ bench_reductions VectorMinCoeff_float/1024 16.1 16.0 -0.47% bench_reductions VectorMinCoeff_float/1048576 37091.7 37240.6 +0.40% bench_reductions VectorMinCoeff_float/16384 319.3 321.2 +0.59% bench_reductions VectorMinCoeff_float/256 4.1 4.1 +0.15% bench_reductions VectorMinCoeff_float/262144 7046.8 6523.1 -7.43% ▼ bench_reductions VectorMinCoeff_float/4096 69.6 64.3 -7.59% ▼ bench_reductions VectorMinCoeff_float/64 1.3 1.3 -0.67% bench_reductions VectorMinCoeff_float/65536 1280.8 1277.2 -0.29% bench_reductions VectorNorm_double/1024 43.3 43.6 +0.50% bench_reductions VectorNorm_double/1048576 67634.2 68943.0 +1.94% ▲ bench_reductions VectorNorm_double/16384 804.6 740.1 -8.01% ▼ bench_reductions VectorNorm_double/256 10.9 11.1 +1.48% ▲ bench_reductions VectorNorm_double/262144 16731.0 16979.3 +1.48% ▲ bench_reductions VectorNorm_double/4096 178.0 178.8 +0.46% bench_reductions VectorNorm_double/64 3.3 3.3 +0.85% bench_reductions VectorNorm_double/65536 3356.5 2886.4 -14.01% ▼ bench_reductions VectorNorm_float/1024 21.4 21.6 +1.37% ▲ bench_reductions VectorNorm_float/1048576 36636.9 37648.4 +2.76% ▲ bench_reductions VectorNorm_float/16384 405.9 400.5 -1.33% ▼ bench_reductions VectorNorm_float/256 6.2 6.2 +0.08% bench_reductions VectorNorm_float/262144 6883.8 7416.6 +7.74% ▲ bench_reductions VectorNorm_float/4096 89.2 89.1 -0.10% bench_reductions VectorNorm_float/64 2.1 2.0 -2.33% ▼ bench_reductions VectorNorm_float/65536 1594.2 1630.7 +2.29% ▲ bench_reductions VectorProd_double/1024 33.2 31.8 -4.21% ▼ bench_reductions VectorProd_double/1048576 68446.6 67187.9 -1.84% ▼ bench_reductions VectorProd_double/16384 692.2 669.5 -3.28% ▼ bench_reductions VectorProd_double/256 7.2 7.4 +2.25% ▲ bench_reductions VectorProd_double/262144 16738.7 16544.5 -1.16% ▼ bench_reductions VectorProd_double/4096 154.5 156.2 +1.10% ▲ bench_reductions VectorProd_double/64 2.1 2.0 -3.98% ▼ bench_reductions VectorProd_double/65536 2734.1 2680.5 -1.96% ▼ bench_reductions VectorProd_float/1024 15.3 15.1 -1.35% ▼ bench_reductions VectorProd_float/1048576 37313.2 36974.6 -0.91% bench_reductions VectorProd_float/16384 341.7 341.4 -0.08% bench_reductions VectorProd_float/256 3.9 3.9 -0.72% bench_reductions VectorProd_float/262144 7053.4 6525.5 -7.48% ▼ bench_reductions VectorProd_float/4096 71.9 74.0 +2.92% ▲ bench_reductions VectorProd_float/64 1.3 1.3 -0.53% bench_reductions VectorProd_float/65536 1392.0 1350.1 -3.01% ▼ bench_reductions VectorSquaredNorm_double/1024 39.5 40.0 +1.43% ▲ bench_reductions VectorSquaredNorm_double/1048576 67425.7 70274.3 +4.22% ▲ bench_reductions VectorSquaredNorm_double/16384 816.9 814.5 -0.29% bench_reductions VectorSquaredNorm_double/256 9.8 9.7 -0.59% bench_reductions VectorSquaredNorm_double/262144 16684.2 17172.9 +2.93% ▲ bench_reductions VectorSquaredNorm_double/4096 174.4 177.4 +1.74% ▲ bench_reductions VectorSquaredNorm_double/64 2.6 2.7 +0.88% bench_reductions VectorSquaredNorm_double/65536 3420.4 3399.2 -0.62% bench_reductions VectorSquaredNorm_float/1024 20.5 20.7 +0.92% bench_reductions VectorSquaredNorm_float/1048576 37372.5 37357.5 -0.04% bench_reductions VectorSquaredNorm_float/16384 377.4 382.5 +1.37% ▲ bench_reductions VectorSquaredNorm_float/256 5.3 5.4 +0.78% bench_reductions VectorSquaredNorm_float/262144 7175.9 7316.2 +1.95% ▲ bench_reductions VectorSquaredNorm_float/4096 85.5 86.7 +1.36% ▲ bench_reductions VectorSquaredNorm_float/64 1.7 1.7 -0.34% bench_reductions VectorSquaredNorm_float/65536 1480.8 1499.3 +1.25% ▲ bench_reductions VectorSum_double/1024 33.2 33.2 +0.23% bench_reductions VectorSum_double/1048576 66040.4 71747.3 +8.64% ▲ bench_reductions VectorSum_double/16384 709.1 713.5 +0.63% bench_reductions VectorSum_double/256 7.3 7.4 +1.45% ▲ bench_reductions VectorSum_double/262144 16444.3 16756.2 +1.90% ▲ bench_reductions VectorSum_double/4096 152.1 154.0 +1.27% ▲ bench_reductions VectorSum_double/64 2.0 2.0 +0.82% bench_reductions VectorSum_double/65536 2794.8 2792.9 -0.07% bench_reductions VectorSum_float/1024 14.9 14.9 +0.09% bench_reductions VectorSum_float/1048576 37097.2 37044.3 -0.14% bench_reductions VectorSum_float/16384 398.5 398.0 -0.14% bench_reductions VectorSum_float/256 3.8 3.8 +0.53% bench_reductions VectorSum_float/262144 7135.8 6667.1 -6.57% ▼ bench_reductions VectorSum_float/4096 71.8 71.2 -0.79% bench_reductions VectorSum_float/64 1.3 1.3 +0.56% bench_reductions VectorSum_float/65536 1586.6 1608.8 +1.40% ▲ bench_fixed_size BM_Batch3x3Gemm/100 573.3 582.7 +1.64% ▲ bench_fixed_size BM_Batch3x3Gemm/1000 5939.0 6034.3 +1.60% ▲ bench_fixed_size BM_Batch3x3Gemm/10000 61323.3 61935.6 +1.00% bench_fixed_size BM_BatchTransform4xN/1 2.0 2.0 +1.62% ▲ bench_fixed_size BM_BatchTransform4xN/16 15.0 15.2 +1.46% ▲ bench_fixed_size BM_BatchTransform4xN/4 4.7 4.8 +1.42% ▲ bench_fixed_size BM_BatchTransform4xN/64 56.6 57.5 +1.50% ▲ bench_fixed_size BM_BatchTransform4xN/8 8.1 8.3 +1.67% ▲ bench_fixed_size FixedDet_2x2 0.2 0.2 +0.04% bench_fixed_size FixedDet_3x3 0.2 0.2 -0.13% bench_fixed_size FixedDet_4x4 3.0 3.0 +0.58% bench_fixed_size FixedGemm_2x2 0.2 0.2 +0.50% bench_fixed_size FixedGemm_3x3 6.9 7.0 +1.57% ▲ bench_fixed_size FixedGemm_4x4 2.2 2.3 +1.77% ▲ bench_fixed_size FixedInverse_2x2 0.5 0.5 +0.43% bench_fixed_size FixedInverse_3x3 2.0 2.0 -0.61% bench_fixed_size FixedInverse_4x4 6.0 6.0 +0.59% bench_fixed_size_double BM_Batch3x3Gemm/100 267.4 271.8 +1.61% ▲ bench_fixed_size_double BM_Batch3x3Gemm/1000 2679.5 2702.5 +0.86% bench_fixed_size_double BM_Batch3x3Gemm/10000 27762.2 28071.9 +1.12% ▲ bench_fixed_size_double BM_BatchTransform4xN/1 2.1 2.1 -0.70% bench_fixed_size_double BM_BatchTransform4xN/16 15.1 15.3 +1.45% ▲ bench_fixed_size_double BM_BatchTransform4xN/4 4.7 4.8 +0.88% bench_fixed_size_double BM_BatchTransform4xN/64 58.1 58.6 +0.84% bench_fixed_size_double BM_BatchTransform4xN/8 8.2 8.9 +8.81% ▲ bench_fixed_size_double FixedDet_2x2 0.2 0.2 +0.14% bench_fixed_size_double FixedDet_3x3 1.1 1.1 +1.71% ▲ bench_fixed_size_double FixedDet_4x4 3.1 3.0 -0.51% bench_fixed_size_double FixedGemm_2x2 0.3 0.3 +0.65% bench_fixed_size_double FixedGemm_3x3 2.6 2.6 +0.56% bench_fixed_size_double FixedGemm_4x4 2.4 2.4 +0.55% bench_fixed_size_double FixedInverse_2x2 0.4 0.4 +0.43% bench_fixed_size_double FixedInverse_3x3 4.4 4.4 +0.13% bench_fixed_size_double FixedInverse_4x4 7.2 7.3 +1.58% ▲ bench_trsm TRSM_Left_double_Lower/256/1 7935.8 8017.7 +1.03% ▲ bench_trsm TRSM_Left_double_Lower/256/16 15704.6 15914.0 +1.33% ▲ bench_trsm TRSM_Left_double_Lower/256/64 67457.6 68039.5 +0.86% bench_trsm TRSM_Left_double_Lower/512/1 39199.8 39861.5 +1.69% ▲ bench_trsm TRSM_Left_double_Lower/512/16 67026.8 68321.1 +1.93% ▲ bench_trsm TRSM_Left_double_Lower/512/64 271500.5 275576.3 +1.50% ▲ bench_trsm TRSM_Left_double_Lower/64/1 649.0 656.1 +1.08% ▲ bench_trsm TRSM_Left_double_Lower/64/16 1207.2 1233.0 +2.14% ▲ bench_trsm TRSM_Left_double_Lower/64/64 4720.7 4791.1 +1.49% ▲ bench_trsm TRSM_Left_float_Lower/256/1 7907.9 8014.8 +1.35% ▲ bench_trsm TRSM_Left_float_Lower/256/16 8131.1 8247.5 +1.43% ▲ bench_trsm TRSM_Left_float_Lower/256/64 32145.5 32693.2 +1.70% ▲ bench_trsm TRSM_Left_float_Lower/512/1 30019.2 30564.0 +1.81% ▲ bench_trsm TRSM_Left_float_Lower/512/16 31127.9 31619.5 +1.58% ▲ bench_trsm TRSM_Left_float_Lower/512/64 134503.3 139319.8 +3.58% ▲ bench_trsm TRSM_Left_float_Lower/64/1 672.1 671.8 -0.06% bench_trsm TRSM_Left_float_Lower/64/16 694.8 705.2 +1.49% ▲ bench_trsm TRSM_Left_float_Lower/64/64 2439.9 2459.3 +0.79% bench_trsm TRSM_Right_double_Lower/256/1 7867.1 7826.7 -0.51% bench_trsm TRSM_Right_double_Lower/256/16 15003.0 14913.1 -0.60% bench_trsm TRSM_Right_double_Lower/256/64 58727.3 59490.0 +1.30% ▲ bench_trsm TRSM_Right_double_Lower/512/1 31126.7 30747.2 -1.22% ▼ bench_trsm TRSM_Right_double_Lower/512/16 60374.5 59923.4 -0.75% bench_trsm TRSM_Right_double_Lower/512/64 235760.5 238783.2 +1.28% ▲ bench_trsm TRSM_Right_double_Lower/64/1 606.3 600.8 -0.91% bench_trsm TRSM_Right_double_Lower/64/16 1037.5 1031.1 -0.62% bench_trsm TRSM_Right_double_Lower/64/64 4005.0 3972.1 -0.82% bench_trsm TRSM_Right_float_Lower/256/1 7803.0 7836.0 +0.42% bench_trsm TRSM_Right_float_Lower/256/16 7598.3 7630.0 +0.42% bench_trsm TRSM_Right_float_Lower/256/64 29265.2 29365.4 +0.34% bench_trsm TRSM_Right_float_Lower/512/1 29729.9 29902.7 +0.58% bench_trsm TRSM_Right_float_Lower/512/16 29682.5 29945.1 +0.88% bench_trsm TRSM_Right_float_Lower/512/64 117387.7 116958.2 -0.37% bench_trsm TRSM_Right_float_Lower/64/1 596.7 597.4 +0.12% bench_trsm TRSM_Right_float_Lower/64/16 573.4 573.0 -0.07% bench_trsm TRSM_Right_float_Lower/64/64 1940.6 1946.2 +0.28% bench_trsm TRSV_double_Lower/128 885.6 982.7 +10.96% ▲ bench_trsm TRSV_double_Lower/32 187.7 199.7 +6.44% ▲ bench_trsm TRSV_double_Lower/512 10402.6 11068.0 +6.40% ▲ bench_trsm TRSV_float_Lower/128 754.9 887.9 +17.61% ▲ bench_trsm TRSV_float_Lower/32 167.1 185.8 +11.15% ▲ bench_trsm TRSV_float_Lower/512 5766.2 6849.2 +18.78% ▲ bench_trmv TRMV_double_Lower/1024 38573.3 39622.2 +2.72% ▲ bench_trmv TRMV_double_Lower/128 613.1 620.1 +1.13% ▲ bench_trmv TRMV_double_Lower/16 40.0 40.3 +0.56% bench_trmv TRMV_double_Lower/2048 310812.0 321935.3 +3.58% ▲ bench_trmv TRMV_double_Lower/256 2098.1 2153.7 +2.65% ▲ bench_trmv TRMV_double_Lower/32 87.0 88.0 +1.08% ▲ bench_trmv TRMV_double_Lower/512 9860.9 9683.6 -1.80% ▼ bench_trmv TRMV_double_Lower/64 211.8 213.0 +0.57% bench_trmv TRMV_double_Lower/8 19.8 19.5 -1.58% ▼ bench_trmv TRMV_double_UnitLower/1024 38480.5 38430.8 -0.13% bench_trmv TRMV_double_UnitLower/128 589.4 589.7 +0.05% bench_trmv TRMV_double_UnitLower/16 39.4 39.7 +0.68% bench_trmv TRMV_double_UnitLower/2048 314347.7 319882.7 +1.76% ▲ bench_trmv TRMV_double_UnitLower/256 2063.4 2070.9 +0.36% bench_trmv TRMV_double_UnitLower/32 83.3 85.0 +2.04% ▲ bench_trmv TRMV_double_UnitLower/512 9752.1 9519.9 -2.38% ▼ bench_trmv TRMV_double_UnitLower/64 202.9 203.1 +0.10% bench_trmv TRMV_double_UnitLower/8 19.7 19.8 +0.71% bench_trmv TRMV_double_UnitUpper/1024 39773.1 39416.2 -0.90% bench_trmv TRMV_double_UnitUpper/128 601.8 596.4 -0.91% bench_trmv TRMV_double_UnitUpper/16 38.6 38.3 -0.91% bench_trmv TRMV_double_UnitUpper/2048 320043.1 370058.3 +15.63% ▲ bench_trmv TRMV_double_UnitUpper/256 2097.4 2084.2 -0.63% bench_trmv TRMV_double_UnitUpper/32 82.9 83.5 +0.67% bench_trmv TRMV_double_UnitUpper/512 9457.9 10050.7 +6.27% ▲ bench_trmv TRMV_double_UnitUpper/64 210.5 209.4 -0.53% bench_trmv TRMV_double_UnitUpper/8 19.4 19.5 +0.80% bench_trmv TRMV_double_Upper/1024 39159.5 39470.1 +0.79% bench_trmv TRMV_double_Upper/128 609.5 612.0 +0.42% bench_trmv TRMV_double_Upper/16 39.4 39.7 +0.65% bench_trmv TRMV_double_Upper/2048 306240.0 333125.3 +8.78% ▲ bench_trmv TRMV_double_Upper/256 2095.6 2108.8 +0.63% bench_trmv TRMV_double_Upper/32 83.3 84.4 +1.36% ▲ bench_trmv TRMV_double_Upper/512 9552.7 9869.0 +3.31% ▲ bench_trmv TRMV_double_Upper/64 209.3 210.5 +0.57% bench_trmv TRMV_double_Upper/8 20.4 20.5 +0.58% bench_trmv TRMV_float_Lower/1024 27333.3 26215.2 -4.09% ▼ bench_trmv TRMV_float_Lower/128 540.6 535.6 -0.92% bench_trmv TRMV_float_Lower/16 45.4 44.8 -1.32% ▼ bench_trmv TRMV_float_Lower/2048 96811.3 93008.6 -3.93% ▼ bench_trmv TRMV_float_Lower/256 1555.1 1562.9 +0.51% bench_trmv TRMV_float_Lower/32 94.8 93.6 -1.24% ▼ bench_trmv TRMV_float_Lower/512 5698.9 5216.7 -8.46% ▼ bench_trmv TRMV_float_Lower/64 209.6 210.2 +0.29% bench_trmv TRMV_float_Lower/8 22.2 22.6 +1.54% ▲ bench_trmv TRMV_float_UnitLower/1024 27241.0 26973.9 -0.98% bench_trmv TRMV_float_UnitLower/128 530.8 542.9 +2.28% ▲ bench_trmv TRMV_float_UnitLower/16 44.1 44.5 +0.73% bench_trmv TRMV_float_UnitLower/2048 95510.1 95958.5 +0.47% bench_trmv TRMV_float_UnitLower/256 1534.1 1565.5 +2.05% ▲ bench_trmv TRMV_float_UnitLower/32 92.1 92.5 +0.53% bench_trmv TRMV_float_UnitLower/512 5441.0 5176.6 -4.86% ▼ bench_trmv TRMV_float_UnitLower/64 203.5 211.0 +3.70% ▲ bench_trmv TRMV_float_UnitLower/8 22.3 22.8 +2.24% ▲ bench_trmv TRMV_float_UnitUpper/1024 20204.3 24021.0 +18.89% ▲ bench_trmv TRMV_float_UnitUpper/128 443.3 464.2 +4.72% ▲ bench_trmv TRMV_float_UnitUpper/16 37.9 41.0 +8.19% ▲ bench_trmv TRMV_float_UnitUpper/2048 87226.6 87823.4 +0.68% bench_trmv TRMV_float_UnitUpper/256 1250.6 1297.6 +3.76% ▲ bench_trmv TRMV_float_UnitUpper/32 77.4 87.3 +12.78% ▲ bench_trmv TRMV_float_UnitUpper/512 4432.4 4463.5 +0.70% bench_trmv TRMV_float_UnitUpper/64 167.4 186.1 +11.20% ▲ bench_trmv TRMV_float_UnitUpper/8 19.9 20.4 +2.52% ▲ bench_trmv TRMV_float_Upper/1024 20507.4 24302.7 +18.51% ▲ bench_trmv TRMV_float_Upper/128 462.5 473.8 +2.43% ▲ bench_trmv TRMV_float_Upper/16 39.6 39.9 +1.00% bench_trmv TRMV_float_Upper/2048 85229.1 91553.5 +7.42% ▲ bench_trmv TRMV_float_Upper/256 1300.0 1329.9 +2.30% ▲ bench_trmv TRMV_float_Upper/32 87.0 84.0 -3.49% ▼ bench_trmv TRMV_float_Upper/512 4701.7 4525.7 -3.74% ▼ bench_trmv TRMV_float_Upper/64 179.8 181.4 +0.93% bench_trmv TRMV_float_Upper/8 20.3 20.4 +0.34% bench_triangular_product TRMM_Left_double_Lower/1024 16988986.8 18346262.7 +7.99% ▲ bench_triangular_product TRMM_Left_double_Lower/128 38644.1 38987.4 +0.89% bench_triangular_product TRMM_Left_double_Lower/256 278590.1 279609.4 +0.37% bench_triangular_product TRMM_Left_double_Lower/512 2033478.4 2072286.6 +1.91% ▲ bench_triangular_product TRMM_Left_double_Lower/64 6261.9 6293.8 +0.51% bench_triangular_product TRMM_Left_double_Upper/1024 16915669.5 17261509.9 +2.04% ▲ bench_triangular_product TRMM_Left_double_Upper/128 38574.3 39146.1 +1.48% ▲ bench_triangular_product TRMM_Left_double_Upper/256 277355.6 279616.5 +0.82% bench_triangular_product TRMM_Left_double_Upper/512 2041020.8 2076415.3 +1.73% ▲ bench_triangular_product TRMM_Left_double_Upper/64 6263.8 6334.7 +1.13% ▲ bench_triangular_product TRMM_Left_float_Lower/1024 8656390.8 8619707.7 -0.42% bench_triangular_product TRMM_Left_float_Lower/128 23228.2 23582.4 +1.52% ▲ bench_triangular_product TRMM_Left_float_Lower/256 148974.3 151086.7 +1.42% ▲ bench_triangular_product TRMM_Left_float_Lower/512 1075967.9 1075817.4 -0.01% bench_triangular_product TRMM_Left_float_Lower/64 4119.1 4127.6 +0.21% bench_triangular_product TRMM_Left_float_Upper/1024 8601343.7 8763818.4 +1.89% ▲ bench_triangular_product TRMM_Left_float_Upper/128 23073.7 23122.5 +0.21% bench_triangular_product TRMM_Left_float_Upper/256 149201.7 149504.7 +0.20% bench_triangular_product TRMM_Left_float_Upper/512 1070202.6 1092099.8 +2.05% ▲ bench_triangular_product TRMM_Left_float_Upper/64 4072.7 4084.6 +0.29% bench_triangular_product TRMM_Right_double_Lower/1024 16344401.4 16446695.8 +0.63% bench_triangular_product TRMM_Right_double_Lower/128 36150.5 36502.1 +0.97% bench_triangular_product TRMM_Right_double_Lower/256 266064.5 268329.9 +0.85% bench_triangular_product TRMM_Right_double_Lower/512 2003757.2 2048420.9 +2.23% ▲ bench_triangular_product TRMM_Right_double_Lower/64 5560.1 5621.7 +1.11% ▲ bench_triangular_product TRMM_Right_double_Upper/1024 16998547.1 17182655.0 +1.08% ▲ bench_triangular_product TRMM_Right_double_Upper/128 36345.7 36219.0 -0.35% bench_triangular_product TRMM_Right_double_Upper/256 268348.9 265502.8 -1.06% ▼ bench_triangular_product TRMM_Right_double_Upper/512 2063857.1 2031211.5 -1.58% ▼ bench_triangular_product TRMM_Right_double_Upper/64 5554.8 5569.1 +0.26% bench_triangular_product TRMM_Right_float_Lower/1024 8187634.0 8498791.2 +3.80% ▲ bench_triangular_product TRMM_Right_float_Lower/128 20584.3 20595.3 +0.05% bench_triangular_product TRMM_Right_float_Lower/256 138120.3 138212.7 +0.07% bench_triangular_product TRMM_Right_float_Lower/512 1037499.6 1044861.1 +0.71% bench_triangular_product TRMM_Right_float_Lower/64 3319.7 3337.8 +0.54% bench_triangular_product TRMM_Right_float_Upper/1024 8509560.3 8709233.5 +2.35% ▲ bench_triangular_product TRMM_Right_float_Upper/128 20617.0 20644.9 +0.14% bench_triangular_product TRMM_Right_float_Upper/256 138715.9 139684.5 +0.70% bench_triangular_product TRMM_Right_float_Upper/512 1027494.5 1052578.9 +2.44% ▲ bench_triangular_product TRMM_Right_float_Upper/64 3329.6 3331.1 +0.04% bench_selfadjoint_product RankUpdate_double/128/16 4962.7 5000.9 +0.77% bench_selfadjoint_product RankUpdate_double/128/256 74029.5 75019.5 +1.34% ▲ bench_selfadjoint_product RankUpdate_double/128/64 18972.6 19024.3 +0.27% bench_selfadjoint_product RankUpdate_double/256/16 17327.9 17467.0 +0.80% bench_selfadjoint_product RankUpdate_double/256/256 270805.7 272972.5 +0.80% bench_selfadjoint_product RankUpdate_double/256/64 66595.8 67087.2 +0.74% bench_selfadjoint_product RankUpdate_double/512/16 69050.2 67172.0 -2.72% ▼ bench_selfadjoint_product RankUpdate_double/512/256 1008644.2 1018862.4 +1.01% ▲ bench_selfadjoint_product RankUpdate_double/512/64 252749.4 255730.3 +1.18% ▲ bench_selfadjoint_product RankUpdate_double/64/16 1540.7 1568.6 +1.81% ▲ bench_selfadjoint_product RankUpdate_double/64/256 22975.6 23184.7 +0.91% bench_selfadjoint_product RankUpdate_double/64/64 5841.5 5849.3 +0.13% bench_selfadjoint_product RankUpdate_float/128/16 3009.0 3016.6 +0.25% bench_selfadjoint_product RankUpdate_float/128/256 43429.7 44017.7 +1.35% ▲ bench_selfadjoint_product RankUpdate_float/128/64 10901.7 10963.5 +0.57% bench_selfadjoint_product RankUpdate_float/256/16 9727.9 9693.0 -0.36% bench_selfadjoint_product RankUpdate_float/256/256 144548.1 148067.4 +2.43% ▲ bench_selfadjoint_product RankUpdate_float/256/64 36540.8 36847.8 +0.84% bench_selfadjoint_product RankUpdate_float/512/16 34027.5 34317.5 +0.85% bench_selfadjoint_product RankUpdate_float/512/256 528932.5 532808.3 +0.73% bench_selfadjoint_product RankUpdate_float/512/64 133800.5 134140.6 +0.25% bench_selfadjoint_product RankUpdate_float/64/16 1038.3 1038.5 +0.02% bench_selfadjoint_product RankUpdate_float/64/256 13644.5 13776.2 +0.96% bench_selfadjoint_product RankUpdate_float/64/64 3460.7 3443.8 -0.49% bench_selfadjoint_product SYMM_Left_double/1024 31313423.7 31896644.1 +1.86% ▲ bench_selfadjoint_product SYMM_Left_double/128 63985.6 64773.9 +1.23% ▲ bench_selfadjoint_product SYMM_Left_double/256 490380.9 488367.5 -0.41% bench_selfadjoint_product SYMM_Left_double/512 3819699.0 3861806.9 +1.10% ▲ bench_selfadjoint_product SYMM_Left_double/64 8688.9 8759.2 +0.81% bench_selfadjoint_product SYMM_Left_float/1024 15162481.0 15286161.4 +0.82% bench_selfadjoint_product SYMM_Left_float/128 31765.6 32098.0 +1.05% ▲ bench_selfadjoint_product SYMM_Left_float/256 256447.2 258820.5 +0.93% bench_selfadjoint_product SYMM_Left_float/512 1891761.1 1905225.4 +0.71% bench_selfadjoint_product SYMM_Left_float/64 4414.7 4486.4 +1.62% ▲ bench_selfadjoint_product SYMM_Right_double/1024 31633335.9 32010692.0 +1.19% ▲ bench_selfadjoint_product SYMM_Right_double/128 62330.7 62701.3 +0.59% bench_selfadjoint_product SYMM_Right_double/256 485787.0 485319.1 -0.10% bench_selfadjoint_product SYMM_Right_double/512 3819352.7 3853814.3 +0.90% bench_selfadjoint_product SYMM_Right_double/64 8438.9 8446.9 +0.09% bench_selfadjoint_product SYMM_Right_float/1024 15227232.8 15321116.2 +0.62% bench_selfadjoint_product SYMM_Right_float/128 31383.0 31751.0 +1.17% ▲ bench_selfadjoint_product SYMM_Right_float/256 240798.9 244540.7 +1.55% ▲ bench_selfadjoint_product SYMM_Right_float/512 1879660.6 1904209.4 +1.31% ▲ bench_selfadjoint_product SYMM_Right_float/64 4312.9 4388.7 +1.76% ▲ bench_dot Dot_cdouble/1024 200.3 201.9 +0.82% bench_dot Dot_cdouble/1048576 607010.8 591717.3 -2.52% ▼ bench_dot Dot_cdouble/16384 3207.8 3226.4 +0.58% bench_dot Dot_cdouble/256 51.9 51.8 -0.20% bench_dot Dot_cdouble/262144 75530.4 69072.6 -8.55% ▼ bench_dot Dot_cdouble/4096 805.1 806.4 +0.16% bench_dot Dot_cdouble/64 14.0 15.3 +9.28% ▲ bench_dot Dot_cdouble/65536 16911.4 16947.9 +0.22% bench_dot Dot_cfloat/1024 88.3 89.2 +1.00% bench_dot Dot_cfloat/1048576 192374.7 222690.5 +15.76% ▲ bench_dot Dot_cfloat/16384 1363.0 1375.4 +0.91% bench_dot Dot_cfloat/256 24.9 25.0 +0.05% bench_dot Dot_cfloat/262144 35694.8 36443.2 +2.10% ▲ bench_dot Dot_cfloat/4096 345.6 349.5 +1.12% ▲ bench_dot Dot_cfloat/64 6.8 6.8 +0.97% bench_dot Dot_cfloat/65536 6885.8 6647.2 -3.46% ▼ bench_dot Dot_double/1024 55.1 55.3 +0.49% bench_dot Dot_double/1048576 161168.6 160179.0 -0.61% bench_dot Dot_double/16384 1189.7 1206.5 +1.41% ▲ bench_dot Dot_double/256 13.8 14.3 +3.33% ▲ bench_dot Dot_double/262144 36150.7 36312.9 +0.45% bench_dot Dot_double/4096 304.6 296.5 -2.65% ▼ bench_dot Dot_double/64 3.5 3.6 +1.09% ▲ bench_dot Dot_double/65536 5666.9 6561.4 +15.78% ▲ bench_dot Dot_float/1024 27.6 27.6 +0.06% bench_dot Dot_float/1048576 64164.1 63161.7 -1.56% ▼ bench_dot Dot_float/16384 602.5 604.7 +0.38% bench_dot Dot_float/256 7.1 7.0 -0.74% bench_dot Dot_float/262144 16230.0 16567.9 +2.08% ▲ bench_dot Dot_float/4096 109.8 110.8 +0.85% bench_dot Dot_float/64 2.0 2.0 +1.19% ▲ bench_dot Dot_float/65536 1997.3 2029.0 +1.59% ▲ bench_dot SquaredNorm_double/1024 40.0 40.0 +0.05% bench_dot SquaredNorm_double/1048576 79659.1 71832.5 -9.83% ▼ bench_dot SquaredNorm_double/16384 754.6 753.5 -0.14% bench_dot SquaredNorm_double/256 9.7 10.1 +4.03% ▲ bench_dot SquaredNorm_double/262144 16949.8 17529.8 +3.42% ▲ bench_dot SquaredNorm_double/4096 176.7 175.9 -0.41% bench_dot SquaredNorm_double/64 2.7 2.7 +1.31% ▲ bench_dot SquaredNorm_double/65536 2921.8 2931.3 +0.33% bench_dot SquaredNorm_float/1024 20.6 20.8 +0.67% bench_dot SquaredNorm_float/1048576 38066.8 39128.3 +2.79% ▲ bench_dot SquaredNorm_float/16384 406.3 410.1 +0.93% bench_dot SquaredNorm_float/256 5.4 5.4 +0.66% bench_dot SquaredNorm_float/262144 6783.2 7661.6 +12.95% ▲ bench_dot SquaredNorm_float/4096 86.0 86.7 +0.84% bench_dot SquaredNorm_float/64 1.7 1.7 +0.77% bench_dot SquaredNorm_float/65536 1591.5 1613.5 +1.38% ▲ bench_block_ops BlockRead_double/1024/128 1770.5 1774.5 +0.22% bench_block_ops BlockRead_double/1024/16 29.7 30.6 +3.10% ▲ bench_block_ops BlockRead_double/1024/64 269.8 278.4 +3.22% ▲ bench_block_ops BlockRead_double/256/128 1668.0 1743.0 +4.50% ▲ bench_block_ops BlockRead_double/256/16 28.4 29.0 +2.32% ▲ bench_block_ops BlockRead_double/256/64 274.7 290.2 +5.66% ▲ bench_block_ops BlockRead_double/512/128 1760.7 1753.2 -0.43% bench_block_ops BlockRead_double/512/16 30.4 31.1 +2.35% ▲ bench_block_ops BlockRead_double/512/64 267.7 268.1 +0.13% bench_block_ops BlockRead_float/1024/128 876.8 894.3 +1.99% ▲ bench_block_ops BlockRead_float/1024/16 28.3 29.0 +2.44% ▲ bench_block_ops BlockRead_float/1024/64 169.1 155.8 -7.86% ▼ bench_block_ops BlockRead_float/256/128 788.4 811.6 +2.94% ▲ bench_block_ops BlockRead_float/256/16 27.5 28.3 +2.96% ▲ bench_block_ops BlockRead_float/256/64 160.3 159.7 -0.36% bench_block_ops BlockRead_float/512/128 867.1 900.8 +3.89% ▲ bench_block_ops BlockRead_float/512/16 27.4 28.1 +2.62% ▲ bench_block_ops BlockRead_float/512/64 150.7 268.7 +78.29% ▲ bench_block_ops BlockWrite_double/1024/128 1837.1 1802.8 -1.87% ▼ bench_block_ops BlockWrite_double/1024/16 42.8 41.6 -2.79% ▼ bench_block_ops BlockWrite_double/1024/64 262.9 270.3 +2.80% ▲ bench_block_ops BlockWrite_double/256/128 1664.3 1734.3 +4.21% ▲ bench_block_ops BlockWrite_double/256/16 33.1 33.5 +0.95% bench_block_ops BlockWrite_double/256/64 283.1 291.7 +3.06% ▲ bench_block_ops BlockWrite_double/512/128 1783.4 1800.2 +0.94% bench_block_ops BlockWrite_double/512/16 49.5 49.8 +0.60% bench_block_ops BlockWrite_double/512/64 263.3 270.4 +2.70% ▲ bench_block_ops BlockWrite_float/1024/128 831.4 839.9 +1.02% ▲ bench_block_ops BlockWrite_float/1024/16 60.9 61.1 +0.28% bench_block_ops BlockWrite_float/1024/64 177.7 181.1 +1.92% ▲ bench_block_ops BlockWrite_float/256/128 845.7 864.0 +2.17% ▲ bench_block_ops BlockWrite_float/256/16 58.8 60.2 +2.26% ▲ bench_block_ops BlockWrite_float/256/64 174.5 174.0 -0.28% bench_block_ops BlockWrite_float/512/128 829.6 874.3 +5.38% ▲ bench_block_ops BlockWrite_float/512/16 59.0 60.3 +2.25% ▲ bench_block_ops BlockWrite_float/512/64 173.8 170.0 -2.17% ▼ bench_block_ops LeftCols_double/1024/128 15901.9 16059.3 +0.99% bench_block_ops LeftCols_double/1024/16 1180.9 1202.3 +1.82% ▲ bench_block_ops LeftCols_double/1024/64 5570.8 5671.7 +1.81% ▲ bench_block_ops LeftCols_double/256/128 1827.0 1853.5 +1.45% ▲ bench_block_ops LeftCols_double/256/16 281.6 288.2 +2.36% ▲ bench_block_ops LeftCols_double/256/64 1181.4 1203.7 +1.89% ▲ bench_block_ops LeftCols_double/512/128 6104.8 5455.9 -10.63% ▼ bench_block_ops LeftCols_double/512/16 592.0 608.0 +2.72% ▲ bench_block_ops LeftCols_double/512/64 1826.8 1853.3 +1.45% ▲ bench_block_ops LeftCols_float/1024/128 5515.0 5747.2 +4.21% ▲ bench_block_ops LeftCols_float/1024/16 592.9 599.8 +1.16% ▲ bench_block_ops LeftCols_float/1024/64 1859.2 1853.9 -0.29% bench_block_ops LeftCols_float/256/128 1198.6 1194.9 -0.31% bench_block_ops LeftCols_float/256/16 81.7 72.5 -11.22% ▼ bench_block_ops LeftCols_float/256/64 592.7 601.5 +1.48% ▲ bench_block_ops LeftCols_float/512/128 1866.0 1851.7 -0.77% bench_block_ops LeftCols_float/512/16 282.1 283.8 +0.59% bench_block_ops LeftCols_float/512/64 1182.9 1196.5 +1.15% ▲ bench_block_ops TopRows_double/1024/128 18376.9 17198.1 -6.41% ▼ bench_block_ops TopRows_double/1024/16 2587.8 2647.5 +2.30% ▲ bench_block_ops TopRows_double/1024/64 8110.4 8186.4 +0.94% bench_block_ops TopRows_double/256/128 3555.3 3390.5 -4.63% ▼ bench_block_ops TopRows_double/256/16 470.6 475.5 +1.05% ▲ bench_block_ops TopRows_double/256/64 1794.9 1789.5 -0.30% bench_block_ops TopRows_double/512/128 7764.6 7091.0 -8.68% ▼ bench_block_ops TopRows_double/512/16 956.4 977.2 +2.17% ▲ bench_block_ops TopRows_double/512/64 3712.1 3712.8 +0.02% bench_block_ops TopRows_float/1024/128 8418.5 8643.3 +2.67% ▲ bench_block_ops TopRows_float/1024/16 2025.5 2037.5 +0.59% bench_block_ops TopRows_float/1024/64 4597.7 4603.9 +0.13% bench_block_ops TopRows_float/256/128 1726.2 1670.9 -3.21% ▼ bench_block_ops TopRows_float/256/16 455.4 454.0 -0.30% bench_block_ops TopRows_float/256/64 923.2 943.6 +2.21% ▲ bench_block_ops TopRows_float/512/128 3658.2 3756.5 +2.69% ▲ bench_block_ops TopRows_float/512/16 904.6 932.3 +3.07% ▲ bench_block_ops TopRows_float/512/64 1913.1 1913.1 +0.00% Total: 693 cases