wgblas
    Preparing search index...

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

    Benchmark results for isamax on Nvidia Geforce Gtx 1650.

    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0221 0.0058 0.0169 0.0076 76.2%
    64 0.0216 0.0118 0.0165 0.0155 76.3%
    128 0.0219 0.0234 0.0165 0.0310 75.5%
    512 0.0220 0.0932 0.0162 0.1266 73.6%
    1024 0.0218 0.1880 0.0168 0.2438 77.1%
    4096 0.0219 0.7496 0.0179 0.9176 81.7%
    16384 0.0222 2.9489 0.0181 3.6184 81.5%
    65536 0.0228 11.5218 0.0181 14.5120 79.4%
    262144 0.0290 36.2077 0.0195 53.8947 67.2%
    1048576 0.0478 87.7910 0.0400 104.8996 83.7%
    4194304 0.1180 142.1798 0.1094 153.3230 92.7%
    16777216 0.4116 163.0249 0.4021 166.8910 97.7%

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

    isamax-default GB/s chart

    isamax-default ms chart

    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 0.0222 0.0058 0.0168 0.0076 75.7%
    64 0.0217 0.0118 0.0123 0.0208 56.6%
    128 0.0215 0.0238 0.0169 0.0302 78.8%
    512 0.0218 0.0940 0.0169 0.1213 77.5%
    1024 0.0217 0.1884 0.0171 0.2399 78.5%
    4096 0.0214 0.7670 0.0181 0.9030 84.9%
    16384 0.0228 2.8704 0.0182 3.6025 79.7%
    65536 0.0284 9.2148 0.0191 13.6990 67.3%
    262144 0.0462 22.7161 0.0421 24.9281 91.1%
    1048576 0.1131 37.0784 0.1061 39.5212 93.8%
    4194304 0.3812 44.0116 0.3755 44.6849 98.5%

    isamax-stride4 GB/s chart

    isamax-stride4 ms chart

    Nvidia Geforce Gtx 1650 — stride = 32
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    32 0.0223 0.0057 0.0169 0.0076 75.5%
    64 0.0214 0.0120 0.0164 0.0156 76.7%
    128 0.0216 0.0237 0.0169 0.0304 78.1%
    512 0.0215 0.0955 0.0168 0.1217 78.4%
    1024 0.0224 0.1829 0.0169 0.2429 75.3%
    4096 0.0216 0.7602 0.0185 0.8866 85.7%
    16384 0.0291 2.2543 0.0210 3.1267 72.1%
    65536 0.0460 5.7047 0.0395 6.6332 86.0%
    262144 0.1147 9.1429 2.3382 0.4485 2038.5%
    1048576 0.3869 10.8409 0.3756 11.1669 97.1%

    isamax-stride32 GB/s chart

    isamax-stride32 ms chart

    Nvidia Geforce Gtx 1650 — stride = 256
    n wgblas ms wgblas GB/s cuBLAS ms cuBLAS GB/s efficiency
    512 0.0224 0.0916 0.0169 0.1214 75.4%
    1024 0.0224 0.1826 0.0169 0.2417 75.5%
    4096 0.0228 0.7176 0.0182 0.9014 79.6%
    16384 0.0345 1.8989 0.0292 2.2456 84.6%
    65536 0.0613 4.2789 0.0509 5.1506 83.1%

    isamax-stride256 GB/s chart

    isamax-stride256 ms chart

    See also:

    Settings
    Member Visibility

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