Skip to content

Fix CUDA Reduction kernel for ArgMax/ArgMix for when reduction dim=1 #6490

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

Merged
ashbhandare merged 3 commits into from
Feb 2, 2021
Merged

Fix CUDA Reduction kernel for ArgMax/ArgMix for when reduction dim=1 #6490

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
091771a
Fix for when reduction dim=1
ashbhandare Jan 28, 2021
3bcac4a
Disable test for AMD GPUs
ashbhandare Jan 29, 2021
eff7b47
Specify Async
ashbhandare Feb 1, 2021
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
40 changes: 23 additions & 17 deletions onnxruntime/core/providers/cuda/reduction/reduction_ops.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,7 +154,7 @@ Status ReduceKernel<allow_multi_axes>::ReduceKernelShared(
m, n, false);
}
case ApplicableMatrixReduction::Columns:
// don't call reduce_matrix_columns() since it will reset initial output data
// don't call reduce_matrix_columns() since it will reset initial output data
default:
break;
}
Expand Down Expand Up @@ -600,24 +600,30 @@ Status ReduceComputeCore(CUDAExecutionProvider& cuda_ep, const Tensor& input, Pr
}
}
} else { // For ArgMax & ArgMin ops, use the indicies as the output with int64 type
if (temp_X) {
auto temp_output = cuda_ep.GetScratchBuffer<float>(output_count);
CUDNN_RETURN_IF_ERROR(cudnnReduceTensor(
cuda_ep.PerThreadCudnnHandle(), reduce_desc, indices_cuda.get(), indices_bytes,
workspace_cuda.get(), workspace_bytes,
&one, input_tensor, temp_X.get(),
&zero, output_tensor, temp_output.get()));
// cudnnReduceTensor has issue if input and output has same size, which will happen if the axis to be reduced has dim value of 1.
// the output is zeros of the output size
if (input_count == output_count) {
CUDA_RETURN_IF_ERROR(cudaMemset(output.template MutableData<int64_t>(), static_cast<int64_t>(0), output_count * sizeof(int64_t)));
} else {
auto temp_output = cuda_ep.GetScratchBuffer<CudaT>(output_count);
CUDNN_RETURN_IF_ERROR(cudnnReduceTensor(
cuda_ep.PerThreadCudnnHandle(), reduce_desc, indices_cuda.get(), indices_bytes,
workspace_cuda.get(), workspace_bytes,
&one, input_tensor, reinterpret_cast<const CudaT*>(input.template Data<T>()),
&zero, output_tensor, temp_output.get()));
}
if (temp_X) {
auto temp_output = cuda_ep.GetScratchBuffer<float>(output_count);
CUDNN_RETURN_IF_ERROR(cudnnReduceTensor(
cuda_ep.PerThreadCudnnHandle(), reduce_desc, indices_cuda.get(), indices_bytes,
workspace_cuda.get(), workspace_bytes,
&one, input_tensor, temp_X.get(),
&zero, output_tensor, temp_output.get()));
} else {
auto temp_output = cuda_ep.GetScratchBuffer<CudaT>(output_count);
CUDNN_RETURN_IF_ERROR(cudnnReduceTensor(
cuda_ep.PerThreadCudnnHandle(), reduce_desc, indices_cuda.get(), indices_bytes,
workspace_cuda.get(), workspace_bytes,
&one, input_tensor, reinterpret_cast<const CudaT*>(input.template Data<T>()),
&zero, output_tensor, temp_output.get()));
}

// CUDA reduction index is uint32_t for now, cast it to int64_t according to ONNX spec
Impl_Cast<uint32_t, int64_t>(reinterpret_cast<uint32_t*>(indices_cuda.get()), output.template MutableData<int64_t>(), output_count);
// CUDA reduction index is uint32_t for now, cast it to int64_t according to ONNX spec
Impl_Cast<uint32_t, int64_t>(reinterpret_cast<uint32_t*>(indices_cuda.get()), output.template MutableData<int64_t>(), output_count);
}
}

if (calculate_log) {
Expand Down
12 changes: 12 additions & 0 deletions onnxruntime/test/providers/cpu/reduction/reduction_ops_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2060,6 +2060,18 @@ TEST(ReductionOpTest, ArgMax2D_select_last) {
test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider});
}

TEST(ReductionOpTest, ArgMax2D_dim1) {
OpTester test("ArgMax", 11);
test.AddAttribute("axis", (int64_t)1);
test.AddInput<float>("data", {3, 1},
{1.0f,
6.0f,
9.0f});
test.AddOutput<int64_t>("reduced", {3, 1},
{0, 0, 0});
test.Run(OpTester::ExpectResult::kExpectSuccess, "", {kTensorrtExecutionProvider});
}

TEST(ReductionOpTest, ArgMin) {
OpTester test("ArgMin");
test.AddAttribute("axis", (int64_t)0);
Expand Down