nvprof - profiling data are not recorded

Question

I am trying to profile my CUDA program, using the nvprof tool.

Here is my code:

#include <iostream>
#include <math.h>
#include <cuda_profiler_api.h>

// Kernel function to add the elements of two arrays
__global__
void add(int n, float *x, float *y)
{
  int index = blockIdx.x * blockDim.x + threadIdx.x;
  int stride = blockDim.x * gridDim.x;
  for (int i = index; i < n; i += stride)
    y[i] = x[i] + y[i];
}

int main(void)
{
  int N = 1<<10;
  float *x, *y;

  // Allocate Unified Memory – accessible from CPU or GPU
  cudaMallocManaged(&x, N*sizeof(float));
  cudaMallocManaged(&y, N*sizeof(float));

  // initialize x and y arrays on the host
  for (int i = 0; i < N; i++) {
    x[i] = 1.0f;
    y[i] = 2.0f;
  }

  // Run kernel on 1M elements on the GPU
int blockSize = 256;
int numBlocks = (N + blockSize - 1) / blockSize;
add<<<numBlocks, blockSize>>>(N, x, y);

  // Wait for GPU to finish before accessing on host
  cudaDeviceSynchronize();

  // Check for errors (all values should be 3.0f)
  float maxError = 0.0f;
  for (int i = 0; i < N; i++)
    maxError = fmax(maxError, fabs(y[i]-3.0f));
  std::cout << "Max error: " << maxError << std::endl;

  // Free memory
  cudaFree(x);
  cudaFree(y);

  cudaProfilerStop();
  cudaDeviceReset();

  return 0;
}

I compiled it using the command nvcc add.cu -o add_cuda.

I then run it (as root) using nvprof ./add_cuda --unified-memory-profiling off or nvprof and get the following output:

==15318== NVPROF is profiling process 15318, command: ./add_cuda
Max error: 0
==15318== Profiling application: ./add_cuda
==15318== Profiling result:
No kernels were profiled.
No API activities were profiled.
==15318== Warning: Some profiling data are not recorded. Make sure cudaProfilerStop() or cuProfilerStop() is called before application exit to flush profile data.
======== Error: Application received signal 139

I searched online for a solution nvprof not picking up any API calls or kernels, https://devtalk.nvidia.com/default/topic/1010691/visual-profiler/nvprof-error-code-139-but-memcheck-ok/, but really nothing helped.

How can i get nvprof to work?

Thanks!

Fedora 29 64-bit

nvprof: NVIDIA (R) Cuda command line profiler
Copyright (c) 2012 - 2019 NVIDIA Corporation
Release version 10.1.168 (21)

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Wed_Apr_24_19:10:27_PDT_2019
Cuda compilation tools, release 10.1, V10.1.168

nvidia-smi

Mon Jul  1 13:24:54 2019       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 430.26       Driver Version: 430.26       CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 108...  Off  | 00000000:03:00.0  On |                  N/A |
|  0%   37C    P8    20W / 250W |    253MiB / 11175MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      1741      G   /usr/libexec/Xorg                            154MiB |
|    0      2161      G   cinnamon                                      96MiB |
+-----------------------------------------------------------------------------+

Answer 1

I get the same result when using nvprof version 10.1 but it works with 10.0.

Try downloading for instance CUDA 10.0 and use that version of nvprof instead. That works for me even if compiled with nvcc version 10.1.