wgblas
    Preparing search index...

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

    Benchmark results for sswap on Nvidia Geforce Gtx 1650.

    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0064 0.0806 0.0027 0.1893 42.6%
    64 0.0064 0.1604 0.0028 0.3721 43.1%
    128 0.0064 0.3200 0.0028 0.7356 43.5%
    512 0.0064 1.2864 0.0028 2.8764 44.7%
    1024 0.0064 2.5793 0.0028 5.8514 44.1%
    4096 0.0065 10.0392 0.0029 22.7556 44.1%
    16384 0.0074 35.2344 0.0039 67.7025 52.0%
    65536 0.0091 115.7880 0.0068 153.1215 75.6%
    262144 0.0289 145.1517 0.0284 147.8533 98.2%
    1048576 0.1020 164.4825 0.1021 164.3794 100.1%
    4194304 0.3928 170.8405 0.3988 168.2972 101.5%
    16777216 1.5606 172.0105 1.6036 167.3972 102.8%

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

    sswap-default GB/s chart

    sswap-default ms chart

    • sswap.js — WebGPU benchmark script
    • sswap.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 0.0071 0.0717 0.0029 0.1788 40.1%
    64 0.0069 0.1478 0.0028 0.3596 41.1%
    128 0.0072 0.2857 0.0029 0.6995 40.8%
    512 0.0072 1.1353 0.0031 2.6528 42.8%
    1024 0.0073 2.2456 0.0031 5.2784 42.5%
    4096 0.0080 8.1431 0.0034 19.1402 42.5%
    16384 0.0103 25.3622 0.0054 48.6172 52.2%
    65536 0.0287 36.5714 0.0269 39.0095 93.8%
    262144 0.1005 41.7493 0.1004 41.7560 100.0%
    1048576 0.3893 43.0910 0.3937 42.6146 101.1%
    4194304 1.5422 43.5143 1.5677 42.8082 101.6%

    sswap-stride4 GB/s chart

    sswap-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 0.0031 0.1667 50.0%
    64 0.0061 0.1667 0.0031 0.3351 49.7%
    128 0.0061 0.3333 0.0034 0.6009 55.5%
    512 0.0063 1.2995 0.0043 1.9176 67.8%
    1024 0.0065 2.5222 0.0042 3.9084 64.5%
    4096 0.0094 7.0017 0.0058 11.2527 62.2%
    16384 0.0352 7.4540 0.0338 7.7649 96.0%
    65536 0.1385 7.5729 0.1336 7.8486 96.5%
    262144 0.5427 7.7283 0.5253 7.9849 96.8%
    1048576 2.1908 7.6579 2.0900 8.0272 95.4%

    sswap-stride32 GB/s chart

    sswap-stride32 ms chart

    Nvidia Geforce Gtx 1650 — stride = 256
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    512 0.0063 1.2929 0.0042 1.9692 65.7%
    1024 0.0066 2.4976 0.0042 3.8788 64.4%
    4096 0.0103 6.3504 0.0093 7.0621 89.9%
    16384 0.0638 4.1094 0.0698 3.7561 109.4%
    65536 0.2908 3.6056 0.2972 3.5284 102.2%

    sswap-stride256 GB/s chart

    sswap-stride256 ms chart

    See also:

    Settings
    Member Visibility

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