wgblas
    Preparing search index...

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

    Benchmark results for ssymv on Nvidia Geforce Gtx 1650.

    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0083 0.2994 0.0041 0.6142 48.8%
    64 0.0087 1.0499 0.0045 2.0142 52.1%
    128 0.0102 3.3750 0.0055 6.2791 53.7%
    256 0.0164 8.2188 0.0063 21.3063 38.6%
    512 0.0287 18.5357 0.0108 49.3551 37.6%
    1024 0.0722 29.2417 0.0331 63.8143 45.8%
    1280 0.1147 28.7277 0.0477 69.0080 41.6%
    2048 0.2575 32.6920 0.1097 76.7440 42.6%
    4096 1.2717 26.4300 0.4049 83.0199 31.8%

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

    ssymv-default GB/s chart

    ssymv-default ms chart

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

    Unless noted otherwise, every result above uses a tight lda (no padding). Padding the row stride changes throughput here — the exact mechanism and shape of that effect is routine-specific — collapsed below by default, expand a pad value to see its table and chart.

    Nvidia Geforce Gtx 1650 — pad = 0
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0078 0.3184 0.0040 0.6190 51.4%
    64 0.0068 1.3271 0.0044 2.0882 63.6%
    128 0.0081 4.2772 0.0056 6.2069 68.9%
    256 0.0122 11.0446 0.0062 21.6907 50.9%
    512 0.0245 21.6956 0.0106 50.0996 43.3%
    1024 0.0696 30.3235 0.0328 64.4375 47.1%
    1280 0.1106 29.7917 0.0476 69.2401 43.0%
    2048 0.2477 33.9779 0.1086 77.4786 43.9%
    4096 1.1981 28.0536 0.4029 83.4287 33.6%

    ssymv-pad0 GB/s chart

    ssymv-pad0 ms chart

    Nvidia Geforce Gtx 1650 — pad = 1
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0066 0.3786 0.0040 0.6240 60.7%
    64 0.0068 1.3396 0.0044 2.0882 64.2%
    128 0.0081 4.2688 0.0054 6.4286 66.4%
    256 0.0105 12.8489 0.0063 21.3604 60.2%
    512 0.0184 28.8333 0.0113 46.9816 61.4%
    1024 0.0532 39.6538 0.0335 63.0220 62.9%
    1280 0.0920 35.8184 0.0492 66.9223 53.5%
    2048 0.1884 44.6739 0.1114 75.5645 59.1%
    4096 1.0898 30.8428 0.3966 84.7549 36.4%

    ssymv-pad1 GB/s chart

    ssymv-pad1 ms chart

    Nvidia Geforce Gtx 1650 — pad = 8
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0066 0.3786 0.0036 0.6842 55.3%
    64 0.0070 1.3057 0.0041 2.2362 58.4%
    128 0.0082 4.2188 0.0048 7.2483 58.2%
    256 0.0108 12.5238 0.0058 23.2486 53.9%
    512 0.0198 26.8087 0.0096 55.4524 48.3%
    1024 0.0577 36.5866 0.0300 70.3454 52.0%
    1280 0.0933 35.3147 0.0460 71.6493 49.3%
    2048 0.2012 41.8387 0.1061 79.3365 52.7%
    4096 1.1387 29.5180 0.3935 85.4270 34.6%

    ssymv-pad8 GB/s chart

    ssymv-pad8 ms chart

    Nvidia Geforce Gtx 1650 — pad = 32
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0066 0.3786 0.0037 0.6812 55.6%
    64 0.0068 1.3333 0.0040 2.2903 58.2%
    128 0.0082 4.2188 0.0048 7.2241 58.4%
    256 0.0121 11.1323 0.0057 23.6404 47.1%
    512 0.0240 22.1736 0.0096 55.3600 40.1%
    1024 0.0701 30.1022 0.0298 70.8363 42.5%
    1280 0.1065 30.9421 0.0437 75.3182 41.1%
    2048 0.2583 32.5887 0.1034 81.3861 40.0%
    4096 1.1898 28.2494 0.3903 86.1098 32.8%

    ssymv-pad32 GB/s chart

    ssymv-pad32 ms chart

    Nvidia Geforce Gtx 1650 — pad = 64
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0066 0.3786 0.0036 0.6903 54.9%
    64 0.0068 1.3302 0.0041 2.2362 59.5%
    128 0.0082 4.2188 0.0048 7.1523 59.0%
    256 0.0122 11.0013 0.0058 23.2486 47.3%
    512 0.0240 22.1440 0.0096 55.3600 40.0%
    1024 0.0718 29.4046 0.0301 70.1584 41.9%
    1280 0.1067 30.8772 0.0448 73.5954 42.0%
    2048 0.2561 32.8697 0.1035 81.2981 40.4%
    4096 1.1927 28.1815 0.3928 85.5697 32.9%

    ssymv-pad64 GB/s chart

    ssymv-pad64 ms chart

    Nvidia Geforce Gtx 1650 — pad = 128
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0066 0.3805 0.0037 0.6724 56.6%
    64 0.0070 1.2909 0.0041 2.2188 58.2%
    128 0.0082 4.2188 0.0047 7.2973 57.8%
    256 0.0123 10.9583 0.0057 23.6404 46.4%
    512 0.0238 22.3677 0.0095 55.7315 40.1%
    1024 0.0714 29.5693 0.0302 69.9724 42.3%
    1280 0.1062 31.0214 0.0449 73.3333 42.3%
    2048 0.2692 31.2659 0.1047 80.3912 38.9%
    4096 1.2349 27.2183 0.4133 81.3325 33.5%

    ssymv-pad128 GB/s chart

    ssymv-pad128 ms chart

    See also:

    Settings
    Member Visibility

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