Skip to content

added malloc pitch on merged pool embedding #123

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: shbiswas/uptream_develop_0815
Choose a base branch
from
Open

added malloc pitch on merged pool embedding #123

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
78 changes: 57 additions & 21 deletions fbgemm_gpu/src/merge_pooled_embedding_ops/merge_pooled_embedding_ops_gpu.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,15 +222,21 @@ void all_to_one(
.output_idx = i,
.transfer_cuda_event =
std::make_unique<at::cuda::CUDAEvent>(cudaEventDisableTiming)});
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
copy_stream));
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dstPitchBytes,//.stride(0) * dst.element_size(),
src.data_ptr(),
srcPitchBytes,//.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
at::cuda::getCurrentCUDAStream(src.get_device())));
two_hop_transfers.back().transfer_cuda_event->record(copy_stream);
is_two_hop_transfer.push_back(true);
} else {
Expand Down Expand Up @@ -279,15 +285,21 @@ void all_to_one(

auto& dst = output_tensors[i];
// on source device, launch memcpy.
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
dstPitchBytes,//.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
srcPitchBytes,//.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
copy_stream));
at::cuda::getCurrentCUDAStream(src.get_device())));
}
}

Expand All @@ -313,11 +325,17 @@ void all_to_one(
const auto output_index = two_hop_transfer.output_idx;
auto& dst = output_tensors.at(output_index);
// on source device, launch memcpy.
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
dstPitchBytes,//dst.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
srcPitchBytes,//src.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
Expand All @@ -333,11 +351,17 @@ void all_to_one(
// single device memcpy, not that src_device == dst_device.
at::cuda::CUDAStream copy_stream =
at::cuda::getCurrentCUDAStream(target_device_index);
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
dstPitchBytes,//dst.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
srcPitchBytes,//src.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
Expand Down Expand Up @@ -456,11 +480,17 @@ Tensor sum_reduce_to_one(

// on source device, launch memcpy.
auto& dst = copied_tensors[i];
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
dstPitchBytes,//dst.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
srcPitchBytes,//src.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
Expand Down Expand Up @@ -538,11 +568,17 @@ Tensor sum_reduce_to_one(
dst_ready.block(copy_stream);

auto& dst = copied_tensors[i];
void* tmp1 = nullptr;
void* tmp2 = nullptr;
size_t srcPitchBytes;
size_t dstPitchBytes;
AT_CUDA_CHECK(cudaMallocPitch(&tmp1, &srcPitchBytes, src.size(1) * src.element_size(), src.size(0)));
AT_CUDA_CHECK(cudaMallocPitch(&tmp2, &dstPitchBytes, dst.size(1) * dst.element_size(), dst.size(0)));
AT_CUDA_CHECK(cudaMemcpy2DAsync(
dst.data_ptr(),
dst.stride(0) * dst.element_size(),
dstPitchBytes,//dst.stride(0) * dst.element_size(),
src.data_ptr(),
src.stride(0) * src.element_size(),
srcPitchBytes,//src.stride(0) * src.element_size(),
src.size(1) * src.element_size(),
src.size(0),
cudaMemcpyDeviceToDevice,
Expand Down Expand Up @@ -741,4 +777,4 @@ TORCH_LIBRARY_FRAGMENT(fbgemm, m) {
"merge_pooled_embeddings", fbgemm_gpu::merge_pooled_embeddings);
DISPATCH_TO_CUDA("all_to_one_device", fbgemm_gpu::all_to_one_device);
DISPATCH_TO_CUDA("sum_reduce_to_one", fbgemm_gpu::sum_reduce_to_one_device);
}
}