在Windows下运行nvprof --metrics命令会出现错误:CUDA分析错误

问题描述 投票:1回答:1

在Windows下运行nvprof --metrics命令会出现错误:

==6580== NVPROF is profiling process 6580, command: Project1.exe
==6580== Error: Internal profiling error 4292:1.
======== Error: CUDA profiling error.

error1

如果仅使用nvprof命令,将不会报告错误:

F:\vstest\Project1\x64\Release>nvprof Project1.exe
==384== NVPROF is profiling process 384, command: Project1.exe
sumMatrixOnGPU2D <<<(512,512), (32,32)>>> elapsed 22 ms
==384== Profiling application: Project1.exe
==384== Profiling result:
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
 GPU activities:   61.28%  538.11ms         2  269.06ms  260.98ms  277.13ms  [CUDA memcpy HtoD]
                   36.29%  318.68ms         1  318.68ms  318.68ms  318.68ms  [CUDA memcpy DtoH]
                    2.43%  21.364ms         1  21.364ms  21.364ms  21.364ms  sumMatrixOnGPU2D(float*, float*, float*, int, int)
      API calls:   56.77%  1.29771s         3  432.57ms  47.895ms  1.19911s  cudaMalloc
                   37.53%  857.94ms         3  285.98ms  261.20ms  319.19ms  cudaMemcpy
                    2.56%  58.617ms         1  58.617ms  58.617ms  58.617ms  cudaDeviceReset
                    2.13%  48.594ms         3  16.198ms  14.312ms  17.671ms  cudaFree
                    0.95%  21.732ms         2  10.866ms  275.60us  21.456ms  cudaDeviceSynchronize
                    0.02%  512.70us         1  512.70us  512.70us  512.70us  cudaLaunchKernel
                    0.02%  359.30us        96  3.7420us     100ns  204.60us  cuDeviceGetAttribute
                    0.02%  347.80us         1  347.80us  347.80us  347.80us  cudaGetDeviceProperties
                    0.01%  180.60us         1  180.60us  180.60us  180.60us  cuDeviceGetPCIBusId
                    0.00%  32.100us         1  32.100us  32.100us  32.100us  cuDeviceTotalMem
                    0.00%  13.400us         1  13.400us  13.400us  13.400us  cudaSetDevice
                    0.00%  4.0000us         3  1.3330us     200ns  3.5000us  cuDeviceGetCount
                    0.00%  3.9000us         1  3.9000us  3.9000us  3.9000us  cudaGetLastError
                    0.00%  1.1000us         2     550ns     200ns     900ns  cuDeviceGet
                    0.00%  1.0000us         1  1.0000us  1.0000us  1.0000us  cuDeviceGetName
                    0.00%     300ns         1     300ns     300ns     300ns  cuDeviceGetUuid
                    0.00%     300ns         1     300ns     300ns     300ns  cuDeviceGetLuid

我想问什么问题,如何使用命令nvprof --metrics

cuda metrics nvprof
1个回答
0
投票

我找到了答案。

我正在添加对我有用的解决方案,以供其他人参考。

您需要打开NVIDIA控制面板(在桌面上单击鼠标右键,然后选择它)桌面(从顶部菜单)->启用开发人员设置。然后,选择开发人员(从侧面树起)->管理GPU性能计数器->允许所有用户访问GPU性能计数器。

“现在很难找到。
© www.soinside.com 2019 - 2024. All rights reserved.