Table of Contents

Benchmarks

This document presents benchmarks conducted to evaluate various operations on a specific system configuration. The system details are as follows:

BenchmarkDotNet v0.13.12, Windows 11 (10.0.22631.3296/23H2/2023Update/SunValley3)
AMD Ryzen 9 7940HS w/ Radeon 780M Graphics, 1 CPU, 16 logical and 8 physical cores
.NET SDK 8.0.101
  [Host]    : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI
  Scalar    : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT
  Vector128 : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX
  Vector256 : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX2
  Vector512 : .NET 8.0.1 (8.0.123.58001), X64 RyuJIT AVX-512F+CD+BW+DQ+VL+VBMI

The system supports vectorization up to 512 bits. Benchmarks are conducted both with and without vectorization for each supported vectorization size.

Benchmark Jobs

Four distinct jobs are included in the benchmarks:

  • Scalar: No SIMD support
  • Vector128: using 128-bit SIMD support
  • Vector256: using 256-bit SIMD support
  • Vector512: using 512-bit SIMD support

Benchmark Scenarios

The benchmarks encompass the following scenarios:

  • Baseline_*: Simple iteration without explicit optimizations
  • LINQ_*: using LINQ (when available)
  • System_*: using System.Numerics.Tensors
  • NetFabric_*: using NetFabric.Numerics.Tensors

The source code for the benchmarks can be accessed here.

Test Cases and Considerations

Benchmarks were conducted on small source spans (5 items) as well as larger ones (100 items). Various scenarios of operators and their applications were covered. The baseline for each scenario involved equivalent operations using a for loop, with no optimizations applied other than those automatically added by the JIT compiler.

It's important to note that while these benchmarks provide insights into the performance characteristics of the library, they are not exhaustive and are not intended to serve as a comprehensive performance analysis. They aim to offer a general understanding of the library's performance under different scenarios.

Results

Negate

Applying a vectorizable unary operator on a span.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 45.573 ns 0.2966 ns baseline
System_Double Scalar Double 100 29.756 ns 0.1381 ns 1.53x faster
NetFabric_Double Scalar Double 100 26.658 ns 0.1550 ns 1.71x faster
Baseline_Double Vector128 Double 100 47.152 ns 2.1215 ns 1.06x slower
System_Double Vector128 Double 100 12.626 ns 0.1176 ns 3.61x faster
NetFabric_Double Vector128 Double 100 14.744 ns 0.2733 ns 3.09x faster
Baseline_Double Vector256 Double 100 32.666 ns 0.3011 ns 1.40x faster
System_Double Vector256 Double 100 10.927 ns 0.0521 ns 4.17x faster
NetFabric_Double Vector256 Double 100 9.304 ns 0.0298 ns 4.90x faster
Baseline_Double Vector512 Double 100 47.302 ns 1.9834 ns 1.05x slower
System_Double Vector512 Double 100 11.154 ns 0.0398 ns 4.09x faster
NetFabric_Double Vector512 Double 100 9.813 ns 0.0464 ns 4.64x faster
Baseline_Float Scalar Float 100 45.061 ns 0.2763 ns baseline
System_Float Scalar Float 100 29.691 ns 0.6771 ns 1.52x faster
NetFabric_Float Scalar Float 100 27.657 ns 0.4055 ns 1.63x faster
Baseline_Float Vector128 Float 100 47.998 ns 2.1608 ns 1.05x slower
System_Float Vector128 Float 100 9.232 ns 0.0872 ns 4.88x faster
NetFabric_Float Vector128 Float 100 9.161 ns 0.0973 ns 4.92x faster
Baseline_Float Vector256 Float 100 32.591 ns 0.2715 ns 1.38x faster
System_Float Vector256 Float 100 8.053 ns 0.0349 ns 5.59x faster
NetFabric_Float Vector256 Float 100 7.088 ns 0.0503 ns 6.35x faster
Baseline_Float Vector512 Float 100 47.134 ns 2.0154 ns 1.05x slower
System_Float Vector512 Float 100 8.638 ns 0.0700 ns 5.22x faster
NetFabric_Float Vector512 Float 100 8.214 ns 0.0763 ns 5.48x faster
Baseline_Half Scalar Half 100 767.767 ns 2.5521 ns baseline
System_Half Scalar Half 100 760.971 ns 1.9508 ns 1.01x faster
NetFabric_Half Scalar Half 100 758.768 ns 2.0240 ns 1.01x faster
Baseline_Half Vector128 Half 100 642.664 ns 0.7075 ns 1.19x faster
System_Half Vector128 Half 100 639.995 ns 2.0392 ns 1.20x faster
NetFabric_Half Vector128 Half 100 644.216 ns 2.7896 ns 1.19x faster
Baseline_Half Vector256 Half 100 647.412 ns 4.6062 ns 1.19x faster
System_Half Vector256 Half 100 639.948 ns 4.1512 ns 1.20x faster
NetFabric_Half Vector256 Half 100 642.387 ns 3.0127 ns 1.20x faster
Baseline_Half Vector512 Half 100 644.374 ns 2.6637 ns 1.19x faster
System_Half Vector512 Half 100 644.819 ns 6.5954 ns 1.19x faster
NetFabric_Half Vector512 Half 100 649.468 ns 3.9853 ns 1.18x faster
Baseline_Int Scalar Int 100 35.906 ns 0.4704 ns baseline
System_Int Scalar Int 100 29.608 ns 0.7775 ns 1.21x faster
NetFabric_Int Scalar Int 100 30.740 ns 0.2080 ns 1.17x faster
Baseline_Int Vector128 Int 100 36.527 ns 0.3254 ns 1.02x slower
System_Int Vector128 Int 100 9.831 ns 0.0608 ns 3.65x faster
NetFabric_Int Vector128 Int 100 13.548 ns 0.0619 ns 2.65x faster
Baseline_Int Vector256 Int 100 36.767 ns 0.4214 ns 1.03x slower
System_Int Vector256 Int 100 7.482 ns 0.0695 ns 4.80x faster
NetFabric_Int Vector256 Int 100 12.252 ns 0.0617 ns 2.93x faster
Baseline_Int Vector512 Int 100 36.467 ns 0.1628 ns 1.02x slower
System_Int Vector512 Int 100 12.493 ns 0.0274 ns 2.87x faster
NetFabric_Int Vector512 Int 100 11.466 ns 0.0437 ns 3.13x faster
Baseline_Long Scalar Long 100 36.301 ns 0.2350 ns baseline
System_Long Scalar Long 100 29.256 ns 0.1246 ns 1.24x faster
NetFabric_Long Scalar Long 100 31.201 ns 0.1614 ns 1.16x faster
Baseline_Long Vector128 Long 100 36.455 ns 0.2352 ns 1.00x slower
System_Long Vector128 Long 100 14.563 ns 0.1236 ns 2.49x faster
NetFabric_Long Vector128 Long 100 18.382 ns 0.1982 ns 1.98x faster
Baseline_Long Vector256 Long 100 36.484 ns 0.3588 ns 1.00x slower
System_Long Vector256 Long 100 11.685 ns 0.0431 ns 3.11x faster
NetFabric_Long Vector256 Long 100 10.807 ns 0.0716 ns 3.36x faster
Baseline_Long Vector512 Long 100 36.663 ns 0.2058 ns 1.01x slower
System_Long Vector512 Long 100 10.474 ns 0.0600 ns 3.47x faster
NetFabric_Long Vector512 Long 100 10.635 ns 0.1332 ns 3.41x faster
Baseline_Short Scalar Short 100 40.951 ns 0.3544 ns baseline
System_Short Scalar Short 100 40.886 ns 0.4036 ns 1.00x faster
NetFabric_Short Scalar Short 100 36.349 ns 0.2416 ns 1.13x faster
Baseline_Short Vector128 Short 100 41.558 ns 0.2594 ns 1.01x slower
System_Short Vector128 Short 100 7.253 ns 0.0736 ns 5.65x faster
NetFabric_Short Vector128 Short 100 8.133 ns 0.0460 ns 5.04x faster
Baseline_Short Vector256 Short 100 41.490 ns 0.3011 ns 1.01x slower
System_Short Vector256 Short 100 5.168 ns 0.0518 ns 7.93x faster
NetFabric_Short Vector256 Short 100 6.172 ns 0.0951 ns 6.64x faster
Baseline_Short Vector512 Short 100 41.288 ns 0.1954 ns 1.01x slower
System_Short Vector512 Short 100 5.517 ns 0.0320 ns 7.43x faster
NetFabric_Short Vector512 Short 100 6.571 ns 0.0211 ns 6.24x faster

Add

Applying a vectorizable binary operator on two spans.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 42.838 ns 0.5474 ns baseline
System_Double Scalar Double 100 30.240 ns 0.1343 ns 1.42x faster
NetFabric_Double Scalar Double 100 27.300 ns 0.1176 ns 1.57x faster
Baseline_Double Vector128 Double 100 47.762 ns 0.6866 ns 1.12x slower
System_Double Vector128 Double 100 20.282 ns 0.0841 ns 2.11x faster
NetFabric_Double Vector128 Double 100 20.364 ns 0.1156 ns 2.10x faster
Baseline_Double Vector256 Double 100 47.177 ns 0.3005 ns 1.10x slower
System_Double Vector256 Double 100 16.653 ns 0.0518 ns 2.57x faster
NetFabric_Double Vector256 Double 100 13.296 ns 0.0712 ns 3.22x faster
Baseline_Double Vector512 Double 100 47.769 ns 0.4299 ns 1.12x slower
System_Double Vector512 Double 100 17.823 ns 0.0964 ns 2.40x faster
NetFabric_Double Vector512 Double 100 13.988 ns 0.0918 ns 3.06x faster
Baseline_Float Scalar Float 100 46.910 ns 0.2536 ns baseline
System_Float Scalar Float 100 30.420 ns 0.1175 ns 1.54x faster
NetFabric_Float Scalar Float 100 27.097 ns 0.1320 ns 1.73x faster
Baseline_Float Vector128 Float 100 40.948 ns 0.6031 ns 1.14x faster
System_Float Vector128 Float 100 14.220 ns 0.0738 ns 3.30x faster
NetFabric_Float Vector128 Float 100 12.781 ns 0.0379 ns 3.67x faster
Baseline_Float Vector256 Float 100 40.751 ns 0.3792 ns 1.15x faster
System_Float Vector256 Float 100 10.541 ns 0.0632 ns 4.45x faster
NetFabric_Float Vector256 Float 100 10.306 ns 0.0569 ns 4.55x faster
Baseline_Float Vector512 Float 100 40.859 ns 0.5954 ns 1.15x faster
System_Float Vector512 Float 100 11.316 ns 0.0517 ns 4.15x faster
NetFabric_Float Vector512 Float 100 10.274 ns 0.0545 ns 4.57x faster
Baseline_Half Scalar Half 100 999.391 ns 0.9589 ns baseline
System_Half Scalar Half 100 997.099 ns 3.8572 ns 1.00x faster
NetFabric_Half Scalar Half 100 995.374 ns 3.8658 ns 1.00x faster
Baseline_Half Vector128 Half 100 911.213 ns 3.2180 ns 1.10x faster
System_Half Vector128 Half 100 885.748 ns 2.5029 ns 1.13x faster
NetFabric_Half Vector128 Half 100 887.482 ns 3.0970 ns 1.13x faster
Baseline_Half Vector256 Half 100 915.089 ns 6.6291 ns 1.09x faster
System_Half Vector256 Half 100 887.685 ns 3.2944 ns 1.13x faster
NetFabric_Half Vector256 Half 100 887.629 ns 3.2086 ns 1.13x faster
Baseline_Half Vector512 Half 100 912.210 ns 2.0226 ns 1.10x faster
System_Half Vector512 Half 100 886.560 ns 2.3231 ns 1.13x faster
NetFabric_Half Vector512 Half 100 885.757 ns 1.8336 ns 1.13x faster
Baseline_Int Scalar Int 100 47.679 ns 0.3331 ns baseline
System_Int Scalar Int 100 37.232 ns 0.1545 ns 1.28x faster
NetFabric_Int Scalar Int 100 33.680 ns 0.1329 ns 1.42x faster
Baseline_Int Vector128 Int 100 50.704 ns 0.1286 ns 1.06x slower
System_Int Vector128 Int 100 15.799 ns 0.0674 ns 3.02x faster
NetFabric_Int Vector128 Int 100 13.488 ns 0.1433 ns 3.54x faster
Baseline_Int Vector256 Int 100 48.321 ns 0.3998 ns 1.01x slower
System_Int Vector256 Int 100 12.430 ns 0.0301 ns 3.84x faster
NetFabric_Int Vector256 Int 100 10.519 ns 0.1974 ns 4.53x faster
Baseline_Int Vector512 Int 100 48.420 ns 0.3729 ns 1.02x slower
System_Int Vector512 Int 100 13.326 ns 0.0623 ns 3.58x faster
NetFabric_Int Vector512 Int 100 10.261 ns 0.0552 ns 4.65x faster
Baseline_Long Scalar Long 100 47.667 ns 0.3977 ns baseline
System_Long Scalar Long 100 36.912 ns 0.1346 ns 1.29x faster
NetFabric_Long Scalar Long 100 35.274 ns 0.1377 ns 1.35x faster
Baseline_Long Vector128 Long 100 47.490 ns 0.6803 ns 1.00x faster
System_Long Vector128 Long 100 20.943 ns 0.0900 ns 2.28x faster
NetFabric_Long Vector128 Long 100 25.661 ns 0.1937 ns 1.86x faster
Baseline_Long Vector256 Long 100 48.330 ns 0.6507 ns 1.01x slower
System_Long Vector256 Long 100 17.558 ns 0.0552 ns 2.71x faster
NetFabric_Long Vector256 Long 100 14.723 ns 0.1270 ns 3.24x faster
Baseline_Long Vector512 Long 100 47.401 ns 0.5282 ns 1.01x faster
System_Long Vector512 Long 100 17.561 ns 0.0920 ns 2.71x faster
NetFabric_Long Vector512 Long 100 13.943 ns 0.0869 ns 3.42x faster
Baseline_Short Scalar Short 100 52.775 ns 0.2422 ns baseline
System_Short Scalar Short 100 45.858 ns 0.1945 ns 1.15x faster
NetFabric_Short Scalar Short 100 43.392 ns 0.1510 ns 1.22x faster
Baseline_Short Vector128 Short 100 54.653 ns 1.5272 ns 1.04x slower
System_Short Vector128 Short 100 9.671 ns 0.0582 ns 5.45x faster
NetFabric_Short Vector128 Short 100 11.470 ns 0.1263 ns 4.60x faster
Baseline_Short Vector256 Short 100 52.536 ns 0.3891 ns 1.00x faster
System_Short Vector256 Short 100 6.239 ns 0.0472 ns 8.46x faster
NetFabric_Short Vector256 Short 100 9.148 ns 0.0540 ns 5.77x faster
Baseline_Short Vector512 Short 100 52.830 ns 0.3333 ns 1.00x slower
System_Short Vector512 Short 100 7.545 ns 0.0464 ns 7.00x faster
NetFabric_Short Vector512 Short 100 9.329 ns 0.0400 ns 5.66x faster

Min

Applying a vectorizable binary operator with propagation of NaN.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 77.822 ns 0.3262 ns baseline
System_Double Scalar Double 100 68.871 ns 0.4840 ns 1.13x faster
NetFabric_Double Scalar Double 100 86.981 ns 0.6853 ns 1.12x slower
Baseline_Double Vector128 Double 100 96.596 ns 1.7143 ns 1.24x slower
System_Double Vector128 Double 100 34.331 ns 0.2659 ns 2.27x faster
NetFabric_Double Vector128 Double 100 39.739 ns 0.2777 ns 1.96x faster
Baseline_Double Vector256 Double 100 82.876 ns 1.2040 ns 1.06x slower
System_Double Vector256 Double 100 20.034 ns 0.1042 ns 3.89x faster
NetFabric_Double Vector256 Double 100 21.825 ns 0.1232 ns 3.57x faster
Baseline_Double Vector512 Double 100 51.904 ns 0.2927 ns 1.50x faster
System_Double Vector512 Double 100 27.181 ns 0.1602 ns 2.86x faster
NetFabric_Double Vector512 Double 100 21.818 ns 0.0917 ns 3.57x faster
Baseline_Float Scalar Float 100 78.342 ns 0.8300 ns baseline
System_Float Scalar Float 100 68.697 ns 0.4501 ns 1.14x faster
NetFabric_Float Scalar Float 100 85.124 ns 0.6446 ns 1.09x slower
Baseline_Float Vector128 Float 100 101.167 ns 0.5612 ns 1.29x slower
System_Float Vector128 Float 100 19.894 ns 0.1286 ns 3.94x faster
NetFabric_Float Vector128 Float 100 21.724 ns 0.1460 ns 3.61x faster
Baseline_Float Vector256 Float 100 102.121 ns 1.1186 ns 1.30x slower
System_Float Vector256 Float 100 13.085 ns 0.1873 ns 5.99x faster
NetFabric_Float Vector256 Float 100 14.734 ns 0.0741 ns 5.32x faster
Baseline_Float Vector512 Float 100 51.630 ns 0.2217 ns 1.52x faster
System_Float Vector512 Float 100 13.855 ns 0.0906 ns 5.65x faster
NetFabric_Float Vector512 Float 100 13.948 ns 0.0720 ns 5.61x faster
Baseline_Half Scalar Half 100 883.888 ns 2.8508 ns baseline
System_Half Scalar Half 100 880.562 ns 1.4729 ns 1.00x faster
NetFabric_Half Scalar Half 100 893.825 ns 2.5456 ns 1.01x slower
Baseline_Half Vector128 Half 100 757.449 ns 4.0445 ns 1.17x faster
System_Half Vector128 Half 100 763.809 ns 2.6164 ns 1.16x faster
NetFabric_Half Vector128 Half 100 760.726 ns 4.4007 ns 1.16x faster
Baseline_Half Vector256 Half 100 756.132 ns 3.4296 ns 1.17x faster
System_Half Vector256 Half 100 764.713 ns 4.1691 ns 1.16x faster
NetFabric_Half Vector256 Half 100 759.633 ns 2.6422 ns 1.16x faster
Baseline_Half Vector512 Half 100 865.753 ns 2.2094 ns 1.02x faster
System_Half Vector512 Half 100 863.612 ns 2.5026 ns 1.02x faster
NetFabric_Half Vector512 Half 100 867.104 ns 1.8447 ns 1.02x faster
Baseline_Int Scalar Int 100 51.976 ns 0.3824 ns baseline
System_Int Scalar Int 100 35.131 ns 0.2025 ns 1.48x faster
NetFabric_Int Scalar Int 100 33.484 ns 0.1669 ns 1.55x faster
Baseline_Int Vector128 Int 100 51.267 ns 0.2203 ns 1.01x faster
System_Int Vector128 Int 100 14.950 ns 0.1110 ns 3.48x faster
NetFabric_Int Vector128 Int 100 12.695 ns 0.0619 ns 4.09x faster
Baseline_Int Vector256 Int 100 52.505 ns 0.4324 ns 1.01x slower
System_Int Vector256 Int 100 12.779 ns 0.0542 ns 4.07x faster
NetFabric_Int Vector256 Int 100 10.157 ns 0.0634 ns 5.12x faster
Baseline_Int Vector512 Int 100 52.323 ns 0.1651 ns 1.01x slower
System_Int Vector512 Int 100 13.103 ns 0.0907 ns 3.97x faster
NetFabric_Int Vector512 Int 100 10.275 ns 0.0453 ns 5.06x faster
Baseline_Long Scalar Long 100 52.713 ns 0.4538 ns baseline
System_Long Scalar Long 100 34.924 ns 0.3090 ns 1.51x faster
NetFabric_Long Scalar Long 100 33.423 ns 0.0927 ns 1.58x faster
Baseline_Long Vector128 Long 100 75.910 ns 2.1183 ns 1.45x slower
System_Long Vector128 Long 100 21.505 ns 0.0783 ns 2.45x faster
NetFabric_Long Vector128 Long 100 25.719 ns 0.2887 ns 2.05x faster
Baseline_Long Vector256 Long 100 75.608 ns 1.5240 ns 1.43x slower
System_Long Vector256 Long 100 17.885 ns 0.0486 ns 2.95x faster
NetFabric_Long Vector256 Long 100 15.290 ns 0.0587 ns 3.45x faster
Baseline_Long Vector512 Long 100 74.174 ns 0.3423 ns 1.41x slower
System_Long Vector512 Long 100 18.898 ns 0.1029 ns 2.79x faster
NetFabric_Long Vector512 Long 100 14.279 ns 0.1114 ns 3.69x faster
Baseline_Short Scalar Short 100 53.586 ns 0.3659 ns baseline
System_Short Scalar Short 100 46.125 ns 0.3830 ns 1.16x faster
NetFabric_Short Scalar Short 100 41.945 ns 0.5680 ns 1.28x faster
Baseline_Short Vector128 Short 100 59.540 ns 0.5334 ns 1.11x slower
System_Short Vector128 Short 100 10.624 ns 0.0495 ns 5.04x faster
NetFabric_Short Vector128 Short 100 12.153 ns 0.0755 ns 4.41x faster
Baseline_Short Vector256 Short 100 59.430 ns 0.4117 ns 1.11x slower
System_Short Vector256 Short 100 6.224 ns 0.0330 ns 8.62x faster
NetFabric_Short Vector256 Short 100 9.478 ns 0.0505 ns 5.65x faster
Baseline_Short Vector512 Short 100 59.545 ns 0.4133 ns 1.11x slower
System_Short Vector512 Short 100 7.552 ns 0.0554 ns 7.10x faster
NetFabric_Short Vector512 Short 100 8.903 ns 0.0664 ns 6.02x faster

Add Value

Applying a vectorizable binary operator on a span and a fixed scalar value.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 30.491 ns 0.2496 ns baseline
System_Double Scalar Double 100 30.656 ns 0.2597 ns 1.01x slower
NetFabric_Double Scalar Double 100 23.163 ns 0.2204 ns 1.32x faster
Baseline_Double Vector128 Double 100 28.773 ns 0.2517 ns 1.06x faster
System_Double Vector128 Double 100 14.443 ns 0.0950 ns 2.11x faster
NetFabric_Double Vector128 Double 100 15.381 ns 0.1444 ns 1.99x faster
Baseline_Double Vector256 Double 100 28.832 ns 0.2141 ns 1.06x faster
System_Double Vector256 Double 100 10.916 ns 0.0433 ns 2.79x faster
NetFabric_Double Vector256 Double 100 9.413 ns 0.1037 ns 3.24x faster
Baseline_Double Vector512 Double 100 28.806 ns 0.3041 ns 1.06x faster
System_Double Vector512 Double 100 11.155 ns 0.0419 ns 2.73x faster
NetFabric_Double Vector512 Double 100 10.368 ns 0.0690 ns 2.94x faster
Baseline_Float Scalar Float 100 31.094 ns 0.3901 ns baseline
System_Float Scalar Float 100 29.603 ns 0.3048 ns 1.05x faster
NetFabric_Float Scalar Float 100 23.029 ns 0.1304 ns 1.35x faster
Baseline_Float Vector128 Float 100 48.725 ns 1.8406 ns 1.56x slower
System_Float Vector128 Float 100 9.178 ns 0.0628 ns 3.39x faster
NetFabric_Float Vector128 Float 100 9.732 ns 0.1437 ns 3.20x faster
Baseline_Float Vector256 Float 100 28.264 ns 0.2363 ns 1.10x faster
System_Float Vector256 Float 100 8.336 ns 0.0831 ns 3.73x faster
NetFabric_Float Vector256 Float 100 6.873 ns 0.0598 ns 4.52x faster
Baseline_Float Vector512 Float 100 48.878 ns 1.4262 ns 1.57x slower
System_Float Vector512 Float 100 8.364 ns 0.0216 ns 3.71x faster
NetFabric_Float Vector512 Float 100 8.289 ns 0.0368 ns 3.75x faster
Baseline_Half Scalar Half 100 992.662 ns 4.3637 ns baseline
System_Half Scalar Half 100 983.645 ns 3.8075 ns 1.01x faster
NetFabric_Half Scalar Half 100 1,000.728 ns 3.2709 ns 1.01x slower
Baseline_Half Vector128 Half 100 894.462 ns 2.2043 ns 1.11x faster
System_Half Vector128 Half 100 875.850 ns 2.5469 ns 1.13x faster
NetFabric_Half Vector128 Half 100 877.771 ns 2.2985 ns 1.13x faster
Baseline_Half Vector256 Half 100 895.050 ns 1.3509 ns 1.11x faster
System_Half Vector256 Half 100 875.080 ns 2.0255 ns 1.13x faster
NetFabric_Half Vector256 Half 100 879.065 ns 1.8564 ns 1.13x faster
Baseline_Half Vector512 Half 100 895.426 ns 1.7073 ns 1.11x faster
System_Half Vector512 Half 100 877.551 ns 3.3208 ns 1.13x faster
NetFabric_Half Vector512 Half 100 876.934 ns 2.6832 ns 1.13x faster
Baseline_Int Scalar Int 100 44.760 ns 0.6570 ns baseline
System_Int Scalar Int 100 36.937 ns 0.1208 ns 1.21x faster
NetFabric_Int Scalar Int 100 33.508 ns 0.1485 ns 1.34x faster
Baseline_Int Vector128 Int 100 35.064 ns 0.3169 ns 1.28x faster
System_Int Vector128 Int 100 11.550 ns 0.0417 ns 3.88x faster
NetFabric_Int Vector128 Int 100 10.248 ns 0.0604 ns 4.37x faster
Baseline_Int Vector256 Int 100 34.246 ns 0.1734 ns 1.31x faster
System_Int Vector256 Int 100 8.310 ns 0.0370 ns 5.39x faster
NetFabric_Int Vector256 Int 100 7.636 ns 0.0398 ns 5.86x faster
Baseline_Int Vector512 Int 100 34.726 ns 0.1851 ns 1.29x faster
System_Int Vector512 Int 100 12.269 ns 0.0535 ns 3.65x faster
NetFabric_Int Vector512 Int 100 10.999 ns 0.0325 ns 4.07x faster
Baseline_Long Scalar Long 100 45.567 ns 1.5187 ns baseline
System_Long Scalar Long 100 36.564 ns 0.2023 ns 1.24x faster
NetFabric_Long Scalar Long 100 33.437 ns 0.1774 ns 1.36x faster
Baseline_Long Vector128 Long 100 34.554 ns 0.3545 ns 1.31x faster
System_Long Vector128 Long 100 13.107 ns 0.0771 ns 3.46x faster
NetFabric_Long Vector128 Long 100 24.331 ns 0.2522 ns 1.86x faster
Baseline_Long Vector256 Long 100 34.498 ns 0.4009 ns 1.32x faster
System_Long Vector256 Long 100 11.127 ns 0.0865 ns 4.07x faster
NetFabric_Long Vector256 Long 100 9.689 ns 0.0644 ns 4.67x faster
Baseline_Long Vector512 Long 100 34.803 ns 0.1882 ns 1.30x faster
System_Long Vector512 Long 100 10.955 ns 0.0720 ns 4.14x faster
NetFabric_Long Vector512 Long 100 10.202 ns 0.0799 ns 4.44x faster
Baseline_Short Scalar Short 100 39.386 ns 0.4163 ns baseline
System_Short Scalar Short 100 40.853 ns 0.1484 ns 1.04x slower
NetFabric_Short Scalar Short 100 36.261 ns 0.1594 ns 1.09x faster
Baseline_Short Vector128 Short 100 39.172 ns 0.2904 ns 1.01x faster
System_Short Vector128 Short 100 6.859 ns 0.0512 ns 5.74x faster
NetFabric_Short Vector128 Short 100 7.834 ns 0.1299 ns 5.03x faster
Baseline_Short Vector256 Short 100 39.163 ns 0.5241 ns 1.01x faster
System_Short Vector256 Short 100 5.024 ns 0.0321 ns 7.84x faster
NetFabric_Short Vector256 Short 100 6.008 ns 0.0401 ns 6.55x faster
Baseline_Short Vector512 Short 100 38.894 ns 0.5157 ns 1.01x faster
System_Short Vector512 Short 100 5.725 ns 0.0396 ns 6.88x faster
NetFabric_Short Vector512 Short 100 5.886 ns 0.0306 ns 6.69x faster

AddMultiply

Applying a vectorizable ternary operator on three spans.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 53.194 ns 0.5459 ns baseline
System_Double Scalar Double 100 40.778 ns 0.1742 ns 1.30x faster
NetFabric_Double Scalar Double 100 38.343 ns 0.1443 ns 1.39x faster
Baseline_Double Vector128 Double 100 53.476 ns 0.4960 ns 1.01x slower
System_Double Vector128 Double 100 28.606 ns 0.1557 ns 1.86x faster
NetFabric_Double Vector128 Double 100 28.967 ns 0.1387 ns 1.84x faster
Baseline_Double Vector256 Double 100 53.079 ns 0.2441 ns 1.00x faster
System_Double Vector256 Double 100 19.916 ns 0.0737 ns 2.67x faster
NetFabric_Double Vector256 Double 100 18.981 ns 0.0716 ns 2.80x faster
Baseline_Double Vector512 Double 100 52.691 ns 0.4285 ns 1.01x faster
System_Double Vector512 Double 100 22.735 ns 0.1027 ns 2.34x faster
NetFabric_Double Vector512 Double 100 18.626 ns 0.0902 ns 2.86x faster
Baseline_Float Scalar Float 100 51.615 ns 0.2995 ns baseline
System_Float Scalar Float 100 40.463 ns 0.1519 ns 1.28x faster
NetFabric_Float Scalar Float 100 38.171 ns 0.0796 ns 1.35x faster
Baseline_Float Vector128 Float 100 52.654 ns 0.1693 ns 1.02x slower
System_Float Vector128 Float 100 19.467 ns 0.0949 ns 2.65x faster
NetFabric_Float Vector128 Float 100 18.488 ns 0.0830 ns 2.79x faster
Baseline_Float Vector256 Float 100 52.849 ns 0.3430 ns 1.02x slower
System_Float Vector256 Float 100 16.036 ns 0.0519 ns 3.22x faster
NetFabric_Float Vector256 Float 100 14.800 ns 0.0487 ns 3.49x faster
Baseline_Float Vector512 Float 100 53.026 ns 0.2116 ns 1.03x slower
System_Float Vector512 Float 100 16.049 ns 0.1000 ns 3.22x faster
NetFabric_Float Vector512 Float 100 15.217 ns 0.0473 ns 3.39x faster
Baseline_Half Scalar Half 100 1,618.204 ns 3.5398 ns baseline
System_Half Scalar Half 100 2,381.759 ns 2.3965 ns 1.47x slower
NetFabric_Half Scalar Half 100 2,375.177 ns 2.8007 ns 1.47x slower
Baseline_Half Vector128 Half 100 1,455.613 ns 6.4194 ns 1.11x faster
System_Half Vector128 Half 100 2,278.200 ns 14.9670 ns 1.41x slower
NetFabric_Half Vector128 Half 100 2,272.810 ns 9.7295 ns 1.40x slower
Baseline_Half Vector256 Half 100 1,459.872 ns 9.3435 ns 1.11x faster
System_Half Vector256 Half 100 2,272.492 ns 14.8960 ns 1.40x slower
NetFabric_Half Vector256 Half 100 2,270.518 ns 5.6785 ns 1.40x slower
Baseline_Half Vector512 Half 100 1,454.237 ns 6.5861 ns 1.11x faster
System_Half Vector512 Half 100 2,275.960 ns 12.6515 ns 1.41x slower
NetFabric_Half Vector512 Half 100 2,277.877 ns 13.1558 ns 1.41x slower
Baseline_Int Scalar Int 100 59.095 ns 0.4008 ns baseline
System_Int Scalar Int 100 52.993 ns 0.2280 ns 1.12x faster
NetFabric_Int Scalar Int 100 45.341 ns 0.1997 ns 1.30x faster
Baseline_Int Vector128 Int 100 59.250 ns 0.5185 ns 1.00x slower
System_Int Vector128 Int 100 19.522 ns 0.2302 ns 3.03x faster
NetFabric_Int Vector128 Int 100 18.133 ns 0.0982 ns 3.26x faster
Baseline_Int Vector256 Int 100 59.131 ns 0.4722 ns 1.00x slower
System_Int Vector256 Int 100 15.924 ns 0.0793 ns 3.71x faster
NetFabric_Int Vector256 Int 100 13.997 ns 0.0616 ns 4.22x faster
Baseline_Int Vector512 Int 100 58.863 ns 0.3824 ns 1.00x faster
System_Int Vector512 Int 100 14.588 ns 0.1058 ns 4.05x faster
NetFabric_Int Vector512 Int 100 14.360 ns 0.1070 ns 4.11x faster
Baseline_Long Scalar Long 100 59.422 ns 0.7016 ns baseline
System_Long Scalar Long 100 52.869 ns 0.4472 ns 1.12x faster
NetFabric_Long Scalar Long 100 45.050 ns 0.1657 ns 1.32x faster
Baseline_Long Vector128 Long 100 59.398 ns 0.5420 ns 1.00x faster
System_Long Vector128 Long 100 353.783 ns 1.3794 ns 5.95x slower
NetFabric_Long Vector128 Long 100 258.917 ns 1.1508 ns 4.36x slower
Baseline_Long Vector256 Long 100 59.045 ns 0.2597 ns 1.01x faster
System_Long Vector256 Long 100 363.682 ns 1.2358 ns 6.12x slower
NetFabric_Long Vector256 Long 100 147.048 ns 0.6580 ns 2.48x slower
Baseline_Long Vector512 Long 100 59.265 ns 0.5245 ns 1.00x faster
System_Long Vector512 Long 100 22.449 ns 0.1203 ns 2.65x faster
NetFabric_Long Vector512 Long 100 144.334 ns 1.3654 ns 2.43x slower
Baseline_Short Scalar Short 100 85.215 ns 0.5693 ns baseline
System_Short Scalar Short 100 79.092 ns 0.4696 ns 1.08x faster
NetFabric_Short Scalar Short 100 80.144 ns 0.5933 ns 1.06x faster
Baseline_Short Vector128 Short 100 83.962 ns 0.3168 ns 1.02x faster
System_Short Vector128 Short 100 12.624 ns 0.0586 ns 6.75x faster
NetFabric_Short Vector128 Short 100 13.848 ns 0.1233 ns 6.15x faster
Baseline_Short Vector256 Short 100 84.149 ns 0.3873 ns 1.01x faster
System_Short Vector256 Short 100 9.702 ns 0.0640 ns 8.78x faster
NetFabric_Short Vector256 Short 100 11.736 ns 0.0789 ns 7.26x faster
Baseline_Short Vector512 Short 100 84.477 ns 0.7474 ns 1.01x faster
System_Short Vector512 Short 100 19.894 ns 0.1089 ns 4.28x faster
NetFabric_Short Vector512 Short 100 12.510 ns 0.0460 ns 6.81x faster

DegreesToRadians

Applying a vectorizable ternary operator on a span and two fixed scalar values.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 105.363 ns 0.6721 ns baseline
System_Double Scalar Double 100 104.955 ns 0.9296 ns 1.00x faster
NetFabric_Double Scalar Double 100 108.631 ns 0.4020 ns 1.03x slower
Baseline_Double Vector128 Double 100 104.213 ns 0.4783 ns 1.01x faster
System_Double Vector128 Double 100 70.537 ns 0.5717 ns 1.49x faster
NetFabric_Double Vector128 Double 100 50.372 ns 0.3891 ns 2.09x faster
Baseline_Double Vector256 Double 100 104.494 ns 0.6641 ns 1.01x faster
System_Double Vector256 Double 100 50.886 ns 0.3590 ns 2.07x faster
NetFabric_Double Vector256 Double 100 23.781 ns 0.1212 ns 4.43x faster
Baseline_Double Vector512 Double 100 105.602 ns 0.7539 ns 1.00x slower
System_Double Vector512 Double 100 183.516 ns 0.8624 ns 1.74x slower
NetFabric_Double Vector512 Double 100 23.738 ns 0.0542 ns 4.44x faster
Baseline_Float Scalar Float 100 64.749 ns 0.2191 ns baseline
System_Float Scalar Float 100 65.010 ns 0.2194 ns 1.00x slower
NetFabric_Float Scalar Float 100 65.193 ns 0.3212 ns 1.01x slower
Baseline_Float Vector128 Float 100 64.669 ns 0.2393 ns 1.00x faster
System_Float Vector128 Float 100 39.316 ns 0.1564 ns 1.65x faster
NetFabric_Float Vector128 Float 100 15.905 ns 0.0988 ns 4.07x faster
Baseline_Float Vector256 Float 100 64.664 ns 0.3021 ns 1.00x faster
System_Float Vector256 Float 100 29.404 ns 0.3665 ns 2.20x faster
NetFabric_Float Vector256 Float 100 9.523 ns 0.0493 ns 6.80x faster
Baseline_Float Vector512 Float 100 64.599 ns 0.2632 ns 1.00x faster
System_Float Vector512 Float 100 77.022 ns 0.7295 ns 1.19x slower
NetFabric_Float Vector512 Float 100 12.352 ns 0.0653 ns 5.24x faster
Baseline_Half Scalar Half 100 999.253 ns 6.6407 ns baseline
System_Half Scalar Half 100 1,009.173 ns 4.2703 ns 1.01x slower
NetFabric_Half Scalar Half 100 2,561.898 ns 7.8650 ns 2.56x slower
Baseline_Half Vector128 Half 100 905.334 ns 3.8491 ns 1.10x faster
System_Half Vector128 Half 100 892.921 ns 1.9791 ns 1.12x faster
NetFabric_Half Vector128 Half 100 2,420.872 ns 9.8031 ns 2.42x slower
Baseline_Half Vector256 Half 100 903.875 ns 2.1924 ns 1.11x faster
System_Half Vector256 Half 100 892.802 ns 1.8886 ns 1.12x faster
NetFabric_Half Vector256 Half 100 2,418.932 ns 8.0544 ns 2.42x slower
Baseline_Half Vector512 Half 100 906.720 ns 4.1213 ns 1.10x faster
System_Half Vector512 Half 100 894.048 ns 3.8254 ns 1.12x faster
NetFabric_Half Vector512 Half 100 2,424.812 ns 13.1376 ns 2.43x slower

Sum

Applying a vectorizable aggregation operator on a span.

It additionally compares with the performance of LINQ's Sum(). However, it's worth noting that this method lacks support for the types short and Half. In such instances, LINQ's Aggregate() is employed instead.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 47.326 ns 0.5851 ns baseline
System_Double Scalar Double 100 39.147 ns 0.7875 ns 1.21x faster
NetFabric_Double Scalar Double 100 61.619 ns 0.2195 ns 1.30x slower
Baseline_Double Vector128 Double 100 42.618 ns 0.3416 ns 1.11x faster
System_Double Vector128 Double 100 11.973 ns 0.1163 ns 3.95x faster
NetFabric_Double Vector128 Double 100 29.314 ns 0.5716 ns 1.62x faster
Baseline_Double Vector256 Double 100 42.449 ns 0.2493 ns 1.11x faster
System_Double Vector256 Double 100 6.136 ns 0.0649 ns 7.71x faster
NetFabric_Double Vector256 Double 100 11.261 ns 0.1015 ns 4.20x faster
Baseline_Double Vector512 Double 100 42.334 ns 0.3927 ns 1.12x faster
System_Double Vector512 Double 100 5.625 ns 0.0293 ns 8.42x faster
NetFabric_Double Vector512 Double 100 13.771 ns 0.1643 ns 3.44x faster
Baseline_Float Scalar Float 100 42.699 ns 0.2814 ns baseline
System_Float Scalar Float 100 38.877 ns 0.1420 ns 1.10x faster
NetFabric_Float Scalar Float 100 61.195 ns 1.1528 ns 1.43x slower
Baseline_Float Vector128 Float 100 42.135 ns 0.1684 ns 1.01x faster
System_Float Vector128 Float 100 6.284 ns 0.0545 ns 6.79x faster
NetFabric_Float Vector128 Float 100 10.830 ns 0.0863 ns 3.94x faster
Baseline_Float Vector256 Float 100 42.783 ns 0.2612 ns 1.00x slower
System_Float Vector256 Float 100 4.567 ns 0.0313 ns 9.35x faster
NetFabric_Float Vector256 Float 100 7.253 ns 0.0512 ns 5.88x faster
Baseline_Float Vector512 Float 100 42.215 ns 0.2125 ns 1.01x faster
System_Float Vector512 Float 100 4.829 ns 0.0459 ns 8.84x faster
NetFabric_Float Vector512 Float 100 8.068 ns 0.0579 ns 5.29x faster
Baseline_Half Scalar Half 100 1,248.404 ns 3.8695 ns baseline
System_Half Scalar Half 100 1,261.046 ns 3.3137 ns 1.01x slower
NetFabric_Half Scalar Half 100 1,246.705 ns 5.5448 ns 1.00x faster
Baseline_Half Vector128 Half 100 1,209.489 ns 3.9070 ns 1.03x faster
System_Half Vector128 Half 100 1,226.410 ns 1.3758 ns 1.02x faster
NetFabric_Half Vector128 Half 100 1,213.822 ns 8.9948 ns 1.03x faster
Baseline_Half Vector256 Half 100 1,208.729 ns 2.2773 ns 1.03x faster
System_Half Vector256 Half 100 1,228.855 ns 3.7491 ns 1.02x faster
NetFabric_Half Vector256 Half 100 1,208.195 ns 1.4436 ns 1.03x faster
Baseline_Half Vector512 Half 100 1,236.813 ns 12.7384 ns 1.01x faster
System_Half Vector512 Half 100 1,249.052 ns 11.0453 ns 1.00x slower
NetFabric_Half Vector512 Half 100 1,251.404 ns 26.3882 ns 1.01x slower
Baseline_Int Scalar Int 100 27.755 ns 0.1998 ns baseline
LINQ_Int Scalar Int 100 26.892 ns 0.1653 ns 1.03x faster
System_Int Scalar Int 100 29.587 ns 0.1430 ns 1.07x slower
NetFabric_Int Scalar Int 100 26.959 ns 0.1393 ns 1.03x faster
Baseline_Int Vector128 Int 100 27.303 ns 0.2011 ns 1.02x faster
LINQ_Int Vector128 Int 100 8.549 ns 0.0409 ns 3.25x faster
System_Int Vector128 Int 100 5.547 ns 0.0283 ns 5.00x faster
NetFabric_Int Vector128 Int 100 11.656 ns 0.0453 ns 2.38x faster
Baseline_Int Vector256 Int 100 27.162 ns 0.1590 ns 1.02x faster
LINQ_Int Vector256 Int 100 6.058 ns 0.0549 ns 4.58x faster
System_Int Vector256 Int 100 4.551 ns 0.0410 ns 6.10x faster
NetFabric_Int Vector256 Int 100 6.314 ns 0.0908 ns 4.40x faster
Baseline_Int Vector512 Int 100 45.141 ns 0.3433 ns 1.63x slower
LINQ_Int Vector512 Int 100 5.990 ns 0.0493 ns 4.64x faster
System_Int Vector512 Int 100 4.382 ns 0.0712 ns 6.34x faster
NetFabric_Int Vector512 Int 100 6.208 ns 0.0640 ns 4.47x faster
Baseline_Long Scalar Long 100 45.088 ns 0.3162 ns baseline
LINQ_Long Scalar Long 100 26.957 ns 0.1937 ns 1.67x faster
System_Long Scalar Long 100 30.365 ns 0.2084 ns 1.48x faster
NetFabric_Long Scalar Long 100 27.270 ns 0.2998 ns 1.65x faster
Baseline_Long Vector128 Long 100 45.053 ns 0.1710 ns 1.00x faster
LINQ_Long Vector128 Long 100 28.189 ns 0.6928 ns 1.62x faster
System_Long Vector128 Long 100 8.667 ns 0.0958 ns 5.20x faster
NetFabric_Long Vector128 Long 100 23.970 ns 0.2298 ns 1.88x faster
Baseline_Long Vector256 Long 100 44.878 ns 0.2091 ns 1.00x faster
LINQ_Long Vector256 Long 100 8.948 ns 0.0359 ns 5.04x faster
System_Long Vector256 Long 100 6.441 ns 0.0544 ns 7.00x faster
NetFabric_Long Vector256 Long 100 11.967 ns 0.0714 ns 3.77x faster
Baseline_Long Vector512 Long 100 27.126 ns 0.1424 ns 1.66x faster
LINQ_Long Vector512 Long 100 9.445 ns 0.0874 ns 4.77x faster
System_Long Vector512 Long 100 9.783 ns 0.2136 ns 4.62x faster
NetFabric_Long Vector512 Long 100 11.440 ns 0.1473 ns 3.94x faster
Baseline_Short Scalar Short 100 39.626 ns 0.1431 ns baseline
System_Short Scalar Short 100 41.646 ns 0.1554 ns 1.05x slower
NetFabric_Short Scalar Short 100 45.522 ns 0.4657 ns 1.15x slower
Baseline_Short Vector128 Short 100 45.966 ns 0.2654 ns 1.16x slower
System_Short Vector128 Short 100 4.729 ns 0.0337 ns 8.38x faster
NetFabric_Short Vector128 Short 100 7.206 ns 0.0413 ns 5.50x faster
Baseline_Short Vector256 Short 100 45.998 ns 0.2975 ns 1.16x slower
System_Short Vector256 Short 100 3.947 ns 0.0294 ns 10.04x faster
NetFabric_Short Vector256 Short 100 4.827 ns 0.0566 ns 8.21x faster
Baseline_Short Vector512 Short 100 38.637 ns 0.2469 ns 1.03x faster
System_Short Vector512 Short 100 4.740 ns 0.0400 ns 8.36x faster
NetFabric_Short Vector512 Short 100 4.960 ns 0.0394 ns 7.99x faster

Sum2D

Applying a vectorizable aggregation operator on a span with two contiguos values for each element.

It also compares to the performance of LINQ's Aggregate(), as LINQ's Sum() does not support non-native numeric types.

Method Job Categories Count Mean StdDev Median Ratio
Baseline_Double Scalar Double 100 52.96 ns 0.183 ns 52.99 ns baseline
NetFabric_Double Scalar Double 100 35.17 ns 0.153 ns 35.18 ns 1.51x faster
Baseline_Double Vector128 Double 100 53.88 ns 0.522 ns 53.94 ns 1.02x slower
NetFabric_Double Vector128 Double 100 48.78 ns 0.257 ns 48.70 ns 1.09x faster
Baseline_Double Vector256 Double 100 54.02 ns 0.499 ns 53.96 ns 1.02x slower
NetFabric_Double Vector256 Double 100 24.94 ns 0.093 ns 24.97 ns 2.12x faster
Baseline_Double Vector512 Double 100 53.78 ns 0.319 ns 53.67 ns 1.01x slower
NetFabric_Double Vector512 Double 100 25.16 ns 0.064 ns 25.19 ns 2.10x faster
Baseline_Float Scalar Float 100 222.75 ns 0.541 ns 222.69 ns baseline
NetFabric_Float Scalar Float 100 36.00 ns 0.063 ns 36.00 ns 6.18x faster
Baseline_Float Vector128 Float 100 222.88 ns 0.807 ns 222.87 ns 1.00x slower
NetFabric_Float Vector128 Float 100 28.59 ns 0.117 ns 28.60 ns 7.79x faster
Baseline_Float Vector256 Float 100 222.69 ns 0.407 ns 222.74 ns 1.00x slower
NetFabric_Float Vector256 Float 100 16.06 ns 0.072 ns 16.05 ns 13.87x faster
Baseline_Float Vector512 Float 100 222.47 ns 0.646 ns 222.54 ns 1.00x faster
NetFabric_Float Vector512 Float 100 16.09 ns 0.047 ns 16.08 ns 13.84x faster
Baseline_Half Scalar Half 100 2,009.62 ns 1.739 ns 2,009.29 ns baseline
NetFabric_Half Scalar Half 100 2,031.01 ns 2.899 ns 2,029.84 ns 1.01x slower
Baseline_Half Vector128 Half 100 1,791.11 ns 3.574 ns 1,791.90 ns 1.12x faster
NetFabric_Half Vector128 Half 100 1,789.25 ns 6.415 ns 1,789.96 ns 1.12x faster
Baseline_Half Vector256 Half 100 1,790.95 ns 2.931 ns 1,791.40 ns 1.12x faster
NetFabric_Half Vector256 Half 100 1,786.29 ns 8.328 ns 1,785.79 ns 1.13x faster
Baseline_Half Vector512 Half 100 1,792.03 ns 5.916 ns 1,792.21 ns 1.12x faster
NetFabric_Half Vector512 Half 100 1,784.97 ns 2.908 ns 1,785.13 ns 1.13x faster
Baseline_Int Scalar Int 100 155.52 ns 1.035 ns 155.73 ns baseline
NetFabric_Int Scalar Int 100 46.21 ns 0.312 ns 46.16 ns 3.37x faster
Baseline_Int Vector128 Int 100 155.49 ns 0.525 ns 155.47 ns 1.00x faster
NetFabric_Int Vector128 Int 100 18.09 ns 0.191 ns 18.02 ns 8.60x faster
Baseline_Int Vector256 Int 100 155.41 ns 0.830 ns 155.68 ns 1.00x faster
NetFabric_Int Vector256 Int 100 16.31 ns 0.083 ns 16.30 ns 9.54x faster
Baseline_Int Vector512 Int 100 155.47 ns 0.662 ns 155.39 ns 1.00x faster
NetFabric_Int Vector512 Int 100 15.23 ns 0.163 ns 15.22 ns 10.21x faster
Baseline_Long Scalar Long 100 46.74 ns 0.214 ns 46.76 ns baseline
NetFabric_Long Scalar Long 100 44.23 ns 0.180 ns 44.27 ns 1.06x faster
Baseline_Long Vector128 Long 100 47.42 ns 0.634 ns 47.11 ns 1.01x slower
NetFabric_Long Vector128 Long 100 33.77 ns 0.260 ns 33.78 ns 1.38x faster
Baseline_Long Vector256 Long 100 46.83 ns 0.193 ns 46.80 ns 1.00x slower
NetFabric_Long Vector256 Long 100 24.60 ns 0.166 ns 24.55 ns 1.90x faster
Baseline_Long Vector512 Long 100 46.75 ns 0.240 ns 46.73 ns 1.00x slower
NetFabric_Long Vector512 Long 100 16.93 ns 0.201 ns 16.87 ns 2.76x faster
Baseline_Short Scalar Short 100 206.23 ns 0.565 ns 206.34 ns baseline
NetFabric_Short Scalar Short 100 70.70 ns 3.636 ns 69.14 ns 2.81x faster
Baseline_Short Vector128 Short 100 206.65 ns 0.640 ns 206.62 ns 1.00x slower
NetFabric_Short Vector128 Short 100 15.54 ns 0.122 ns 15.52 ns 13.27x faster
Baseline_Short Vector256 Short 100 206.70 ns 0.756 ns 206.53 ns 1.00x slower
NetFabric_Short Vector256 Short 100 19.11 ns 0.077 ns 19.12 ns 10.79x faster
Baseline_Short Vector512 Short 100 206.66 ns 1.016 ns 206.27 ns 1.00x slower
NetFabric_Short Vector512 Short 100 17.90 ns 0.213 ns 17.78 ns 11.52x faster

Sum3D

Applying a vectorizable aggregation operator on a span with three contiguos values for each element.

It also compares to the performance of LINQ's Aggregate(), as LINQ's Sum() does not support non-native numeric types.

Method Job Categories Count Mean StdDev Median Ratio
Baseline_Double Scalar Double 100 63.13 ns 1.191 ns 62.88 ns baseline
NetFabric_Double Scalar Double 100 64.18 ns 0.660 ns 64.02 ns 1.02x slower
Baseline_Double Vector128 Double 100 64.47 ns 0.678 ns 64.20 ns 1.02x slower
NetFabric_Double Vector128 Double 100 136.92 ns 2.452 ns 137.40 ns 2.17x slower
Baseline_Double Vector256 Double 100 65.29 ns 1.422 ns 65.45 ns 1.04x slower
NetFabric_Double Vector256 Double 100 77.85 ns 3.706 ns 77.32 ns 1.20x slower
Baseline_Double Vector512 Double 100 65.87 ns 1.049 ns 66.15 ns 1.04x slower
NetFabric_Double Vector512 Double 100 78.31 ns 1.251 ns 78.17 ns 1.24x slower
Baseline_Float Scalar Float 100 64.87 ns 1.669 ns 65.03 ns baseline
NetFabric_Float Scalar Float 100 66.72 ns 0.964 ns 67.09 ns 1.02x slower
Baseline_Float Vector128 Float 100 66.07 ns 1.175 ns 66.50 ns 1.01x slower
NetFabric_Float Vector128 Float 100 70.49 ns 1.354 ns 70.22 ns 1.08x slower
Baseline_Float Vector256 Float 100 64.62 ns 0.888 ns 64.97 ns 1.01x faster
NetFabric_Float Vector256 Float 100 47.88 ns 0.768 ns 47.76 ns 1.37x faster
Baseline_Float Vector512 Float 100 65.29 ns 1.584 ns 65.19 ns 1.01x slower
NetFabric_Float Vector512 Float 100 47.94 ns 0.577 ns 48.03 ns 1.36x faster
Baseline_Half Scalar Half 100 3,027.82 ns 11.188 ns 3,028.59 ns baseline
NetFabric_Half Scalar Half 100 3,011.36 ns 27.446 ns 3,002.45 ns 1.01x faster
Baseline_Half Vector128 Half 100 2,697.99 ns 23.967 ns 2,692.15 ns 1.12x faster
NetFabric_Half Vector128 Half 100 2,648.99 ns 16.402 ns 2,648.01 ns 1.14x faster
Baseline_Half Vector256 Half 100 2,674.79 ns 14.089 ns 2,675.71 ns 1.13x faster
NetFabric_Half Vector256 Half 100 2,657.25 ns 22.551 ns 2,652.22 ns 1.14x faster
Baseline_Half Vector512 Half 100 2,678.09 ns 13.898 ns 2,678.40 ns 1.13x faster
NetFabric_Half Vector512 Half 100 2,652.20 ns 17.853 ns 2,650.41 ns 1.14x faster
Baseline_Int Scalar Int 100 55.13 ns 0.827 ns 54.86 ns baseline
NetFabric_Int Scalar Int 100 77.42 ns 2.504 ns 77.02 ns 1.41x slower
Baseline_Int Vector128 Int 100 56.04 ns 0.775 ns 55.79 ns 1.02x slower
NetFabric_Int Vector128 Int 100 57.54 ns 0.452 ns 57.52 ns 1.04x slower
Baseline_Int Vector256 Int 100 56.18 ns 0.528 ns 56.14 ns 1.02x slower
NetFabric_Int Vector256 Int 100 46.41 ns 1.023 ns 46.17 ns 1.18x faster
Baseline_Int Vector512 Int 100 56.17 ns 0.538 ns 55.81 ns 1.02x slower
NetFabric_Int Vector512 Int 100 47.44 ns 1.304 ns 47.39 ns 1.17x faster
Baseline_Long Scalar Long 100 55.58 ns 0.647 ns 55.60 ns baseline
NetFabric_Long Scalar Long 100 79.05 ns 3.349 ns 78.05 ns 1.45x slower
Baseline_Long Vector128 Long 100 53.84 ns 0.598 ns 53.97 ns 1.03x faster
NetFabric_Long Vector128 Long 100 115.99 ns 4.999 ns 114.15 ns 2.20x slower
Baseline_Long Vector256 Long 100 54.18 ns 1.042 ns 53.96 ns 1.03x faster
NetFabric_Long Vector256 Long 100 65.29 ns 0.788 ns 64.94 ns 1.17x slower
Baseline_Long Vector512 Long 100 53.62 ns 0.473 ns 53.47 ns 1.04x faster
NetFabric_Long Vector512 Long 100 64.74 ns 0.961 ns 64.99 ns 1.17x slower
Baseline_Short Scalar Short 100 104.98 ns 2.734 ns 103.79 ns baseline
NetFabric_Short Scalar Short 100 116.80 ns 0.871 ns 116.76 ns 1.11x slower
Baseline_Short Vector128 Short 100 105.19 ns 1.461 ns 104.58 ns 1.00x slower
NetFabric_Short Vector128 Short 100 54.18 ns 2.162 ns 53.45 ns 1.93x faster
Baseline_Short Vector256 Short 100 104.76 ns 0.832 ns 104.62 ns 1.00x faster
NetFabric_Short Vector256 Short 100 51.95 ns 0.549 ns 51.96 ns 2.02x faster
Baseline_Short Vector512 Short 100 104.62 ns 0.647 ns 104.67 ns 1.00x faster
NetFabric_Short Vector512 Short 100 52.01 ns 0.386 ns 51.97 ns 2.02x faster

Sum4D

Applying a vectorizable aggregation operator on a span with four contiguos values for each element.

It also compares to the performance of LINQ's Aggregate(), as LINQ's Sum() does not support non-native numeric types.

Method Job Categories Count Mean StdDev Median Ratio
Baseline_Double Scalar Double 100 70.27 ns 1.159 ns 70.54 ns baseline
NetFabric_Double Scalar Double 100 74.65 ns 1.107 ns 74.23 ns 1.06x slower
Baseline_Double Vector128 Double 100 66.23 ns 0.673 ns 66.16 ns 1.06x faster
NetFabric_Double Vector128 Double 100 266.95 ns 3.067 ns 267.10 ns 3.80x slower
Baseline_Double Vector256 Double 100 70.21 ns 2.883 ns 70.53 ns 1.05x faster
NetFabric_Double Vector256 Double 100 51.89 ns 0.490 ns 51.98 ns 1.35x faster
Baseline_Double Vector512 Double 100 71.16 ns 1.484 ns 71.40 ns 1.01x slower
NetFabric_Double Vector512 Double 100 53.24 ns 1.668 ns 52.35 ns 1.32x faster
Baseline_Float Scalar Float 100 70.08 ns 1.180 ns 70.32 ns baseline
NetFabric_Float Scalar Float 100 78.04 ns 0.886 ns 78.15 ns 1.11x slower
Baseline_Float Vector128 Float 100 70.11 ns 1.266 ns 70.35 ns 1.00x slower
NetFabric_Float Vector128 Float 100 51.27 ns 0.503 ns 51.20 ns 1.37x faster
Baseline_Float Vector256 Float 100 70.51 ns 0.976 ns 70.50 ns 1.01x slower
NetFabric_Float Vector256 Float 100 28.12 ns 0.812 ns 27.82 ns 2.50x faster
Baseline_Float Vector512 Float 100 71.39 ns 1.198 ns 71.67 ns 1.02x slower
NetFabric_Float Vector512 Float 100 28.20 ns 0.806 ns 27.82 ns 2.47x faster
Baseline_Half Scalar Half 100 3,619.66 ns 24.005 ns 3,625.74 ns baseline
NetFabric_Half Scalar Half 100 4,107.22 ns 39.515 ns 4,099.78 ns 1.14x slower
Baseline_Half Vector128 Half 100 2,901.28 ns 36.231 ns 2,907.40 ns 1.25x faster
NetFabric_Half Vector128 Half 100 3,621.06 ns 50.761 ns 3,612.47 ns 1.00x slower
Baseline_Half Vector256 Half 100 2,926.03 ns 25.950 ns 2,928.69 ns 1.24x faster
NetFabric_Half Vector256 Half 100 3,641.74 ns 34.245 ns 3,637.19 ns 1.01x slower
Baseline_Half Vector512 Half 100 2,904.28 ns 25.974 ns 2,903.98 ns 1.25x faster
NetFabric_Half Vector512 Half 100 3,620.77 ns 30.743 ns 3,612.92 ns 1.00x slower
Baseline_Int Scalar Int 100 64.53 ns 1.218 ns 64.64 ns baseline
NetFabric_Int Scalar Int 100 91.91 ns 0.867 ns 92.21 ns 1.42x slower
Baseline_Int Vector128 Int 100 66.54 ns 1.323 ns 66.69 ns 1.03x slower
NetFabric_Int Vector128 Int 100 35.50 ns 0.474 ns 35.67 ns 1.82x faster
Baseline_Int Vector256 Int 100 66.49 ns 0.943 ns 66.77 ns 1.03x slower
NetFabric_Int Vector256 Int 100 19.73 ns 0.320 ns 19.69 ns 3.27x faster
Baseline_Int Vector512 Int 100 67.16 ns 2.221 ns 66.95 ns 1.04x slower
NetFabric_Int Vector512 Int 100 19.66 ns 0.254 ns 19.65 ns 3.29x faster
Baseline_Long Scalar Long 100 68.33 ns 1.170 ns 68.47 ns baseline
NetFabric_Long Scalar Long 100 92.38 ns 3.018 ns 91.48 ns 1.36x slower
Baseline_Long Vector128 Long 100 65.33 ns 0.962 ns 65.22 ns 1.05x faster
NetFabric_Long Vector128 Long 100 224.12 ns 2.034 ns 223.69 ns 3.28x slower
Baseline_Long Vector256 Long 100 66.62 ns 1.365 ns 66.62 ns 1.03x faster
NetFabric_Long Vector256 Long 100 36.37 ns 0.685 ns 36.52 ns 1.88x faster
Baseline_Long Vector512 Long 100 66.71 ns 1.552 ns 66.69 ns 1.02x faster
NetFabric_Long Vector512 Long 100 36.98 ns 0.629 ns 37.20 ns 1.85x faster
Baseline_Short Scalar Short 100 96.96 ns 1.010 ns 97.01 ns baseline
NetFabric_Short Scalar Short 100 137.71 ns 2.069 ns 138.09 ns 1.42x slower
Baseline_Short Vector128 Short 100 96.97 ns 0.634 ns 96.82 ns 1.00x slower
NetFabric_Short Vector128 Short 100 22.32 ns 0.225 ns 22.30 ns 4.35x faster
Baseline_Short Vector256 Short 100 96.67 ns 2.173 ns 96.08 ns 1.00x slower
NetFabric_Short Vector256 Short 100 18.00 ns 0.273 ns 17.98 ns 5.38x faster
Baseline_Short Vector512 Short 100 95.26 ns 1.179 ns 94.80 ns 1.02x faster
NetFabric_Short Vector512 Short 100 21.14 ns 0.326 ns 21.10 ns 4.58x faster

Min aggregation

Applying a vectorizable aggregation operator with propagation os NaN.

Method Job Categories Count Mean StdDev Ratio
Baseline_Double Scalar Double 100 88.695 ns 0.4477 ns baseline
System_Double Scalar Double 100 77.290 ns 0.5142 ns 1.15x faster
NetFabric_Double Scalar Double 100 101.372 ns 0.8989 ns 1.14x slower
Baseline_Double Vector128 Double 100 99.235 ns 0.8462 ns 1.12x slower
System_Double Vector128 Double 100 37.595 ns 0.3627 ns 2.36x faster
NetFabric_Double Vector128 Double 100 74.457 ns 0.2241 ns 1.19x faster
Baseline_Double Vector256 Double 100 99.448 ns 1.1002 ns 1.12x slower
System_Double Vector256 Double 100 19.433 ns 0.1401 ns 4.57x faster
NetFabric_Double Vector256 Double 100 37.875 ns 0.2884 ns 2.34x faster
Baseline_Double Vector512 Double 100 72.462 ns 0.2434 ns 1.22x faster
System_Double Vector512 Double 100 28.316 ns 0.1643 ns 3.13x faster
NetFabric_Double Vector512 Double 100 38.506 ns 0.2427 ns 2.30x faster
Baseline_Float Scalar Float 100 100.707 ns 0.3968 ns baseline
System_Float Scalar Float 100 74.501 ns 0.4437 ns 1.35x faster
NetFabric_Float Scalar Float 100 88.893 ns 0.3883 ns 1.13x faster
Baseline_Float Vector128 Float 100 93.727 ns 0.3726 ns 1.07x faster
System_Float Vector128 Float 100 17.961 ns 0.1359 ns 5.61x faster
NetFabric_Float Vector128 Float 100 36.605 ns 0.1388 ns 2.75x faster
Baseline_Float Vector256 Float 100 96.138 ns 0.7183 ns 1.05x faster
System_Float Vector256 Float 100 10.558 ns 0.0520 ns 9.54x faster
NetFabric_Float Vector256 Float 100 30.799 ns 0.1294 ns 3.27x faster
Baseline_Float Vector512 Float 100 72.845 ns 0.2197 ns 1.38x faster
System_Float Vector512 Float 100 14.931 ns 0.2321 ns 6.75x faster
NetFabric_Float Vector512 Float 100 23.989 ns 0.0737 ns 4.20x faster
Baseline_Half Scalar Half 100 1,152.279 ns 5.4053 ns baseline
System_Half Scalar Half 100 180.660 ns 0.6765 ns 6.38x faster
NetFabric_Half Scalar Half 100 1,128.877 ns 4.8094 ns 1.02x faster
Baseline_Half Vector128 Half 100 1,096.543 ns 5.6732 ns 1.05x faster
System_Half Vector128 Half 100 180.565 ns 1.6048 ns 6.38x faster
NetFabric_Half Vector128 Half 100 1,089.019 ns 5.2743 ns 1.06x faster
Baseline_Half Vector256 Half 100 1,096.248 ns 6.2655 ns 1.05x faster
System_Half Vector256 Half 100 179.587 ns 0.4680 ns 6.42x faster
NetFabric_Half Vector256 Half 100 1,083.846 ns 5.0029 ns 1.06x faster
Baseline_Half Vector512 Half 100 1,264.763 ns 3.0543 ns 1.10x slower
System_Half Vector512 Half 100 179.811 ns 0.7011 ns 6.41x faster
NetFabric_Half Vector512 Half 100 1,263.840 ns 3.9612 ns 1.10x slower
Baseline_Int Scalar Int 100 34.827 ns 0.1450 ns baseline
System_Int Scalar Int 100 34.970 ns 0.4225 ns 1.00x slower
NetFabric_Int Scalar Int 100 35.142 ns 0.2962 ns 1.01x slower
Baseline_Int Vector128 Int 100 35.000 ns 0.2081 ns 1.00x slower
System_Int Vector128 Int 100 6.652 ns 0.0848 ns 5.24x faster
NetFabric_Int Vector128 Int 100 14.140 ns 0.0707 ns 2.46x faster
Baseline_Int Vector256 Int 100 34.743 ns 0.2514 ns 1.00x faster
System_Int Vector256 Int 100 3.935 ns 0.0226 ns 8.85x faster
NetFabric_Int Vector256 Int 100 9.600 ns 0.0822 ns 3.63x faster
Baseline_Int Vector512 Int 100 35.223 ns 0.3104 ns 1.01x slower
System_Int Vector512 Int 100 2.844 ns 0.0248 ns 12.25x faster
NetFabric_Int Vector512 Int 100 9.713 ns 0.0722 ns 3.59x faster
Baseline_Long Scalar Long 100 34.862 ns 0.2250 ns baseline
System_Long Scalar Long 100 35.469 ns 0.5156 ns 1.02x slower
NetFabric_Long Scalar Long 100 35.329 ns 0.2556 ns 1.01x slower
Baseline_Long Vector128 Long 100 71.426 ns 0.3344 ns 2.05x slower
System_Long Vector128 Long 100 20.049 ns 0.1214 ns 1.74x faster
NetFabric_Long Vector128 Long 100 27.073 ns 0.1602 ns 1.29x faster
Baseline_Long Vector256 Long 100 71.732 ns 0.5353 ns 2.06x slower
System_Long Vector256 Long 100 8.965 ns 0.0369 ns 3.89x faster
NetFabric_Long Vector256 Long 100 14.426 ns 0.1072 ns 2.42x faster
Baseline_Long Vector512 Long 100 71.470 ns 0.3791 ns 2.05x slower
System_Long Vector512 Long 100 6.234 ns 0.0301 ns 5.59x faster
NetFabric_Long Vector512 Long 100 13.922 ns 0.0954 ns 2.50x faster
Baseline_Short Scalar Short 100 39.547 ns 0.1179 ns baseline
System_Short Scalar Short 100 39.819 ns 0.2629 ns 1.01x slower
NetFabric_Short Scalar Short 100 39.546 ns 0.3363 ns 1.00x slower
Baseline_Short Vector128 Short 100 40.087 ns 0.1960 ns 1.01x slower
System_Short Vector128 Short 100 3.567 ns 0.0207 ns 11.09x faster
NetFabric_Short Vector128 Short 100 10.459 ns 0.0921 ns 3.78x faster
Baseline_Short Vector256 Short 100 40.558 ns 0.4019 ns 1.02x slower
System_Short Vector256 Short 100 2.834 ns 0.0335 ns 13.94x faster
NetFabric_Short Vector256 Short 100 12.507 ns 0.1092 ns 3.16x faster
Baseline_Short Vector512 Short 100 40.248 ns 0.3349 ns 1.02x slower
System_Short Vector512 Short 100 2.710 ns 0.0450 ns 14.63x faster
NetFabric_Short Vector512 Short 100 13.168 ns 0.1512 ns 3.00x faster

MinMax

Applying two vectorizable aggregation operators on a single iteration of a span.

Method Job Categories Count Mean StdDev Median Ratio
Baseline_Double Scalar Double 100 179.35 ns 0.721 ns 179.23 ns baseline
NetFabric_Double Scalar Double 100 280.46 ns 1.142 ns 280.48 ns 1.56x slower
Baseline_Double Vector128 Double 100 162.75 ns 0.844 ns 162.67 ns 1.10x faster
NetFabric_Double Vector128 Double 100 29.13 ns 0.211 ns 29.10 ns 6.16x faster
Baseline_Double Vector256 Double 100 163.13 ns 1.237 ns 162.72 ns 1.10x faster
NetFabric_Double Vector256 Double 100 23.43 ns 0.106 ns 23.45 ns 7.65x faster
Baseline_Double Vector512 Double 100 85.79 ns 0.456 ns 85.89 ns 2.09x faster
NetFabric_Double Vector512 Double 100 21.53 ns 0.147 ns 21.57 ns 8.33x faster
Baseline_Float Scalar Float 100 170.32 ns 4.771 ns 168.32 ns baseline
NetFabric_Float Scalar Float 100 278.95 ns 0.886 ns 278.88 ns 1.64x slower
Baseline_Float Vector128 Float 100 159.66 ns 0.784 ns 159.44 ns 1.06x faster
NetFabric_Float Vector128 Float 100 23.01 ns 0.243 ns 23.11 ns 7.38x faster
Baseline_Float Vector256 Float 100 160.10 ns 1.203 ns 160.16 ns 1.07x faster
NetFabric_Float Vector256 Float 100 40.18 ns 0.178 ns 40.20 ns 4.25x faster
Baseline_Float Vector512 Float 100 85.27 ns 0.305 ns 85.35 ns 2.00x faster
NetFabric_Float Vector512 Float 100 17.24 ns 0.068 ns 17.24 ns 9.91x faster
Baseline_Half Scalar Half 100 1,556.85 ns 8.238 ns 1,555.36 ns baseline
NetFabric_Half Scalar Half 100 424.22 ns 2.930 ns 424.24 ns 3.67x faster
Baseline_Half Vector128 Half 100 1,326.75 ns 6.993 ns 1,326.94 ns 1.17x faster
NetFabric_Half Vector128 Half 100 443.36 ns 3.124 ns 442.03 ns 3.51x faster
Baseline_Half Vector256 Half 100 1,329.31 ns 8.359 ns 1,328.91 ns 1.17x faster
NetFabric_Half Vector256 Half 100 445.48 ns 2.852 ns 445.42 ns 3.49x faster
Baseline_Half Vector512 Half 100 1,683.51 ns 5.413 ns 1,684.22 ns 1.08x slower
NetFabric_Half Vector512 Half 100 443.25 ns 3.760 ns 443.83 ns 3.52x faster
Baseline_Int Scalar Int 100 56.99 ns 0.471 ns 56.94 ns baseline
NetFabric_Int Scalar Int 100 56.65 ns 0.396 ns 56.69 ns 1.01x faster
Baseline_Int Vector128 Int 100 74.98 ns 2.101 ns 74.10 ns 1.32x slower
NetFabric_Int Vector128 Int 100 15.97 ns 0.068 ns 15.97 ns 3.57x faster
Baseline_Int Vector256 Int 100 58.38 ns 0.607 ns 58.35 ns 1.02x slower
NetFabric_Int Vector256 Int 100 13.69 ns 0.158 ns 13.66 ns 4.16x faster
Baseline_Int Vector512 Int 100 58.12 ns 0.684 ns 57.98 ns 1.02x slower
NetFabric_Int Vector512 Int 100 13.79 ns 0.087 ns 13.78 ns 4.14x faster
Baseline_Long Scalar Long 100 65.65 ns 6.839 ns 62.71 ns baseline
NetFabric_Long Scalar Long 100 59.87 ns 0.458 ns 59.96 ns 1.11x faster
Baseline_Long Vector128 Long 100 116.27 ns 0.414 ns 116.27 ns 1.77x slower
NetFabric_Long Vector128 Long 100 33.53 ns 0.188 ns 33.51 ns 1.97x faster
Baseline_Long Vector256 Long 100 133.37 ns 0.618 ns 133.28 ns 2.02x slower
NetFabric_Long Vector256 Long 100 19.29 ns 0.081 ns 19.32 ns 3.43x faster
Baseline_Long Vector512 Long 100 133.90 ns 0.909 ns 133.86 ns 2.03x slower
NetFabric_Long Vector512 Long 100 18.21 ns 0.248 ns 18.15 ns 3.63x faster
Baseline_Short Scalar Short 100 57.21 ns 0.209 ns 57.17 ns baseline
NetFabric_Short Scalar Short 100 77.70 ns 1.397 ns 77.86 ns 1.36x slower
Baseline_Short Vector128 Short 100 99.30 ns 0.467 ns 99.44 ns 1.74x slower
NetFabric_Short Vector128 Short 100 15.76 ns 0.133 ns 15.73 ns 3.63x faster
Baseline_Short Vector256 Short 100 73.83 ns 0.381 ns 73.89 ns 1.29x slower
NetFabric_Short Vector256 Short 100 23.80 ns 0.355 ns 23.95 ns 2.40x faster
Baseline_Short Vector512 Short 100 74.44 ns 0.724 ns 74.45 ns 1.30x slower
NetFabric_Short Vector512 Short 100 23.79 ns 0.333 ns 23.66 ns 2.41x faster
Edit this page