CODEWITHSUNDEEP
cudagpuparallel-processing
cnhk
cnhk

Reputation: 1305

How to measure the execution time of every block when using CUDA?

clock() is not accurate enough.

Upvotes: 4

Views: 6553

Answers (3)

ticy
ticy

Reputation: 51

I think long long int clock64() is what you are looking for?

See Cuda Programming Guide, C Language Extensions, B. 11.

Upvotes: 1

Milinda Pathirage
Milinda Pathirage

Reputation: 60

How about using clock() function in every CUDA thread to calculate start and end times. And store it in a array such a way that you can figure out which thread start/stop at which time based on array indices like following:

__global__ void kclock(unsigned int *ts) {
    unsigned int start_time = 0, stop_time = 0;

    start_time = clock();

    // Code we need to measure should go here.

    stop_time = clock();

    ts[(blockIdx.x * blockDim.x + threadIdx.x) * 2] = start_time;
    ts[(blockIdx.x * blockDim.x + threadIdx.x) * 2 + 1] = stop_time; 
}

Then use this array to figure out minimal start time and maximum stop time for block you are considering. For example you can calculate range of indices of time array which corresponds to the (0, 0) block in CUDA and use min/max to calculate the execution time.

Upvotes: 2

KoppeKTop
KoppeKTop

Reputation: 648

Use CUDA events for measure time of kernels or CUDA operations (memcpy etc):

// Prepare
cudaEvent_t start, stop;
cudaEventCreate(&start);
cudaEventCreate(&stop);
// Start record
cudaEventRecord(start, 0);
// Do something on GPU
MyKernel<<<dimGrid, dimBlock>>>(input_data, output_data);
// Stop event
cudaEventRecord(stop, 0);
cudaEventSynchronize(stop);
float elapsedTime;
cudaEventElapsedTime(&elapsedTime, start, stop); // that's our time!
// Clean up:
cudaEventDestroy(start);
cudaEventDestroy(stop);

See CUDA Programming Guide, section 3.2.7.6

Upvotes: 13

Related Questions