wgblas
    Preparing search index...

    Module benchmarks/nvidia-geforce-gtx-1650/srotm

    Benchmark results for srotm on Nvidia Geforce Gtx 1650.

    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0061 0.0842 0.0029 0.1778 47.4%
    64 0.0061 0.1667 0.0028 0.3657 45.6%
    128 0.0061 0.3333 0.0028 0.7314 45.6%
    512 0.0061 1.3333 0.0029 2.8132 47.4%
    1024 0.0061 2.6667 0.0029 5.6264 47.4%
    4096 0.0061 10.6667 0.0031 21.3333 50.0%
    16384 0.0069 38.0139 0.0039 67.4239 56.4%
    65536 0.0085 123.6528 0.0072 145.6356 84.9%
    262144 0.0289 145.3126 0.0274 153.1215 94.9%
    1048576 0.1024 163.8400 0.1008 166.5199 98.4%
    4194304 0.3963 169.3437 0.3936 170.4863 99.3%
    16777216 1.5601 172.0617 1.5664 171.3692 100.4%

    Efficiency = wgblas GB/s ÷ cuBLAS GB/s × 100. 100% means parity with cuBLAS; values above 100% mean wgblas achieved greater throughput.

    srotm-default GB/s chart

    srotm-default ms chart

    • srotm.js — WebGPU benchmark script
    • srotm.c — CUDA / cuBLAS reference script

    Unless noted otherwise, every result above uses unit stride (incx = incy = 1) — the normal case, and the coalesced, best-case GPU access pattern. Real usage sometimes passes a non-unit stride (e.g. operating on a row or column of a larger matrix, where incx = lda), which breaks memory coalescing and costs measurably more. This section sweeps a few representative strides to characterize that cost separately, collapsed below by default — expand a stride to see its table and chart.

    Nvidia Geforce Gtx 1650 — stride = 4
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0061 0.0833 0.0031 0.1633 51.0%
    64 0.0061 0.1667 0.0029 0.3536 47.1%
    128 0.0061 0.3333 0.0037 0.5590 59.6%
    512 0.0061 1.3333 0.0032 2.5347 52.6%
    1024 0.0061 2.6667 0.0035 4.7407 56.3%
    4096 0.0066 9.8937 0.0038 17.3559 57.0%
    16384 0.0085 30.6816 0.0054 48.9075 62.7%
    65536 0.0304 34.5290 0.0270 38.8707 88.8%
    262144 0.1024 40.9600 0.1005 41.7360 98.1%
    1048576 0.3921 42.7833 0.3950 42.4765 100.7%
    4194304 1.5442 43.4589 1.5708 42.7240 101.7%

    srotm-stride4 GB/s chart

    srotm-stride4 ms chart

    Nvidia Geforce Gtx 1650 — stride = 32
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0061 0.0833 0.0032 0.1616 51.6%
    64 0.0061 0.1667 0.0032 0.3168 52.6%
    128 0.0061 0.3333 0.0037 0.5590 59.6%
    512 0.0065 1.2611 0.0040 2.0645 61.1%
    1024 0.0067 2.4498 0.0040 4.0960 59.8%
    4096 0.0099 6.6278 0.0058 11.2527 58.9%
    16384 0.0362 7.2367 0.0348 7.5294 96.1%
    65536 0.1331 7.8769 0.1340 7.8243 100.7%
    262144 0.5156 8.1348 0.5330 7.8694 103.4%
    1048576 2.1496 7.8049 2.1507 7.8007 100.1%

    srotm-stride32 GB/s chart

    srotm-stride32 ms chart

    Nvidia Geforce Gtx 1650 — stride = 256
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    512 0.0065 1.2549 0.0041 2.0157 62.3%
    1024 0.0067 2.4381 0.0038 4.2667 57.1%
    4096 0.0113 5.7935 0.0084 7.7576 74.7%
    16384 0.0635 4.1301 0.0672 3.9000 105.9%
    65536 0.2888 3.6312 0.3006 3.4878 104.1%

    srotm-stride256 GB/s chart

    srotm-stride256 ms chart

    See also:

    Settings
    Member Visibility

    On This Page
    Nvidia Geforce Gtx 1650 — wgblas vs cuBLASSee alsoStride sweep