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 0.0074 0.0694 0.0143 0.0357 194.4%
    64 0.0075 0.1359 0.0143 0.0714 190.3%
    128 0.0075 0.2735 0.0143 0.1429 191.4%
    512 0.0075 1.0917 0.0143 0.5714 191.1%
    1024 0.0075 2.1974 0.0143 1.1429 192.3%
    4096 0.0075 8.7149 0.0143 4.5714 190.6%
    16384 0.0082 32.0000 0.0155 16.9081 189.3%
    65536 0.0103 102.0810 0.0184 56.8889 179.4%
    262144 0.0307 136.5333 0.0420 99.7504 136.9%
    1048576 0.1031 162.6708 0.1188 141.2794 115.1%
    4194304 0.3930 170.7500 0.3958 169.5627 100.7%
    16777216 1.5646 171.5690 1.5605 172.0175 99.7%

    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 0.0029 0.1739 47.9%
    64 0.0061 0.1667 0.0029 0.3575 46.6%
    128 0.0061 0.3333 0.0030 0.6737 49.5%
    512 0.0061 1.3333 0.0034 2.4151 55.2%
    1024 0.0063 2.5990 0.0032 5.1200 50.8%
    4096 0.0070 9.3730 0.0037 17.6552 53.1%
    16384 0.0088 29.9525 0.0053 49.4985 60.5%
    65536 0.0287 36.5714 0.0270 38.8246 94.2%
    262144 0.1023 40.9920 0.1004 41.7626 98.2%
    1048576 0.3973 42.2268 0.3946 42.5213 99.3%
    4194304 1.5454 43.4238 1.5709 42.7192 101.6%

    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 0.0030 0.1693 49.2%
    64 0.0061 0.1667 0.0034 0.3048 54.7%
    128 0.0063 0.3265 0.0035 0.5899 55.4%
    512 0.0066 1.2397 0.0039 2.0984 59.1%
    1024 0.0069 2.3814 0.0039 4.1967 56.7%
    4096 0.0099 6.6278 0.0058 11.2527 58.9%
    16384 0.0356 7.3537 0.0344 7.6134 96.6%
    65536 0.1308 8.0157 0.1345 7.7963 102.8%
    262144 0.5181 8.0961 0.5365 7.8180 103.6%
    1048576 2.1463 7.8168 4.4330 3.7846 206.5%

    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 0.0067 1.2278 0.0041 2.0000 61.4%
    1024 0.0071 2.3167 0.0041 4.0315 57.5%
    4096 0.0114 5.7690 0.0082 8.0000 72.1%
    16384 0.0637 4.1155 0.0676 3.8788 106.1%
    65536 0.2909 3.6046 0.2924 3.5865 100.5%

    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