Skip to content

Commit

Permalink
Bug Fix - Fix stderr message in gpu-copy benchmark (#673)
Browse files Browse the repository at this point in the history 
Fix ordering of args in err messages.
  • Loading branch information
@dpower4
dpower4 authored Nov 28, 2024
1 parent 96f5cce commit 4e6935a
Showing 1 changed file with 21 additions and 21 deletions.
42 changes: 21 additions & 21 deletions superbench/benchmarks/micro_benchmarks/gpu_copy_performance/gpu_copy.cu
View file
Original file line number Diff line number Diff line change
Expand Up @@ -913,7 +913,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank

for (int rank = 0; rank < gpu_count; rank++) {
if (SetGpu(rank)) {
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", rank, cuda_err);
return -1;
}

Expand All @@ -924,7 +924,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
cuda_err = GpuMallocDataBuf(&(src_buffers_gpu[rank]), opts.size);
#endif
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMalloc for src_buffers_gpu[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaMalloc for src_buffers_gpu[%d] error: %d\n", rank, cuda_err);
return -1;
}
if (opts.check_data) {
Expand All @@ -933,7 +933,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
}
cuda_err = cudaMemcpy(src_buffers_gpu[rank], data_buffer_cpu, opts.size, cudaMemcpyDefault);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMemcpy to src_buffers_gpu[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaMemcpy to src_buffers_gpu[%d] error: %d\n", rank, cuda_err);
return -1;
}
}
Expand All @@ -945,7 +945,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
cuda_err = GpuMallocDataBuf(&(dst_buffers_gpu[rank]), opts.size);
#endif
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMalloc for dst_buffers_gpu[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaMalloc for dst_buffers_gpu[%d] error: %d\n", rank, cuda_err);
return -1;
}

Expand All @@ -959,33 +959,33 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
// Prepare events
cuda_err = cudaEventCreate(&(start_events[rank]));
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaEventCreate for start_events[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaEventCreate for start_events[%d] error: %d\n", rank, cuda_err);
return -1;
}
cuda_err = cudaEventCreate(&(stop_events[rank]));
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaEventCreate for stop_events[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaEventCreate for stop_events[%d] error: %d\n", rank, cuda_err);
return -1;
}
}

// Prepare kernel arguments
for (int rank = 0; rank < gpu_count; rank++) {
if (SetGpu(rank)) {
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", rank, cuda_err);
return -1;
}

// Prepare destination buffer args
cuda_err = cudaMalloc(&(dst_buffer_gpu_args[rank]), sizeof(uint8_t *) * gpu_count);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMalloc for dst_buffer_gpu_args[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaMalloc for dst_buffer_gpu_args[%d] error: %d\n", rank, cuda_err);
return -1;
}
cuda_err = cudaMemcpy(dst_buffer_gpu_args[rank], dst_buffers_gpu.data(), sizeof(uint8_t *) * gpu_count,
cudaMemcpyDefault);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMemcpy to dst_buffer_gpu_args[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaMemcpy to dst_buffer_gpu_args[%d] error: %d\n", rank, cuda_err);
return -1;
}
}
Expand All @@ -998,15 +998,15 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
}

if (SetGpu(rank)) {
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::SetGpu for rank %d error: %d\n", rank, cuda_err);
return -1;
}

if (i == opts.num_warm_up) {
cuda_err = cudaEventRecord(start_events[rank], streams[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaEventRecord for start_events[%d] error: %d\n", cuda_err,
rank);
fprintf(stderr, "RunAllToAllBench::cudaEventRecord for start_events[%d] error: %d\n", rank,
cuda_err);
return -1;
}
}
Expand All @@ -1017,8 +1017,8 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
if (i == opts.num_warm_up + opts.num_loops - 1) {
cuda_err = cudaEventRecord(stop_events[rank], streams[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaEventRecord for stop_events[%d] error: %d\n", cuda_err,
rank);
fprintf(stderr, "RunAllToAllBench::cudaEventRecord for stop_events[%d] error: %d\n", rank,
cuda_err);
return -1;
}
}
Expand All @@ -1030,7 +1030,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
}
cuda_err = cudaStreamSynchronize(streams[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaStreamSynchronize streams[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaStreamSynchronize streams[%d] error: %d\n", rank, cuda_err);
return -1;
}
}
Expand All @@ -1045,7 +1045,7 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
float time_in_ms = 0;
cuda_err = cudaEventElapsedTime(&time_in_ms, start_events[rank], stop_events[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaEventElapsedTime for rank %d error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaEventElapsedTime for rank %d error: %d\n", rank, cuda_err);
return -1;
}
double bw = opts.size * (gpu_count - 1) * opts.num_loops / gpu_count / time_in_ms / 1e6;
Expand All @@ -1072,8 +1072,8 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
}
cuda_err = cudaMemcpy(data_buffer_cpu, dst_buffers_gpu[curr_dst_rank], opts.size, cudaMemcpyDefault);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaMemcpy from dst_buffers_gpu[%d] error: %d\n", cuda_err,
curr_dst_rank);
fprintf(stderr, "RunAllToAllBench::cudaMemcpy from dst_buffers_gpu[%d] error: %d\n", curr_dst_rank,
cuda_err);
return -1;
}
for (uint64_t i = 0; i < opts.size / sizeof(uint64_t); i++) {
Expand All @@ -1100,17 +1100,17 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
for (int rank = 0; rank < gpu_count; rank++) {
cuda_err = cudaFree(src_buffers_gpu[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaFree for src_buffers_gpu[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaFree for src_buffers_gpu[%d] error: %d\n", rank, cuda_err);
return -1;
}
cuda_err = cudaFree(dst_buffers_gpu[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaFree for dst_buffers_gpu[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaFree for dst_buffers_gpu[%d] error: %d\n", rank, cuda_err);
return -1;
}
cuda_err = cudaFree(dst_buffer_gpu_args[rank]);
if (cuda_err != cudaSuccess) {
fprintf(stderr, "RunAllToAllBench::cudaFree for dst_buffer_gpu_args[%d] error: %d\n", cuda_err, rank);
fprintf(stderr, "RunAllToAllBench::cudaFree for dst_buffer_gpu_args[%d] error: %d\n", rank, cuda_err);
return -1;
}
}
Expand Down

0 comments on commit 4e6935a

Please sign in to comment.