NASP/cuda-samples
85
0
Fork 0
mirror of https://github.com/NVIDIA/cuda-samples.git synced 2026-10-26 06:38:16 +08:00
Code Issues Packages Projects Releases Wiki Activity

Bug 5196362, 5184356, 5212196, 5214258 and 5214259: Update sameples for CUDA13.0 API changes

Browse Source
This commit is contained in:
shawnz
2025-04-10 18:27:18 +08:00
parent 278f4adbd2
commit 02fdb070ad
27 changed files with 181 additions and 116 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
Samples/4_CUDA_Libraries/simpleCUBLASXT/simpleCUBLASXT.cpp
View File

@@ -84,13 +84,17 @@ void findMultipleBestGPUs(int &num_of_devices, int *device_ids)
cudaDeviceProp deviceProp;
int devices_prohibited = 0;
int computeMode;
int clockRate;
while (current_device < device_count) {
cudaGetDeviceProperties(&deviceProp, current_device);
checkCudaErrors(cudaDeviceGetAttribute(&computeMode, cudaDevAttrComputeMode, current_device));
checkCudaErrors(cudaDeviceGetAttribute(&clockRate, cudaDevAttrClockRate, current_device));
// If this GPU is not running on Compute Mode prohibited,
// then we can add it to the list
int sm_per_multiproc;
if (deviceProp.computeMode != cudaComputeModeProhibited) {
if (computeMode != cudaComputeModeProhibited) {
if (deviceProp.major == 9999 && deviceProp.minor == 9999) {
sm_per_multiproc = 1;
}
@@ -99,7 +103,7 @@ void findMultipleBestGPUs(int &num_of_devices, int *device_ids)
}
gpu_stats[current_device].compute_perf =
(uint64_t)deviceProp.multiProcessorCount * sm_per_multiproc * deviceProp.clockRate;
(uint64_t)deviceProp.multiProcessorCount * sm_per_multiproc * clockRate;
gpu_stats[current_device].device_id = current_device;
}
else {