Skip to content

Fix codegen/cuda gcc-5.4 compilation issues #43223

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
malfet wants to merge 1 commit into from
Closed

Fix codegen/cuda gcc-5.4 compilation issues #43223

Show file tree
Hide file tree
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
6 changes: 4 additions & 2 deletions torch/csrc/jit/codegen/cuda/lower_thread_predicate.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,8 +22,10 @@ namespace fuser {
*/
class TORCH_CUDA_API ThreadPredicateMap {
public:
using SourceMapType =
std::unordered_map<ParallelType, std::unordered_set<const TensorView*>>;
using SourceMapType = std::unordered_map<
ParallelType,
std::unordered_set<const TensorView*>,
TypeHash>;
using MapType = std::unordered_map<
const TensorView*,
std::pair<ir_utils::ParallelTypeBitmap, SourceMapType>>;
Expand Down
16 changes: 8 additions & 8 deletions torch/csrc/jit/codegen/cuda/lower_utils.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -479,13 +479,13 @@ bool isUnrolledFor(const Expr* expr) {
ParallelType::Unroll;
}

const std::unordered_map<ParallelType, int> ParallelTypeBitmap::pt_to_offset_{
{ParallelType::BIDx, 0},
{ParallelType::BIDy, 1},
{ParallelType::BIDz, 2},
{ParallelType::TIDx, 3},
{ParallelType::TIDy, 4},
{ParallelType::TIDz, 5}};
const std::unordered_map<ParallelType, int, TypeHash>
ParallelTypeBitmap::pt_to_offset_{{ParallelType::BIDx, 0},
{ParallelType::BIDy, 1},
{ParallelType::BIDz, 2},
{ParallelType::TIDx, 3},
{ParallelType::TIDy, 4},
{ParallelType::TIDz, 5}};

const std::unordered_map<int, ParallelType> ParallelTypeBitmap::offset_to_pt_ =
{{0, ParallelType::BIDx},
Expand Down Expand Up @@ -554,7 +554,7 @@ bool ParallelTypeBitmap::operator[](size_t pos) const {
std::map<ParallelType, bool> ParallelTypeBitmap::getMap() const {
std::map<ParallelType, bool> map;
for (const auto& pt_offset : pt_to_offset_) {
map.emplace(std::make_pair(pt_offset.first, bitset_[pt_offset.second]));
map.emplace(pt_offset.first, bitset_[pt_offset.second]);
}
return map;
}
Expand Down
2 changes: 1 addition & 1 deletion torch/csrc/jit/codegen/cuda/lower_utils.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,7 +125,7 @@ class ParallelTypeBitmap {
private:
ParallelTypeBitmap(const std::bitset<num_p_type>& bs) : bitset_(bs) {}
std::bitset<num_p_type> bitset_;
const static std::unordered_map<ParallelType, int> pt_to_offset_;
const static std::unordered_map<ParallelType, int, TypeHash> pt_to_offset_;
const static std::unordered_map<int, ParallelType> offset_to_pt_;
};

Expand Down
2 changes: 1 addition & 1 deletion torch/csrc/jit/codegen/cuda/manager.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,7 +91,7 @@ class CudaFusionManager {
int32_t kernel_id = getNextUniqueID();
graph_cache_ids_[repr] = kernel_id;
TORCH_CHECK(
graph_cache_.insert({kernel_id, std::make_unique<GraphCache>(graph)})
graph_cache_.emplace(kernel_id, std::make_unique<GraphCache>(graph))
.second);
}
return graph_cache_ids_[repr];
Expand Down