Skip to content

Commit

Permalink
Refactor away from google bench
Browse files Browse the repository at this point in the history
  • Loading branch information
@johnbowen42
johnbowen42 committed Apr 15, 2024
1 parent 633f6c0 commit dad5812
Showing 1 changed file with 11 additions and 10 deletions.
21 changes: 11 additions & 10 deletions benchmark/benchmark-atomic.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -101,27 +101,28 @@ void TimeAtomicOp(const std::string& test_name, int num_iterations = 2, int arra
}

int main () {
// CUDA benchmarks
TimeAtomicOp<RAJA::cuda_exec<32>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 32, RAJA builtin atomic", 4);
//TimeAtomicOp<RAJA::cuda_exec<32>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 32, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<32>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 32, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<64>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 64, RAJA builtin atomic", 4);
//TimeAtomicOp<RAJA::cuda_exec<64>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 64, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<64>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 64, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<128>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 128, RAJA builtin atomic", 4);
//TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 128, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 128, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<256>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 256, RAJA builtin atomic", 4);
//TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 256, DESUL atomic", 4);
TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 256, DESUL atomic", 4);

TimeAtomicOp<RAJA::cuda_exec<128>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 128, RAJA builtin atomic", 2, 10);
//TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 128, DESUL atomic", 2, 10);
TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 128, DESUL atomic", 2, 10);
TimeAtomicOp<RAJA::cuda_exec<256>, int, CudaAtomicAdd<int>, true>("Benchmark array contention. CUDA Block size 256, RAJA builtin atomic", 2, 10);
//TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 256, DESUL atomic", 2, 10);
TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>, true>("Benchmark array contention. CUDA Block size 256, DESUL atomic", 2, 10);

TimeAtomicOp<RAJA::cuda_exec<128>, double, CudaAtomicAdd<double>>("CUDA Block size 128, RAJA builtin atomic");
//TimeAtomicOp<RAJA::cuda_exec<128>, double, atomicWrapperDesul<double, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>>("CUDA Block size 128, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<128>, double, atomicWrapperDesul<double, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>>("CUDA Block size 128, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<256>, double, CudaAtomicAdd<double>>("CUDA Block size 256, RAJA builtin atomic");
//TimeAtomicOp<RAJA::cuda_exec<256>, double, atomicWrapperDesul<double, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>>("CUDA Block size 256, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<256>, double, atomicWrapperDesul<double, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_add>>("CUDA Block size 256, DESUL atomic");

TimeAtomicOp<RAJA::cuda_exec<128>, int, CudaAtomicMax<int>>("CUDA Block size 128, RAJA builtin atomic");
//TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_max>>("CUDA Block size 128, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<128>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_max>>("CUDA Block size 128, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<256>, int, CudaAtomicMax<int>>("CUDA Block size 256, RAJA builtin atomic");
//TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_max>>("CUDA Block size 256, DESUL atomic");
TimeAtomicOp<RAJA::cuda_exec<256>, int, atomicWrapperDesul<int, RAJA::policy::cuda::cuda_atomic, desul::atomic_fetch_max>>("CUDA Block size 256, DESUL atomic");
}

0 comments on commit dad5812

Please sign in to comment.