Skip to content

solve elementwise_add_grad tenosr copy bug #32095

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

Closed
thisjiang wants to merge 1 commit into from
Closed

solve elementwise_add_grad tenosr copy bug #32095

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
54 changes: 34 additions & 20 deletions paddle/fluid/operators/elementwise/elementwise_add_op.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -112,22 +112,40 @@ elementwise_add_grad(const framework::ExecutionContext& ctx,
const framework::Tensor* out,
const framework::Tensor* dout, framework::Tensor* dx,
framework::Tensor* dy) {
auto* dx_data = dx->mutable_data<T>(ctx.GetPlace());
auto* dy_data = dy->mutable_data<T>(ctx.GetPlace());
auto* dout_data = dout->data<T>();
if (dx_data == dout_data && dy_data != dout_data) {
VLOG(4) << "Special case when dx_data is the same as dout_data, "
"only need copy dout to dy";
framework::TensorCopy(
*dout, ctx.GetPlace(),
ctx.template device_context<platform::DeviceContext>(), dy);
} else if (dx_data != dout_data && dy_data == dout_data) {
VLOG(4) << "Special case when dy_data is the same as dout_data, "
"only need copy dout to dx";
framework::TensorCopy(
*dout, ctx.GetPlace(),
ctx.template device_context<platform::DeviceContext>(), dx);
} else if (dx_data != dout_data && dy_data != dout_data) {
bool dims_same = (dx->dims() == dout->dims()) && (dy->dims() == dout->dims());
bool using_tensorcopy = false;
if (dims_same) {
auto* dx_data = dx->mutable_data<T>(ctx.GetPlace());
auto* dy_data = dy->mutable_data<T>(ctx.GetPlace());
auto* dout_data = dout->data<T>();
bool dx_data_same = (dx_data == dout_data);
bool dy_data_same = (dy_data == dout_data);
if (dx_data_same && !dy_data_same) {
using_tensorcopy = true;
VLOG(4) << "Special case when dx_data is the same as dout_data, "
"only need copy dout to dy";
framework::TensorCopy(
*dout, ctx.GetPlace(),
ctx.template device_context<platform::DeviceContext>(), dy);
} else if (dy_data_same && !dx_data_same) {
using_tensorcopy = true;
VLOG(4) << "Special case when dy_data is the same as dout_data, "
"only need copy dout to dx";
framework::TensorCopy(
*dout, ctx.GetPlace(),
ctx.template device_context<platform::DeviceContext>(), dx);
} else if (dy_data_same && dx_data_same) {
using_tensorcopy = true;
VLOG(4) << "Special case when dy_data is the same as dout_data, "
"and dx_data is the same as dout_data, do not need "
"any operator";
} else {
// need copy dout to two tensor: dx and dy
// using SimpleElemwiseAddGradCUDAKernel faster
}
}

if (!using_tensorcopy) {
auto size = x->numel();
int vec_size = max(static_cast<int>(sizeof(float4) / sizeof(T)), 1);
dim3 block_size = dim3(PADDLE_CUDA_THREAD_SIZE, 1);
Expand All @@ -140,10 +158,6 @@ elementwise_add_grad(const framework::ExecutionContext& ctx,
ctx.template device_context<plat::CUDADeviceContext>().stream()>>>(
dout->data<T>(), size, vec_size, dx->mutable_data<T>(ctx.GetPlace()),
dy->mutable_data<T>(ctx.GetPlace()));
} else {
VLOG(4) << "Special case when dy_data is the same as dout_data, "
"and dx_data is the same as dout_data, do not need "
"any operator";
}
}

Expand Down