Skip to content

[heterps]fix SerializeSelectedRows #36543

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
merged 5 commits into from
Oct 20, 2021
Merged

[heterps]fix SerializeSelectedRows #36543

Changes from 4 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
1f5b08a
bug fix for DeserializeSelectedRows. test=develop
zmxdream Oct 18, 2021
67a4bab
Merge branch 'PaddlePaddle:develop' into bug_fix_DeserializeSelectedRows
zmxdream Oct 18, 2021
d778016
fix bug for SerializeSelectedRows. test=develop
zmxdream Oct 19, 2021
7a3f3d9
merge develop. test=develop
zmxdream Oct 19, 2021
0d980dc
update. test=develop
zmxdream Oct 19, 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
21 changes: 8 additions & 13 deletions paddle/fluid/distributed/service/brpc_utils.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,23 +138,14 @@ void SerializeSelectedRows(framework::Variable* var,
var_data->clear();
var_data->resize(rows->size() * sizeof(int64_t));
char* data_ptr = const_cast<char*>(var_data->data());

if (platform::is_cpu_place(tensor->place())) {
memcpy(data_ptr, &(*rows)[0], rows->size() * sizeof(int64_t));
} else {
#ifdef PADDLE_WITH_CUDA
auto stream =
reinterpret_cast<const platform::CUDADeviceContext&>(ctx).stream();
memory::Copy(platform::CPUPlace(), data_ptr,
BOOST_GET_CONST(platform::CUDAPlace, tensor->place()),
&(*rows)[0], rows->size() * sizeof(int64_t), stream);
#endif
}
// the type of rows is framework::Vector, get data by []
// will trigger data copy operation, if the data is in GPU
// it will be copied to CPU
memcpy(data_ptr, &((*rows)[0]), rows->size() * sizeof(int64_t));
var_msg->set_data_type(static_cast<VarMsg::Type>(tensor->type()));
for (auto& dim : framework::vectorize(tensor->dims())) {
var_msg->add_dims(dim);
}

// IO Buffer
if (platform::is_cpu_place(tensor->place())) {
auto data_len = tensor->numel() * framework::SizeOfType(tensor->type());
Expand Down Expand Up @@ -273,6 +264,10 @@ void DeserializeSelectedRows(framework::Variable* var, const VarMsg& msg,
auto* slr = var->GetMutable<framework::SelectedRows>();
framework::Tensor* tensor = slr->mutable_value();
slr->set_height(msg.slr_height());
// now, the rows of SelectedRows is in CPU
// if current device is CPU, it's ok
// if current device is GPU, the row data is in CPU and the tensor data is GPU
// but it doesn't matter because sum op for selected rows will copy rows to GPU
std::vector<int64_t> tmp_rows(msg.dims()[0]);
memcpy(tmp_rows.data(), msg.data().data(), msg.dims()[0] * sizeof(int64_t));
slr->set_rows(tmp_rows);
Expand Down