Skip to content

mtl5_bench capability card -- ryzen-7-2700x

The platform benchmark’s capability cards for the ryzen-7-2700x machine described on the benchmark systems page: every operation family, measured against this machine’s FMA peak and memory bandwidth (#721). This page shows the latest published session; every session’s raw results stay in the data directory.

Preliminary. Measured outside #737’s methodology: cores not pinned, the machine’s clock policy not applied, one session, so no repeatability band.

session20261108T223729Z (UTC)
commite096dbbc6788
classmedium
runsnative-fast at 1 thread, native-fast at 16 threads, native at 1 thread, native at 16 threads
failed checks0
CPUs16 available, pinning not recorded
clock policynot recorded (governor Sonos)
raw databenchmarks/data/ryzen-7-2700x/mtl5_bench/20261108T223729Z/
earlier sessionsnone

Each column is one run of this session: the build variant and its thread count (1T is one thread). The runs follow below, each with its full entry table.

areanative 1Tnative 16Tnative-fast 1Tnative-fast 16T
blas140%91%85%112%
blas240%41%89%81%
blas326%4%27%5%
factor4%1%3%1%
eigen7%1%6%1%
elementwise10%8%9%9%
sparse37%32%34%33%
direct1%0%1%0%
orderingn/an/an/an/a
krylovn/an/an/an/a
pcn/an/an/an/a
smoother53%47%52%49%
mg51%41%47%46%
sparse_eign/an/an/an/a
mixed20%16%79%76%
integern/an/an/an/a

The median time of each of the 107 entries in each run. The fastest run of each entry is in bold; — is an entry the run did not measure.

areaoptypeshapenative 1Tnative 16Tnative-fast 1Tnative-fast 16T
blas1dotf64n=838860814.194 ms4.706 ms5.268 ms4.357 ms
blas1dotf32n=838860814.387 ms2.809 ms2.517 ms2.250 ms
blas1axpyf64n=838860816.203 ms7.158 ms8.908 ms6.836 ms
blas1axpyf32n=838860814.882 ms3.499 ms4.689 ms3.264 ms
blas1scalef64n=83886087.206 ms4.374 ms6.799 ms4.458 ms
blas1two_normf64n=838860814.603 ms2.687 ms2.823 ms2.295 ms
blas1one_normf64n=83886086.570 ms6.604 ms6.628 ms6.608 ms
blas1infinity_normf64n=83886088.846 ms8.787 ms22.107 ms14.559 ms
blas1sumf64n=83886086.636 ms6.611 ms6.638 ms6.648 ms
blas1dot_k3f64n=8388608116.474 ms115.573 ms119.442 ms117.447 ms
blas2gemvf64n=409614.252 ms4.541 ms5.478 ms4.405 ms
blas2gemvf32n=409613.097 ms2.621 ms2.811 ms2.266 ms
blas2symvf64n=40969.995 ms10.069 ms10.340 ms10.024 ms
blas2trmvf64n=40967.160 ms7.113 ms2.997 ms3.159 ms
blas2trsvf64n=40966.939 ms7.081 ms3.172 ms3.365 ms
blas2gerf64n=409624.014 ms23.955 ms24.293 ms24.890 ms
blas3gemmf64512x512x51228.436 ms3.622 ms21.348 ms4.760 ms
blas3gemmf32512x512x51213.307 ms2.715 ms9.929 ms3.837 ms
blas3gemmtf64512x512x51215.912 ms15.902 ms13.463 ms14.098 ms
blas3symmf64512x512x51228.809 ms28.242 ms22.346 ms5.832 ms
blas3syrkf64512x512x51218.498 ms18.815 ms19.065 ms18.837 ms
blas3syr2kf64512x512x51229.410 ms28.495 ms30.963 ms28.438 ms
blas3trmmf64512x512x51245.065 ms44.256 ms17.742 ms11.427 ms
blas3trsmf64512x512x51245.826 ms46.392 ms21.624 ms11.385 ms
blas3trsmf32512x512x51250.882 ms44.436 ms11.764 ms8.993 ms
blas3kronf6448x48 (x) 48x4819.056 ms17.021 ms16.838 ms16.964 ms
factorluf64512x512148.661 ms50.534 ms149.146 ms49.409 ms
factorluf32512x512150.947 ms58.603 ms51.625 ms82.944 ms
factorcholeskyf64512x51216.425 ms16.523 ms16.437 ms16.451 ms
factorcholeskyf32512x51216.857 ms16.316 ms16.436 ms16.409 ms
factorldltf64512x51216.489 ms15.902 ms16.202 ms15.829 ms
factorldlt_bkf64512x512159.439 ms159.553 ms159.488 ms160.108 ms
factorqrf64512x512512.832 ms310.657 ms511.967 ms312.080 ms
factorlqf64512x512154.419 ms72.861 ms172.004 ms109.561 ms
factorhessenbergf64512x512685.062 ms418.485 ms678.472 ms434.708 ms
factorhouseholderf64512x5123.286 ms0.778 ms3.289 ms0.673 ms
factorinvf64512x512171.541 ms159.077 ms206.977 ms159.895 ms
factorlu_iterative_refinef32>f64512x51264.490 ms63.831 ms153.007 ms53.593 ms
eigeneigenvalue_symmetricf64256x25633.554 ms33.550 ms39.822 ms43.571 ms
eigeneigen_symmetricf64256x256111.391 ms104.711 ms113.651 ms119.587 ms
eigeneigen_symmetric_jacobif64256x256586.004 ms583.927 ms610.571 ms584.781 ms
eigeneigenvaluef64256x25675.484 ms68.055 ms68.393 ms68.048 ms
eigenschurf64256x256173.144 ms233.611 ms220.307 ms185.943 ms
eigensvdf64256x256179.009 ms173.886 ms183.620 ms174.965 ms
eigenexpmf64256x25661.162 ms36.989 ms50.017 ms37.893 ms
eigensylvesterf64256x256434.842 ms445.949 ms393.788 ms390.383 ms
elementwiseexpf64n=838860857.329 ms54.608 ms61.654 ms58.003 ms
elementwiselogf64n=838860875.034 ms68.645 ms70.100 ms74.233 ms
elementwisesinf64n=838860856.559 ms58.953 ms59.464 ms58.646 ms
elementwisesqrtf64n=838860827.251 ms28.626 ms28.556 ms27.444 ms
elementwiseerff64n=838860860.953 ms64.030 ms61.858 ms62.164 ms
sparsespmvf64512^2 grid, nnz 13086721.359 ms0.969 ms1.270 ms0.967 ms
sparsespmv_transf64512^2 grid, nnz 13086723.144 ms3.153 ms3.167 ms3.138 ms
sparsecrs_densef64512^2 grid, nnz 1308672 x 167.615 ms4.561 ms7.504 ms4.519 ms
sparsespgemmf64256^2 grid, nnz 32665621.180 ms20.681 ms21.023 ms19.535 ms
sparselower_solve_levelf64192^2 grid, nnz(L) 10035642.961 ms3.881 ms2.982 ms3.252 ms
directsparse_choleskyf6496^2 grid, nnz 45696185.719 ms174.271 ms172.342 ms166.490 ms
directsparse_ldltf6496^2 grid, nnz 4569653.487 ms53.205 ms52.812 ms49.649 ms
directsupernodal_ldltf6496^2 grid, nnz 4569634.282 ms33.784 ms32.804 ms28.916 ms
directsparse_luf6496^2 grid, nnz 45696112.160 ms109.599 ms100.850 ms107.379 ms
directsupernodal_luf6496^2 grid, nnz 45696121.169 ms134.579 ms127.068 ms131.631 ms
directnative_kluf6496^2 grid, nnz 4569646.807 ms50.168 ms53.139 ms51.293 ms
directsparse_qrf6448^2 grid, nnz 11328353.341 ms420.798 ms402.178 ms388.801 ms
orderingamdf6496^2 grid, nnz 45696, scrambled5.922 ms6.478 ms6.007 ms6.058 ms
orderingcolamdf6496^2 grid, nnz 45696, columns scrambled9.345 ms10.788 ms9.421 ms9.506 ms
orderingrcmf6496^2 grid, nnz 45696, scrambled0.779 ms0.866 ms0.877 ms0.856 ms
orderingblock_triangular_formf6432 planted blocks, n 819227.218 ms32.297 ms26.042 ms26.683 ms
krylovcg+ic0f64128^2 grid, n 1638453.128 ms56.006 ms43.186 ms43.101 ms
krylovminres+diagf64128^2 grid, n 1638467.238 ms70.935 ms51.599 ms51.286 ms
krylovbicg+ilu0f64128^2 grid, n 1638444.273 ms46.187 ms38.306 ms36.871 ms
krylovbicgstab+ilu0f64128^2 grid, n 1638422.740 ms25.129 ms18.357 ms17.931 ms
krylovbicgstab_ell+ilu0f64128^2 grid, n 1638433.048 ms32.520 ms25.135 ms25.154 ms
krylovcgs+ilu0f64128^2 grid, n 1638423.994 ms25.480 ms21.382 ms21.456 ms
krylovgmres30+ilu0f64128^2 grid, n 1638478.125 ms87.694 ms53.390 ms47.519 ms
krylovqmr+ilu0f64128^2 grid, n 1638448.771 ms49.748 ms38.915 ms43.657 ms
krylovtfqmr+ilu0f64128^2 grid, n 1638427.843 ms31.736 ms24.594 ms26.641 ms
krylovidr_s1+ilu0f64128^2 grid, n 1638444.966 ms48.518 ms34.267 ms34.188 ms
krylovidr_s2+ilu0f64128^2 grid, n 1638455.468 ms53.815 ms36.995 ms37.655 ms
krylovidr_s4+ilu0f64128^2 grid, n 1638484.429 ms93.213 ms54.484 ms53.020 ms
krylovidr_s8+ilu0f64128^2 grid, n 1638472.799 ms79.002 ms45.719 ms47.587 ms
pcdiagonalf64128^2 grid, rows scaled0.061 ms0.076 ms0.074 ms0.061 ms
pcilu_0f64128^2 grid4.382 ms4.579 ms4.586 ms4.770 ms
pcilutf64128^2 grid46.639 ms49.207 ms43.890 ms43.149 ms
pcic_0f64128^2 grid3.400 ms3.460 ms4.183 ms3.839 ms
pcildlf64128^2 grid3.559 ms3.748 ms3.554 ms3.596 ms
pcssorf64128^2 grid0.049 ms0.029 ms0.039 ms0.029 ms
pcblock_diagonal4f64128^2 grid, rows scaled1.133 ms1.145 ms1.180 ms1.118 ms
smootherjacobif64512^2 grid, nnz 13086722.358 ms2.272 ms2.253 ms2.286 ms
smoothergauss_seidelf64512^2 grid, nnz 13086722.032 ms2.002 ms2.010 ms2.034 ms
smoothersor1.5f64512^2 grid, nnz 13086722.376 ms2.406 ms2.393 ms2.442 ms
mgpoisson2d_vcyclef64255^2 grid, 6 levels, 10 V(1,1)24.197 ms25.987 ms24.617 ms24.290 ms
sparse_eiglanczosf6464x65 Laplacian, k=4, m=8063.165 ms55.326 ms51.213 ms48.740 ms
sparse_eigarnoldif6464x65 convection-diffusion, k=4, m=8069.704 ms101.304 ms70.644 ms74.842 ms
sparse_eigshift_invertf6464x65 Laplacian, k=4 nearest 0, m=40126.893 ms152.520 ms119.792 ms127.283 ms
sparse_eigpower_iterationf6464x65 Laplacian, 300 steps11.852 ms14.543 ms10.019 ms9.789 ms
mixeddotf32>f64n=838860814.221 ms15.628 ms3.196 ms3.302 ms
mixedgemmf32>f64512x512x51253.319 ms7.016 ms25.600 ms4.258 ms
mixedblock_scaled_doti8+f32n=8388608, block 322.300 ms3.078 ms0.898 ms0.893 ms
integerdoti32n=8388608, |v| <= 154.358 ms4.302 ms2.816 ms3.023 ms
integerdoti16>i32n=8388608, |v| <= 153.988 ms3.894 ms1.203 ms1.229 ms
integerdoti8>i32n=8388608, |v| <= 152.635 ms2.174 ms0.857 ms0.885 ms
integerdotu8i8>i32n=8388608, |v| <= 152.236 ms2.165 ms0.677 ms0.757 ms
integergemmi32512^3, |v| <= 204718.349 ms2.704 ms27.443 ms4.165 ms
integergemmi16>i32512^3, |v| <= 204747.401 ms8.559 ms27.521 ms6.555 ms
integergemmi8>i32512^3, |v| <= 12738.913 ms8.393 ms29.075 ms4.432 ms
integergemm_quadi8>i32512^3, |v| <= 12739.613 ms7.651 ms7.638 ms2.109 ms
integergemm_quadu8i8>i32512^3, |v| <= 12738.405 ms7.734 ms6.695 ms2.294 ms

AMD Ryzen 7 2700X Eight-Core Processor — native, 1 thread, class medium

Section titled “AMD Ryzen 7 2700X Eight-Core Processor — native, 1 thread, class medium”
CPUAMD Ryzen 7 2700X Eight-Core Processor (x86_64, 16 logical cores)
OSWindows 10.0.26300 (26H2)
compilerMSVC 19.51.36256
build/DWIN32 /D_WINDOWS /EHsc /O2 /Ob2 /DNDEBUG, SIMD scalar
commite096dbbc6788
FMA peak (ceiling)f64 31.9 GFLOP/s, f32 63.6 GFLOP/s
… hardware, AVX2+FMAf64 31.9, f32 63.6 GFLOP/s
… this build, scalarf64 1.2, f32 1.2 GFLOP/s
bandwidthread 21.1, update 23.4, triad 19.4 GB/s (ceiling 23.4)
last-level cache8 MiB
areaentriesmedian %roofbest %rooffailed checks
blas11040%80%0
blas2640%48%0
blas31026%32%0
factor124%9%0
eigen87%13%0
elementwise510%21%0
sparse537%86%0
direct71%1%0
ordering4n/an/a0
krylov13n/an/a0
pc7n/an/a0
smoother353%62%0
mg151%51%0
sparse_eig4n/an/a0
mixed320%33%0
integer9n/an/a0
Every entry of this run
areaoptypeshapemedianGFLOP/sGB/s%roofcheck
blas1dotf64n=838860814.194 ms1.189.4640%ok
blas1dotf32n=838860814.387 ms1.174.6620%ok
blas1axpyf64n=838860816.203 ms1.0412.4353%ok
blas1axpyf32n=838860814.882 ms1.136.7629%ok
blas1scalef64n=83886087.206 ms1.1618.6380%ok
blas1two_normf64n=838860814.603 ms1.154.6020%ok
blas1one_normf64n=83886086.570 ms1.2810.2144%ok
blas1infinity_normf64n=83886088.846 ms0.957.5932%ok
blas1sumf64n=83886086.636 ms1.2610.1143%ok
blas1dot_k3f64n=8388608116.474 ms1.581.155%ok
blas2gemvf64n=409614.252 ms2.359.4240%ok
blas2gemvf32n=409613.097 ms2.565.1322%ok
blas2symvf64n=40969.995 ms3.366.7229%ok
blas2trmvf64n=40967.160 ms2.349.3840%ok
blas2trsvf64n=40966.939 ms2.429.6841%ok
blas2gerf64n=409624.014 ms1.4011.1848%ok
blas3gemmf64512x512x51228.436 ms9.440.2230%ok
blas3gemmf32512x512x51213.307 ms20.170.2432%ok
blas3gemmtf64512x512x51215.912 ms8.430.4026%ok
blas3symmf64512x512x51228.809 ms9.320.2229%ok
blas3syrkf64512x512x51218.498 ms7.260.2323%ok
blas3syr2kf64512x512x51229.410 ms9.130.2129%ok
blas3trmmf64512x512x51245.065 ms2.980.129%ok
blas3trsmf64512x512x51245.826 ms2.930.119%ok
blas3trsmf32512x512x51250.882 ms2.640.054%ok
blas3kronf6448x48 (x) 48x4819.056 ms0.282.2310%ok
factorluf64512x512148.661 ms0.600.032%ok
factorluf32512x512150.947 ms0.590.011%ok
factorcholeskyf64512x51216.425 ms2.720.269%ok
factorcholeskyf32512x51216.857 ms2.650.124%ok
factorldltf64512x51216.489 ms2.710.259%ok
factorldlt_bkf64512x512159.439 ms0.280.031%ok
factorqrf64512x512512.832 ms0.350.011%ok
factorlqf64512x512154.419 ms1.160.034%ok
factorhessenbergf64512x512685.062 ms0.650.012%ok
factorhouseholderf64512x5123.286 ms0.321.285%*ok
factorinvf64512x512171.541 ms1.560.025%ok
factorlu_iterative_refinef32>f64512x51264.490 ms1.390.074%ok
eigeneigenvalue_symmetricf64256x25633.554 ms0.670.032%ok
eigeneigen_symmetricf64256x256111.391 ms1.360.014%ok
eigeneigen_symmetric_jacobif64256x256586.004 ms0.260.001%ok
eigeneigenvaluef64256x25675.484 ms2.220.017%ok
eigenschurf64256x256173.144 ms2.420.018%ok
eigensvdf64256x256179.009 ms2.060.016%ok
eigenexpmf64256x25661.162 ms4.110.0213%ok
eigensylvesterf64256x256434.842 ms2.310.007%ok
elementwiseexpf64n=838860857.329 ms0.152.3410%ok
elementwiselogf64n=838860875.034 ms0.111.798%ok
elementwisesinf64n=838860856.559 ms0.152.3710%ok
elementwisesqrtf64n=838860827.251 ms0.314.9321%ok
elementwiseerff64n=838860860.953 ms0.142.209%ok
sparsespmvf64512^2 grid, nnz 13086721.359 ms1.9320.0486%ok
sparsespmv_transf64512^2 grid, nnz 13086723.144 ms0.838.6637%ok
sparsecrs_densef64512^2 grid, nnz 1308672 x 167.615 ms5.5011.5649%ok
sparsespgemmf64256^2 grid, nnz 32665621.180 ms0.150.542%ok
sparselower_solve_levelf64192^2 grid, nnz(L) 10035642.961 ms0.685.6224%ok
directsparse_choleskyf6496^2 grid, nnz 45696185.719 ms0.060.000%ok
directsparse_ldltf6496^2 grid, nnz 4569653.487 ms0.210.011%ok
directsupernodal_ldltf6496^2 grid, nnz 4569634.282 ms0.330.021%ok
directsparse_luf6496^2 grid, nnz 45696112.160 ms0.200.011%ok
directsupernodal_luf6496^2 grid, nnz 45696121.169 ms0.180.011%ok
directnative_kluf6496^2 grid, nnz 4569646.807 ms0.480.021%ok
directsparse_qrf6448^2 grid, nnz 11328353.341 ms0.010.000%ok
orderingamdf6496^2 grid, nnz 45696, scrambled5.922 ms0.000.00n/aok
orderingcolamdf6496^2 grid, nnz 45696, columns scrambled9.345 ms0.000.00n/aok
orderingrcmf6496^2 grid, nnz 45696, scrambled0.779 ms0.000.00n/aok
orderingblock_triangular_formf6432 planted blocks, n 819227.218 ms0.000.00n/aok
krylovcg+ic0f64128^2 grid, n 1638453.128 ms0.000.00n/aok
krylovminres+diagf64128^2 grid, n 1638467.238 ms0.000.00n/aok
krylovbicg+ilu0f64128^2 grid, n 1638444.273 ms0.000.00n/aok
krylovbicgstab+ilu0f64128^2 grid, n 1638422.740 ms0.000.00n/aok
krylovbicgstab_ell+ilu0f64128^2 grid, n 1638433.048 ms0.000.00n/aok
krylovcgs+ilu0f64128^2 grid, n 1638423.994 ms0.000.00n/aok
krylovgmres30+ilu0f64128^2 grid, n 1638478.125 ms0.000.00n/aok
krylovqmr+ilu0f64128^2 grid, n 1638448.771 ms0.000.00n/aok
krylovtfqmr+ilu0f64128^2 grid, n 1638427.843 ms0.000.00n/aok
krylovidr_s1+ilu0f64128^2 grid, n 1638444.966 ms0.000.00n/aok
krylovidr_s2+ilu0f64128^2 grid, n 1638455.468 ms0.000.00n/aok
krylovidr_s4+ilu0f64128^2 grid, n 1638484.429 ms0.000.00n/aok
krylovidr_s8+ilu0f64128^2 grid, n 1638472.799 ms0.000.00n/aok
pcdiagonalf64128^2 grid, rows scaled0.061 ms0.000.00n/aok
pcilu_0f64128^2 grid4.382 ms0.000.00n/aok
pcilutf64128^2 grid46.639 ms0.000.00n/aok
pcic_0f64128^2 grid3.400 ms0.000.00n/aok
pcildlf64128^2 grid3.559 ms0.000.00n/aok
pcssorf64128^2 grid0.049 ms0.000.00n/aok
pcblock_diagonal4f64128^2 grid, rows scaled1.133 ms0.000.00n/aok
smootherjacobif64512^2 grid, nnz 13086722.358 ms1.1112.4453%ok
smoothergauss_seidelf64512^2 grid, nnz 13086722.032 ms1.2914.4362%ok
smoothersor1.5f64512^2 grid, nnz 13086722.376 ms1.1012.3553%ok
mgpoisson2d_vcyclef64255^2 grid, 6 levels, 10 V(1,1)24.197 ms1.5912.0151%ok
sparse_eiglanczosf6464x65 Laplacian, k=4, m=8063.165 ms0.000.00n/aok
sparse_eigarnoldif6464x65 convection-diffusion, k=4, m=8069.704 ms0.000.00n/aok
sparse_eigshift_invertf6464x65 Laplacian, k=4 nearest 0, m=40126.893 ms0.000.00n/aok
sparse_eigpower_iterationf6464x65 Laplacian, 300 steps11.852 ms0.000.00n/aok
mixeddotf32>f64n=838860814.221 ms1.184.7220%ok
mixedgemmf32>f64512x512x51253.319 ms5.030.0816%ok
mixedblock_scaled_doti8+f32n=8388608, block 322.300 ms7.297.7533%ok
integerdoti32n=8388608, |v| <= 154.358 ms3.8515.40n/aok
integerdoti16>i32n=8388608, |v| <= 153.988 ms4.218.41n/aok
integerdoti8>i32n=8388608, |v| <= 152.635 ms6.376.37n/aok
integerdotu8i8>i32n=8388608, |v| <= 152.236 ms7.507.50n/aok
integergemmi32512^3, |v| <= 204718.349 ms14.630.17n/aok
integergemmi16>i32512^3, |v| <= 204747.401 ms5.660.04n/aok
integergemmi8>i32512^3, |v| <= 12738.913 ms6.900.04n/aok
integergemm_quadi8>i32512^3, |v| <= 12739.613 ms6.780.04n/aok
integergemm_quadu8i8>i32512^3, |v| <= 12738.405 ms6.990.04n/aok

* working set fits in the last-level cache: a cache rate, shown as a % of the DRAM roofline.

AMD Ryzen 7 2700X Eight-Core Processor — native, 16 threads, class medium

Section titled “AMD Ryzen 7 2700X Eight-Core Processor — native, 16 threads, class medium”
CPUAMD Ryzen 7 2700X Eight-Core Processor (x86_64, 16 logical cores)
OSWindows 10.0.26300 (26H2)
compilerMSVC 19.51.36256
build/DWIN32 /D_WINDOWS /EHsc /O2 /Ob2 /DNDEBUG, SIMD scalar
commite096dbbc6788
FMA peak (ceiling)f64 248.9 GFLOP/s, f32 498.5 GFLOP/s
… hardware, AVX2+FMAf64 248.9, f32 498.5 GFLOP/s
… this build, scalarf64 14.9, f32 17.0 GFLOP/s
bandwidthread 26.3, update 27.4, triad 21.2 GB/s (ceiling 27.4)
last-level cache8 MiB
areaentriesmedian %roofbest %rooffailed checksmedian parallel efficiency
blas11091%112%014%
blas2641%108%06%
blas3104%30%06%
factor121%20%010%
eigen81%3%06%
elementwise58%17%06%
sparse532%103%06%
direct70%0%06%
ordering4n/an/a06%
krylov13n/an/a06%
pc7n/an/a06%
smoother347%54%06%
mg141%41%06%
sparse_eig4n/an/a05%
mixed316%21%06%
integer9n/an/a029%
Every entry of this run
areaoptypeshapemedianGFLOP/sGB/s%roofcheckspeedup (efficiency)
blas1dotf64n=83886084.706 ms3.5728.52104%ok3.0x (19%)
blas1dotf32n=83886082.809 ms5.9723.8987%ok5.1x (32%)
blas1axpyf64n=83886087.158 ms2.3428.13103%ok2.3x (14%)
blas1axpyf32n=83886083.499 ms4.7928.77105%ok4.3x (27%)
blas1scalef64n=83886084.374 ms1.9230.69112%ok1.6x (10%)
blas1two_normf64n=83886082.687 ms6.2524.9891%ok5.4x (34%)
blas1one_normf64n=83886086.604 ms1.2710.1637%ok1.0x (6%)
blas1infinity_normf64n=83886088.787 ms0.957.6428%ok1.0x (6%)
blas1sumf64n=83886086.611 ms1.2710.1537%ok1.0x (6%)
blas1dot_k3f64n=8388608115.573 ms1.601.164%ok1.0x (6%)
blas2gemvf64n=40964.541 ms7.3929.57108%ok3.1x (20%)
blas2gemvf32n=40962.621 ms12.8025.6194%ok5.0x (31%)
blas2symvf64n=409610.069 ms3.336.6724%ok1.0x (6%)
blas2trmvf64n=40967.113 ms2.369.4435%ok1.0x (6%)
blas2trsvf64n=40967.081 ms2.379.4935%ok1.0x (6%)
blas2gerf64n=409623.955 ms1.4011.2141%ok1.0x (6%)
blas3gemmf64512x512x5123.622 ms74.111.7430%ok7.9x (49%)
blas3gemmf32512x512x5122.715 ms98.881.1620%ok4.9x (31%)
blas3gemmtf64512x512x51215.902 ms8.440.403%ok1.0x (6%)
blas3symmf64512x512x51228.242 ms9.500.224%ok1.0x (6%)
blas3syrkf64512x512x51218.815 ms7.130.223%ok1.0x (6%)
blas3syr2kf64512x512x51228.495 ms9.420.224%ok1.0x (6%)
blas3trmmf64512x512x51244.256 ms3.030.121%ok1.0x (6%)
blas3trsmf64512x512x51246.392 ms2.890.111%ok1.0x (6%)
blas3trsmf32512x512x51244.436 ms3.020.061%ok1.1x (7%)
blas3kronf6448x48 (x) 48x4817.021 ms0.312.509%ok1.1x (7%)
factorluf64512x51250.534 ms1.770.081%ok2.9x (18%)
factorluf32512x51258.603 ms1.530.040%ok2.6x (16%)
factorcholeskyf64512x51216.523 ms2.710.251%ok1.0x (6%)
factorcholeskyf32512x51216.316 ms2.740.131%ok1.0x (6%)
factorldltf64512x51215.902 ms2.810.261%ok1.0x (6%)
factorldlt_bkf64512x512159.553 ms0.280.030%ok1.0x (6%)
factorqrf64512x512310.657 ms0.580.010%ok1.7x (10%)
factorlqf64512x51272.861 ms2.460.061%ok2.1x (13%)
factorhessenbergf64512x512418.485 ms1.070.010%ok1.6x (10%)
factorhouseholderf64512x5120.778 ms1.355.3920%*ok4.2x (26%)
factorinvf64512x512159.077 ms1.690.031%ok1.1x (7%)
factorlu_iterative_refinef32>f64512x51263.831 ms1.400.071%ok1.0x (6%)
eigeneigenvalue_symmetricf64256x25633.550 ms0.670.030%ok1.0x (6%)
eigeneigen_symmetricf64256x256104.711 ms1.440.011%ok1.1x (7%)
eigeneigen_symmetric_jacobif64256x256583.927 ms0.260.000%ok1.0x (6%)
eigeneigenvaluef64256x25668.055 ms2.470.021%ok1.1x (7%)
eigenschurf64256x256233.611 ms1.800.001%ok0.7x (5%)
eigensvdf64256x256173.886 ms2.120.011%ok1.0x (6%)
eigenexpmf64256x25636.989 ms6.800.033%ok1.7x (10%)
eigensylvesterf64256x256445.949 ms2.260.001%ok1.0x (6%)
elementwiseexpf64n=838860854.608 ms0.152.469%ok1.0x (7%)
elementwiselogf64n=838860868.645 ms0.121.967%ok1.1x (7%)
elementwisesinf64n=838860858.953 ms0.142.288%ok1.0x (6%)
elementwisesqrtf64n=838860828.626 ms0.294.6917%ok1.0x (6%)
elementwiseerff64n=838860864.030 ms0.132.108%ok1.0x (6%)
sparsespmvf64512^2 grid, nnz 13086720.969 ms2.7028.10103%ok1.4x (9%)
sparsespmv_transf64512^2 grid, nnz 13086723.153 ms0.838.6432%ok1.0x (6%)
sparsecrs_densef64512^2 grid, nnz 1308672 x 164.561 ms9.1819.3071%ok1.7x (10%)
sparsespgemmf64256^2 grid, nnz 32665620.681 ms0.160.562%ok1.0x (6%)
sparselower_solve_levelf64192^2 grid, nnz(L) 10035643.881 ms0.524.2916%ok0.8x (5%)
directsparse_choleskyf6496^2 grid, nnz 45696174.271 ms0.060.000%ok1.1x (7%)
directsparse_ldltf6496^2 grid, nnz 4569653.205 ms0.210.010%ok1.0x (6%)
directsupernodal_ldltf6496^2 grid, nnz 4569633.784 ms0.330.020%ok1.0x (6%)
directsparse_luf6496^2 grid, nnz 45696109.599 ms0.200.010%ok1.0x (6%)
directsupernodal_luf6496^2 grid, nnz 45696134.579 ms0.170.010%ok0.9x (6%)
directnative_kluf6496^2 grid, nnz 4569650.168 ms0.450.010%ok0.9x (6%)
directsparse_qrf6448^2 grid, nnz 11328420.798 ms0.000.000%ok0.8x (5%)
orderingamdf6496^2 grid, nnz 45696, scrambled6.478 ms0.000.00n/aok0.9x (6%)
orderingcolamdf6496^2 grid, nnz 45696, columns scrambled10.788 ms0.000.00n/aok0.9x (5%)
orderingrcmf6496^2 grid, nnz 45696, scrambled0.866 ms0.000.00n/aok0.9x (6%)
orderingblock_triangular_formf6432 planted blocks, n 819232.297 ms0.000.00n/aok0.8x (5%)
krylovcg+ic0f64128^2 grid, n 1638456.006 ms0.000.00n/aok0.9x (6%)
krylovminres+diagf64128^2 grid, n 1638470.935 ms0.000.00n/aok0.9x (6%)
krylovbicg+ilu0f64128^2 grid, n 1638446.187 ms0.000.00n/aok1.0x (6%)
krylovbicgstab+ilu0f64128^2 grid, n 1638425.129 ms0.000.00n/aok0.9x (6%)
krylovbicgstab_ell+ilu0f64128^2 grid, n 1638432.520 ms0.000.00n/aok1.0x (6%)
krylovcgs+ilu0f64128^2 grid, n 1638425.480 ms0.000.00n/aok0.9x (6%)
krylovgmres30+ilu0f64128^2 grid, n 1638487.694 ms0.000.00n/aok0.9x (6%)
krylovqmr+ilu0f64128^2 grid, n 1638449.748 ms0.000.00n/aok1.0x (6%)
krylovtfqmr+ilu0f64128^2 grid, n 1638431.736 ms0.000.00n/aok0.9x (5%)
krylovidr_s1+ilu0f64128^2 grid, n 1638448.518 ms0.000.00n/aok0.9x (6%)
krylovidr_s2+ilu0f64128^2 grid, n 1638453.815 ms0.000.00n/aok1.0x (6%)
krylovidr_s4+ilu0f64128^2 grid, n 1638493.213 ms0.000.00n/aok0.9x (6%)
krylovidr_s8+ilu0f64128^2 grid, n 1638479.002 ms0.000.00n/aok0.9x (6%)
pcdiagonalf64128^2 grid, rows scaled0.076 ms0.000.00n/aok0.8x (5%)
pcilu_0f64128^2 grid4.579 ms0.000.00n/aok1.0x (6%)
pcilutf64128^2 grid49.207 ms0.000.00n/aok0.9x (6%)
pcic_0f64128^2 grid3.460 ms0.000.00n/aok1.0x (6%)
pcildlf64128^2 grid3.748 ms0.000.00n/aok0.9x (6%)
pcssorf64128^2 grid0.029 ms0.000.00n/aok1.7x (11%)
pcblock_diagonal4f64128^2 grid, rows scaled1.145 ms0.000.00n/aok1.0x (6%)
smootherjacobif64512^2 grid, nnz 13086722.272 ms1.1512.9147%ok1.0x (6%)
smoothergauss_seidelf64512^2 grid, nnz 13086722.002 ms1.3114.6554%ok1.0x (6%)
smoothersor1.5f64512^2 grid, nnz 13086722.406 ms1.0912.1945%ok1.0x (6%)
mgpoisson2d_vcyclef64255^2 grid, 6 levels, 10 V(1,1)25.987 ms1.4811.1841%ok0.9x (6%)
sparse_eiglanczosf6464x65 Laplacian, k=4, m=8055.326 ms0.000.00n/aok1.1x (7%)
sparse_eigarnoldif6464x65 convection-diffusion, k=4, m=80101.304 ms0.000.00n/aok0.7x (4%)
sparse_eigshift_invertf6464x65 Laplacian, k=4 nearest 0, m=40152.520 ms0.000.00n/aok0.8x (5%)
sparse_eigpower_iterationf6464x65 Laplacian, 300 steps14.543 ms0.000.00n/aok0.8x (5%)
mixeddotf32>f64n=838860815.628 ms1.074.2916%ok0.9x (6%)
mixedgemmf32>f64512x512x5127.016 ms38.260.6015%ok7.6x (47%)
mixedblock_scaled_doti8+f32n=8388608, block 323.078 ms5.455.7921%ok0.7x (5%)
integerdoti32n=8388608, |v| <= 154.302 ms3.9015.60n/aok1.0x (6%)
integerdoti16>i32n=8388608, |v| <= 153.894 ms4.318.62n/aok1.0x (6%)
integerdoti8>i32n=8388608, |v| <= 152.174 ms7.727.72n/aok1.2x (8%)
integerdotu8i8>i32n=8388608, |v| <= 152.165 ms7.757.75n/aok1.0x (6%)
integergemmi32512^3, |v| <= 20472.704 ms99.261.16n/aok6.8x (42%)
integergemmi16>i32512^3, |v| <= 20478.559 ms31.360.25n/aok5.5x (35%)
integergemmi8>i32512^3, |v| <= 1278.393 ms31.980.19n/aok4.6x (29%)
integergemm_quadi8>i32512^3, |v| <= 1277.651 ms35.080.21n/aok5.2x (32%)
integergemm_quadu8i8>i32512^3, |v| <= 1277.734 ms34.710.20n/aok5.0x (31%)

* working set fits in the last-level cache: a cache rate, shown as a % of the DRAM roofline.

AMD Ryzen 7 2700X Eight-Core Processor — native-fast, 1 thread, class medium

Section titled “AMD Ryzen 7 2700X Eight-Core Processor — native-fast, 1 thread, class medium”
CPUAMD Ryzen 7 2700X Eight-Core Processor (x86_64, 16 logical cores)
OSWindows 10.0.26300 (26H2)
compilerMSVC 19.51.36256
build/DWIN32 /D_WINDOWS /EHsc /O2 /Ob2 /DNDEBUG, SIMD SSE2
commite096dbbc6788
FMA peak (ceiling)f64 31.9 GFLOP/s, f32 64.2 GFLOP/s
… hardware, AVX2+FMAf64 31.9, f32 64.2 GFLOP/s
… this build, SSE2f64 31.9, f32 63.9 GFLOP/s
bandwidthread 25.1, update 21.0, triad 18.3 GB/s (ceiling 25.1)
last-level cache8 MiB
areaentriesmedian %roofbest %rooffailed checks
blas11085%106%0
blas2689%98%0
blas31027%42%0
factor123%9%0
eigen86%16%0
elementwise59%19%0
sparse534%85%0
direct71%1%0
ordering4n/an/a0
krylov13n/an/a0
pc7n/an/a0
smoother352%58%0
mg147%47%0
sparse_eig4n/an/a0
mixed379%84%0
integer9n/an/a0
Every entry of this run
areaoptypeshapemedianGFLOP/sGB/s%roofcheck
blas1dotf64n=83886085.268 ms3.1825.48101%ok
blas1dotf32n=83886082.517 ms6.6726.66106%ok
blas1axpyf64n=83886088.908 ms1.8822.6090%ok
blas1axpyf32n=83886084.689 ms3.5821.4785%ok
blas1scalef64n=83886086.799 ms1.2319.7479%ok
blas1two_normf64n=83886082.823 ms5.9423.7795%ok
blas1one_normf64n=83886086.628 ms1.2710.1240%ok
blas1infinity_normf64n=838860822.107 ms0.383.0412%ok
blas1sumf64n=83886086.638 ms1.2610.1140%ok
blas1dot_k3f64n=8388608119.442 ms1.551.125%ok
blas2gemvf64n=40965.478 ms6.1224.5198%ok
blas2gemvf32n=40962.811 ms11.9423.8995%ok
blas2symvf64n=409610.340 ms3.256.5026%ok
blas2trmvf64n=40962.997 ms5.6022.4189%ok
blas2trsvf64n=40963.172 ms5.2921.1884%ok
blas2gerf64n=409624.293 ms1.3811.0544%ok
blas3gemmf64512x512x51221.348 ms12.570.2939%ok
blas3gemmf32512x512x5129.929 ms27.040.3242%ok
blas3gemmtf64512x512x51213.463 ms9.970.4731%ok
blas3symmf64512x512x51222.346 ms12.010.2838%ok
blas3syrkf64512x512x51219.065 ms7.040.2222%ok
blas3syr2kf64512x512x51230.963 ms8.670.2027%ok
blas3trmmf64512x512x51217.742 ms7.560.3024%ok
blas3trsmf64512x512x51221.624 ms6.210.2419%ok
blas3trsmf32512x512x51211.764 ms11.410.2218%ok
blas3kronf6448x48 (x) 48x4816.838 ms0.322.5210%ok
factorluf64512x512149.146 ms0.600.032%ok
factorluf32512x51251.625 ms1.730.043%ok
factorcholeskyf64512x51216.437 ms2.720.269%ok
factorcholeskyf32512x51216.436 ms2.720.134%ok
factorldltf64512x51216.202 ms2.760.269%ok
factorldlt_bkf64512x512159.488 ms0.280.031%ok
factorqrf64512x512511.967 ms0.350.011%ok
factorlqf64512x512172.004 ms1.040.023%ok
factorhessenbergf64512x512678.472 ms0.660.012%ok
factorhouseholderf64512x5123.289 ms0.321.285%*ok
factorinvf64512x512206.977 ms1.300.024%ok
factorlu_iterative_refinef32>f64512x512153.007 ms0.580.032%ok
eigeneigenvalue_symmetricf64256x25639.822 ms0.560.032%ok
eigeneigen_symmetricf64256x256113.651 ms1.330.014%ok
eigeneigen_symmetric_jacobif64256x256610.571 ms0.250.001%ok
eigeneigenvaluef64256x25668.393 ms2.450.028%ok
eigenschurf64256x256220.307 ms1.900.006%ok
eigensvdf64256x256183.620 ms2.010.016%ok
eigenexpmf64256x25650.017 ms5.030.0216%ok
eigensylvesterf64256x256393.788 ms2.560.008%ok
elementwiseexpf64n=838860861.654 ms0.142.189%ok
elementwiselogf64n=838860870.100 ms0.121.918%ok
elementwisesinf64n=838860859.464 ms0.142.269%ok
elementwisesqrtf64n=838860828.556 ms0.294.7019%ok
elementwiseerff64n=838860861.858 ms0.142.179%ok
sparsespmvf64512^2 grid, nnz 13086721.270 ms2.0621.4485%ok
sparsespmv_transf64512^2 grid, nnz 13086723.167 ms0.838.6034%ok
sparsecrs_densef64512^2 grid, nnz 1308672 x 167.504 ms5.5811.7347%ok
sparsespgemmf64256^2 grid, nnz 32665621.023 ms0.150.552%ok
sparselower_solve_levelf64192^2 grid, nnz(L) 10035642.982 ms0.675.5822%ok
directsparse_choleskyf6496^2 grid, nnz 45696172.342 ms0.060.000%ok
directsparse_ldltf6496^2 grid, nnz 4569652.812 ms0.210.011%ok
directsupernodal_ldltf6496^2 grid, nnz 4569632.804 ms0.340.021%ok
directsparse_luf6496^2 grid, nnz 45696100.850 ms0.220.011%ok
directsupernodal_luf6496^2 grid, nnz 45696127.068 ms0.180.011%ok
directnative_kluf6496^2 grid, nnz 4569653.139 ms0.420.011%ok
directsparse_qrf6448^2 grid, nnz 11328402.178 ms0.000.000%ok
orderingamdf6496^2 grid, nnz 45696, scrambled6.007 ms0.000.00n/aok
orderingcolamdf6496^2 grid, nnz 45696, columns scrambled9.421 ms0.000.00n/aok
orderingrcmf6496^2 grid, nnz 45696, scrambled0.877 ms0.000.00n/aok
orderingblock_triangular_formf6432 planted blocks, n 819226.042 ms0.000.00n/aok
krylovcg+ic0f64128^2 grid, n 1638443.186 ms0.000.00n/aok
krylovminres+diagf64128^2 grid, n 1638451.599 ms0.000.00n/aok
krylovbicg+ilu0f64128^2 grid, n 1638438.306 ms0.000.00n/aok
krylovbicgstab+ilu0f64128^2 grid, n 1638418.357 ms0.000.00n/aok
krylovbicgstab_ell+ilu0f64128^2 grid, n 1638425.135 ms0.000.00n/aok
krylovcgs+ilu0f64128^2 grid, n 1638421.382 ms0.000.00n/aok
krylovgmres30+ilu0f64128^2 grid, n 1638453.390 ms0.000.00n/aok
krylovqmr+ilu0f64128^2 grid, n 1638438.915 ms0.000.00n/aok
krylovtfqmr+ilu0f64128^2 grid, n 1638424.594 ms0.000.00n/aok
krylovidr_s1+ilu0f64128^2 grid, n 1638434.267 ms0.000.00n/aok
krylovidr_s2+ilu0f64128^2 grid, n 1638436.995 ms0.000.00n/aok
krylovidr_s4+ilu0f64128^2 grid, n 1638454.484 ms0.000.00n/aok
krylovidr_s8+ilu0f64128^2 grid, n 1638445.719 ms0.000.00n/aok
pcdiagonalf64128^2 grid, rows scaled0.074 ms0.000.00n/aok
pcilu_0f64128^2 grid4.586 ms0.000.00n/aok
pcilutf64128^2 grid43.890 ms0.000.00n/aok
pcic_0f64128^2 grid4.183 ms0.000.00n/aok
pcildlf64128^2 grid3.554 ms0.000.00n/aok
pcssorf64128^2 grid0.039 ms0.000.00n/aok
pcblock_diagonal4f64128^2 grid, rows scaled1.180 ms0.000.00n/aok
smootherjacobif64512^2 grid, nnz 13086722.253 ms1.1613.0252%ok
smoothergauss_seidelf64512^2 grid, nnz 13086722.010 ms1.3014.5958%ok
smoothersor1.5f64512^2 grid, nnz 13086722.393 ms1.0912.2649%ok
mgpoisson2d_vcyclef64255^2 grid, 6 levels, 10 V(1,1)24.617 ms1.5611.8147%ok
sparse_eiglanczosf6464x65 Laplacian, k=4, m=8051.213 ms0.000.00n/aok
sparse_eigarnoldif6464x65 convection-diffusion, k=4, m=8070.644 ms0.000.00n/aok
sparse_eigshift_invertf6464x65 Laplacian, k=4 nearest 0, m=40119.792 ms0.000.00n/aok
sparse_eigpower_iterationf6464x65 Laplacian, 300 steps10.019 ms0.000.00n/aok
mixeddotf32>f64n=83886083.196 ms5.2521.0084%ok
mixedgemmf32>f64512x512x51225.600 ms10.490.1633%ok
mixedblock_scaled_doti8+f32n=8388608, block 320.898 ms18.6719.8479%ok
integerdoti32n=8388608, |v| <= 152.816 ms5.9623.83n/aok
integerdoti16>i32n=8388608, |v| <= 151.203 ms13.9427.89n/aok
integerdoti8>i32n=8388608, |v| <= 150.857 ms19.5719.57n/aok
integerdotu8i8>i32n=8388608, |v| <= 150.677 ms24.7624.76n/aok
integergemmi32512^3, |v| <= 204727.443 ms9.780.11n/aok
integergemmi16>i32512^3, |v| <= 204727.521 ms9.750.08n/aok
integergemmi8>i32512^3, |v| <= 12729.075 ms9.230.05n/aok
integergemm_quadi8>i32512^3, |v| <= 1277.638 ms35.140.21n/aok
integergemm_quadu8i8>i32512^3, |v| <= 1276.695 ms40.090.23n/aok

* working set fits in the last-level cache: a cache rate, shown as a % of the DRAM roofline.

AMD Ryzen 7 2700X Eight-Core Processor — native-fast, 16 threads, class medium

Section titled “AMD Ryzen 7 2700X Eight-Core Processor — native-fast, 16 threads, class medium”
CPUAMD Ryzen 7 2700X Eight-Core Processor (x86_64, 16 logical cores)
OSWindows 10.0.26300 (26H2)
compilerMSVC 19.51.36256
build/DWIN32 /D_WINDOWS /EHsc /O2 /Ob2 /DNDEBUG, SIMD SSE2
commite096dbbc6788
FMA peak (ceiling)f64 246.6 GFLOP/s, f32 494.3 GFLOP/s
… hardware, AVX2+FMAf64 243.3, f32 488.7 GFLOP/s
… this build, SSE2f64 246.6, f32 494.3 GFLOP/s
bandwidthread 25.4, update 26.2, triad 20.9 GB/s (ceiling 26.2)
last-level cache8 MiB
areaentriesmedian %roofbest %rooffailed checksmedian parallel efficiency
blas110112%118%08%
blas2681%116%06%
blas3105%23%010%
factor121%24%010%
eigen81%3%07%
elementwise59%19%06%
sparse533%107%07%
direct70%0%06%
ordering4n/an/a06%
krylov13n/an/a06%
pc7n/an/a07%
smoother349%55%06%
mg146%46%06%
sparse_eig4n/an/a06%
mixed376%77%06%
integer9n/an/a018%
Every entry of this run
areaoptypeshapemedianGFLOP/sGB/s%roofcheckspeedup (efficiency)
blas1dotf64n=83886084.357 ms3.8530.81117%ok1.2x (8%)
blas1dotf32n=83886082.250 ms7.4629.83114%ok1.1x (7%)
blas1axpyf64n=83886086.836 ms2.4529.45112%ok1.3x (8%)
blas1axpyf32n=83886083.264 ms5.1430.84118%ok1.4x (9%)
blas1scalef64n=83886084.458 ms1.8830.11115%ok1.5x (10%)
blas1two_normf64n=83886082.295 ms7.3129.24111%ok1.2x (8%)
blas1one_normf64n=83886086.608 ms1.2710.1639%ok1.0x (6%)
blas1infinity_normf64n=838860814.559 ms0.584.6118%ok1.5x (9%)
blas1sumf64n=83886086.648 ms1.2610.0938%ok1.0x (6%)
blas1dot_k3f64n=8388608117.447 ms1.571.144%ok1.0x (6%)
blas2gemvf64n=40964.405 ms7.6230.48116%ok1.2x (8%)
blas2gemvf32n=40962.266 ms14.8129.63113%ok1.2x (8%)
blas2symvf64n=409610.024 ms3.356.7026%ok1.0x (6%)
blas2trmvf64n=40963.159 ms5.3121.2681%ok0.9x (6%)
blas2trsvf64n=40963.365 ms4.9919.9676%ok0.9x (6%)
blas2gerf64n=409624.890 ms1.3510.7941%ok1.0x (6%)
blas3gemmf64512x512x5124.760 ms56.401.3223%ok4.5x (28%)
blas3gemmf32512x512x5123.837 ms69.970.8214%ok2.6x (16%)
blas3gemmtf64512x512x51214.098 ms9.520.454%ok1.0x (6%)
blas3symmf64512x512x5125.832 ms46.021.0819%ok3.8x (24%)
blas3syrkf64512x512x51218.837 ms7.130.223%ok1.0x (6%)
blas3syr2kf64512x512x51228.438 ms9.440.224%ok1.1x (7%)
blas3trmmf64512x512x51211.427 ms11.750.465%ok1.6x (10%)
blas3trsmf64512x512x51211.385 ms11.790.465%ok1.9x (12%)
blas3trsmf32512x512x5128.993 ms14.930.293%ok1.3x (8%)
blas3kronf6448x48 (x) 48x4816.964 ms0.312.5010%ok1.0x (6%)
factorluf64512x51249.409 ms1.810.081%ok3.0x (19%)
factorluf32512x51282.944 ms1.080.030%ok0.6x (4%)
factorcholeskyf64512x51216.451 ms2.720.251%ok1.0x (6%)
factorcholeskyf32512x51216.409 ms2.730.131%ok1.0x (6%)
factorldltf64512x51215.829 ms2.830.261%ok1.0x (6%)
factorldlt_bkf64512x512160.108 ms0.280.030%ok1.0x (6%)
factorqrf64512x512312.080 ms0.570.010%ok1.6x (10%)
factorlqf64512x512109.561 ms1.630.041%ok1.6x (10%)
factorhessenbergf64512x512434.708 ms1.030.010%ok1.6x (10%)
factorhouseholderf64512x5120.673 ms1.566.2424%*ok4.9x (31%)
factorinvf64512x512159.895 ms1.680.031%ok1.3x (8%)
factorlu_iterative_refinef32>f64512x51253.593 ms1.670.081%ok2.9x (18%)
eigeneigenvalue_symmetricf64256x25643.571 ms0.510.020%ok0.9x (6%)
eigeneigen_symmetricf64256x256119.587 ms1.260.011%ok1.0x (6%)
eigeneigen_symmetric_jacobif64256x256584.781 ms0.260.000%ok1.0x (7%)
eigeneigenvaluef64256x25668.048 ms2.470.021%ok1.0x (6%)
eigenschurf64256x256185.943 ms2.260.011%ok1.2x (7%)
eigensvdf64256x256174.965 ms2.110.011%ok1.0x (7%)
eigenexpmf64256x25637.893 ms6.640.033%ok1.3x (8%)
eigensylvesterf64256x256390.383 ms2.580.001%ok1.0x (6%)
elementwiseexpf64n=838860858.003 ms0.142.319%ok1.1x (7%)
elementwiselogf64n=838860874.233 ms0.111.817%ok0.9x (6%)
elementwisesinf64n=838860858.646 ms0.142.299%ok1.0x (6%)
elementwisesqrtf64n=838860827.444 ms0.314.8919%ok1.0x (7%)
elementwiseerff64n=838860862.164 ms0.132.168%ok1.0x (6%)
sparsespmvf64512^2 grid, nnz 13086720.967 ms2.7128.17107%ok1.3x (8%)
sparsespmv_transf64512^2 grid, nnz 13086723.138 ms0.838.6833%ok1.0x (6%)
sparsecrs_densef64512^2 grid, nnz 1308672 x 164.519 ms9.2719.4874%ok1.7x (10%)
sparsespgemmf64256^2 grid, nnz 32665619.535 ms0.170.592%ok1.1x (7%)
sparselower_solve_levelf64192^2 grid, nnz(L) 10035643.252 ms0.625.1220%ok0.9x (6%)
directsparse_choleskyf6496^2 grid, nnz 45696166.490 ms0.070.000%ok1.0x (6%)
directsparse_ldltf6496^2 grid, nnz 4569649.649 ms0.230.010%ok1.1x (7%)
directsupernodal_ldltf6496^2 grid, nnz 4569628.916 ms0.390.030%ok1.1x (7%)
directsparse_luf6496^2 grid, nnz 45696107.379 ms0.210.010%ok0.9x (6%)
directsupernodal_luf6496^2 grid, nnz 45696131.631 ms0.170.010%ok1.0x (6%)
directnative_kluf6496^2 grid, nnz 4569651.293 ms0.440.010%ok1.0x (6%)
directsparse_qrf6448^2 grid, nnz 11328388.801 ms0.000.000%ok1.0x (6%)
orderingamdf6496^2 grid, nnz 45696, scrambled6.058 ms0.000.00n/aok1.0x (6%)
orderingcolamdf6496^2 grid, nnz 45696, columns scrambled9.506 ms0.000.00n/aok1.0x (6%)
orderingrcmf6496^2 grid, nnz 45696, scrambled0.856 ms0.000.00n/aok1.0x (6%)
orderingblock_triangular_formf6432 planted blocks, n 819226.683 ms0.000.00n/aok1.0x (6%)
krylovcg+ic0f64128^2 grid, n 1638443.101 ms0.000.00n/aok1.0x (6%)
krylovminres+diagf64128^2 grid, n 1638451.286 ms0.000.00n/aok1.0x (6%)
krylovbicg+ilu0f64128^2 grid, n 1638436.871 ms0.000.00n/aok1.0x (6%)
krylovbicgstab+ilu0f64128^2 grid, n 1638417.931 ms0.000.00n/aok1.0x (6%)
krylovbicgstab_ell+ilu0f64128^2 grid, n 1638425.154 ms0.000.00n/aok1.0x (6%)
krylovcgs+ilu0f64128^2 grid, n 1638421.456 ms0.000.00n/aok1.0x (6%)
krylovgmres30+ilu0f64128^2 grid, n 1638447.519 ms0.000.00n/aok1.1x (7%)
krylovqmr+ilu0f64128^2 grid, n 1638443.657 ms0.000.00n/aok0.9x (6%)
krylovtfqmr+ilu0f64128^2 grid, n 1638426.641 ms0.000.00n/aok0.9x (6%)
krylovidr_s1+ilu0f64128^2 grid, n 1638434.188 ms0.000.00n/aok1.0x (6%)
krylovidr_s2+ilu0f64128^2 grid, n 1638437.655 ms0.000.00n/aok1.0x (6%)
krylovidr_s4+ilu0f64128^2 grid, n 1638453.020 ms0.000.00n/aok1.0x (6%)
krylovidr_s8+ilu0f64128^2 grid, n 1638447.587 ms0.000.00n/aok1.0x (6%)
pcdiagonalf64128^2 grid, rows scaled0.061 ms0.000.00n/aok1.2x (8%)
pcilu_0f64128^2 grid4.770 ms0.000.00n/aok1.0x (6%)
pcilutf64128^2 grid43.149 ms0.000.00n/aok1.0x (6%)
pcic_0f64128^2 grid3.839 ms0.000.00n/aok1.1x (7%)
pcildlf64128^2 grid3.596 ms0.000.00n/aok1.0x (6%)
pcssorf64128^2 grid0.029 ms0.000.00n/aok1.3x (8%)
pcblock_diagonal4f64128^2 grid, rows scaled1.118 ms0.000.00n/aok1.1x (7%)
smootherjacobif64512^2 grid, nnz 13086722.286 ms1.1512.8349%ok1.0x (6%)
smoothergauss_seidelf64512^2 grid, nnz 13086722.034 ms1.2914.4255%ok1.0x (6%)
smoothersor1.5f64512^2 grid, nnz 13086722.442 ms1.0712.0146%ok1.0x (6%)
mgpoisson2d_vcyclef64255^2 grid, 6 levels, 10 V(1,1)24.290 ms1.5811.9746%ok1.0x (6%)
sparse_eiglanczosf6464x65 Laplacian, k=4, m=8048.740 ms0.000.00n/aok1.1x (7%)
sparse_eigarnoldif6464x65 convection-diffusion, k=4, m=8074.842 ms0.000.00n/aok0.9x (6%)
sparse_eigshift_invertf6464x65 Laplacian, k=4 nearest 0, m=40127.283 ms0.000.00n/aok0.9x (6%)
sparse_eigpower_iterationf6464x65 Laplacian, 300 steps9.789 ms0.000.00n/aok1.0x (6%)
mixeddotf32>f64n=83886083.302 ms5.0820.3277%ok1.0x (6%)
mixedgemmf32>f64512x512x5124.258 ms63.040.9926%ok6.0x (38%)
mixedblock_scaled_doti8+f32n=8388608, block 320.893 ms18.7919.9776%ok1.0x (6%)
integerdoti32n=8388608, |v| <= 153.023 ms5.5522.20n/aok0.9x (6%)
integerdoti16>i32n=8388608, |v| <= 151.229 ms13.6527.31n/aok1.0x (6%)
integerdoti8>i32n=8388608, |v| <= 150.885 ms18.9518.95n/aok1.0x (6%)
integerdotu8i8>i32n=8388608, |v| <= 150.757 ms22.1822.18n/aok0.9x (6%)
integergemmi32512^3, |v| <= 20474.165 ms64.440.76n/aok6.6x (41%)
integergemmi16>i32512^3, |v| <= 20476.555 ms40.950.32n/aok4.2x (26%)
integergemmi8>i32512^3, |v| <= 1274.432 ms60.560.35n/aok6.6x (41%)
integergemm_quadi8>i32512^3, |v| <= 1272.109 ms127.310.75n/aok3.6x (23%)
integergemm_quadu8i8>i32512^3, |v| <= 1272.294 ms117.030.69n/aok2.9x (18%)

* working set fits in the last-level cache: a cache rate, shown as a % of the DRAM roofline.