wgblas
    Preparing search index...

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

    Benchmark results for scopy on Nvidia Geforce Gtx 1650.

    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0059 0.0435 0.0026 0.0970 44.8%
    64 0.0059 0.0863 0.0026 0.1951 44.2%
    128 0.0061 0.1675 0.0027 0.3855 43.5%
    512 0.0061 0.6702 0.0033 1.2308 54.4%
    1024 0.0060 1.3545 0.0027 3.0476 44.4%
    4096 0.0061 5.3895 0.0028 11.9070 45.3%
    16384 0.0063 20.7919 0.0030 43.5745 47.7%
    65536 0.0081 64.7589 0.0045 116.1986 55.7%
    262144 0.0186 112.9931 0.0160 131.2032 86.1%
    1048576 0.0594 141.2414 0.0531 158.0133 89.4%
    4194304 0.2171 154.5660 0.2045 164.0836 94.2%
    16777216 0.8437 159.0770 0.8041 166.9175 95.3%

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

    scopy-default GB/s chart

    scopy-default ms chart

    • scopy.js — WebGPU benchmark script
    • scopy.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 0.0072 0.0357 0.0028 0.0930 38.4%
    64 0.0073 0.0697 0.0032 0.1576 44.2%
    128 0.0072 0.1432 0.0036 0.2832 50.6%
    512 0.0071 0.5792 0.0033 1.2249 47.3%
    1024 0.0074 1.1058 0.0031 2.6667 41.5%
    4096 0.0078 4.1796 0.0033 9.7990 42.7%
    16384 0.0086 15.2268 0.0041 32.0000 47.6%
    65536 0.0207 25.2840 0.0221 23.6934 106.7%
    262144 0.0782 26.8151 0.0794 26.4045 101.6%
    1048576 0.3069 27.3323 0.3070 27.3209 100.0%
    4194304 1.2207 27.4889 1.2108 27.7126 99.2%

    scopy-stride4 GB/s chart

    scopy-stride4 ms chart

    Nvidia Geforce Gtx 1650 — stride = 32
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0070 0.0365 0.0030 0.0847 43.1%
    64 0.0072 0.0713 0.0031 0.1658 43.0%
    128 0.0070 0.1461 0.0032 0.3168 46.1%
    512 0.0073 0.5602 0.0036 1.1278 49.7%
    1024 0.0076 1.0756 0.0036 2.2555 47.7%
    4096 0.0096 3.4020 0.0045 7.2113 47.2%
    16384 0.0205 6.4000 0.0181 7.2367 88.4%
    65536 0.0695 7.5433 0.0652 8.0412 93.8%
    262144 0.2662 7.8769 0.2519 8.3252 94.6%
    1048576 1.0463 8.0175 0.9955 8.4262 95.1%

    scopy-stride32 GB/s chart

    scopy-stride32 ms chart

    Nvidia Geforce Gtx 1650 — stride = 256
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    512 0.0068 0.5995 0.0043 0.9446 63.5%
    1024 0.0063 1.2929 0.0041 2.0000 64.6%
    4096 0.0082 4.0000 0.0044 7.3669 54.3%
    16384 0.0284 4.6178 0.0319 4.1145 112.2%
    65536 0.1377 3.8080 0.1430 3.6674 103.8%

    scopy-stride256 GB/s chart

    scopy-stride256 ms chart

    See also:

    Settings
    Member Visibility

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