Skip to content

feat(meta_schedule): expand CUDA unroll steps for SM70 optimization #18927

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

Open
jianhua1724 wants to merge 1 commit into
base: main
Choose a base branch
from
+1 −1
Open

feat(meta_schedule): expand CUDA unroll steps for SM70 optimization #18927

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
2 changes: 1 addition & 1 deletion src/s_tir/meta_schedule/schedule_rule/schedule_rule.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -166,7 +166,7 @@ ffi::Array<ScheduleRule> ScheduleRule::DefaultCUDA() {
ScheduleRule::ParallelizeVectorizeUnroll(
/*max_jobs_per_core=*/-1,
/*max_vectorize_extent=*/-1,
/*unroll_max_steps=*/ffi::Array<Integer>{0, 16, 64, 512, 1024},
/*unroll_max_steps=*/ffi::Array<Integer>{0, 16, 32, 64, 128, 256, 512, 1024},
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist bot Mar 25, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

This change expands the unroll search space for all CUDA targets, which can increase auto-tuning time for architectures where these new unroll steps are not beneficial.

To make this optimization more targeted to SM70+ GPUs as intended, I suggest making the choice of unroll steps conditional on the target's compute capability. This avoids slowing down tuning on older architectures. This can be done cleanly using an immediately-invoked lambda expression.

The condition can be sm_version >= 70 if the optimization is beneficial for SM70 and newer, or sm_version == 70 if it's specific to V100.

          /*unroll_max_steps=*/[]() -> ffi::Array<Integer> {
            auto target = tvm::Target::Current(true);
            if (target.defined() && target->kind->name == "cuda") {
              if (const auto* sm_ptr = target->GetAttr<Integer>("sm")) {
                if (sm_ptr->value() >= 70) {
                  return {0, 16, 32, 64, 128, 256, 512, 1024};
                }
              }
            }
            return {0, 16, 64, 512, 1024};
          }(),

/*unroll_explicit=*/true),
ScheduleRule::AutoBind(
/*max_threadblocks=*/256,
Expand Down