Skip to content
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

Jump to bottom

[CUDA] Support memory reuse for dynamic shared memory #9341

Merged
merged 7 commits into from
Oct 31, 2021
Merged

[CUDA] Support memory reuse for dynamic shared memory #9341

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
267982f
reuse shared dyn
jinhongyii Oct 21, 2021
e64e4b1
minor
jinhongyii Oct 21, 2021
430ae5d
format
jinhongyii Oct 21, 2021
073970b
format
jinhongyii Oct 21, 2021
96a9974
address comment and fix
jinhongyii Oct 25, 2021
07464d0
address comment
jinhongyii Oct 28, 2021
eed8d5e
address comment
jinhongyii Oct 30, 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
3 changes: 2 additions & 1 deletion src/driver/driver_api.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -590,7 +590,6 @@ transform::Sequential MixedModulePassManager(IRModule mixed_mod, Target target)
mixed_pass_list.push_back(BindTarget(target));

mixed_pass_list.push_back(tir::transform::VerifyMemory());
mixed_pass_list.push_back(tir::transform::MergeDynamicSharedMemoryAllocations());

if (ShouldAnnotateEntryFunc(target, mixed_mod)) {
mixed_pass_list.push_back(AnnotateEntryFunc(true));
Expand All @@ -603,6 +602,8 @@ transform::Sequential MixedModulePassManager(IRModule mixed_mod, Target target)
}

mixed_pass_list.push_back(tir::transform::ThreadSync("shared"));
mixed_pass_list.push_back(tir::transform::ThreadSync("shared.dyn"));
mixed_pass_list.push_back(tir::transform::MergeDynamicSharedMemoryAllocations());
mixed_pass_list.push_back(tir::transform::ThreadSync("warp"));
mixed_pass_list.push_back(tir::transform::InferFragment());
mixed_pass_list.push_back(tir::transform::LowerThreadAllreduce());
Expand Down
2 changes: 1 addition & 1 deletion src/target/source/codegen_cuda.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -525,7 +525,7 @@ void CodeGenCUDA::PrintStorageSync(const CallNode* op) {
const std::string& sync = op->args[0].as<StringImmNode>()->value;
if (sync == "warp") {
// DO nothing.
} else if (sync == "shared") {
} else if (sync == "shared" || sync == "shared.dyn") {
this->PrintIndent();
this->stream << "__syncthreads();\n";
} else if (sync == "global") {
Expand Down
Loading