Skip to content

Adjusting location of the fence #329

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
gilbertlee-amd merged 1 commit into from
Jun 29, 2026
Merged

Adjusting location of the fence #329

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
12 changes: 6 additions & 6 deletions src/header/TransferBench.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5037,9 +5037,6 @@ static bool IsConfiguredGid(union ibv_gid const& gid)
}
}

// Drain every thread's writes out to system scope
__threadfence_system();

// Wait for all threads to finish this subiteration
if (seType == 1) {
// For warp-level, sync within warp only
Expand All @@ -5058,6 +5055,9 @@ static bool IsConfiguredGid(union ibv_gid const& gid)
}

if (shouldRecordTiming) {
// Previous sync ensures data from all threads in this subexecutor has landed in cache
// so only one thread needs to do system level threadfence.
__threadfence_system();
p.stopCycle = GetTimestamp();
p.startCycle = startCycle;
GetHwId(p.hwId);
Expand Down Expand Up @@ -5218,9 +5218,6 @@ static bool IsConfiguredGid(union ibv_gid const& gid)
}
}

// Drain every thread's writes out to system scope
__threadfence_system();

// Wait for all threads to finish this subiteration
if (seType == 1) {
// For warp-level, sync within warp only
Expand All @@ -5239,6 +5236,9 @@ static bool IsConfiguredGid(union ibv_gid const& gid)
}

if (shouldRecordTiming) {
// Previous sync ensures data from all threads in this subexecutor has landed in cache
// so only one thread needs to do system level threadfence.
__threadfence_system();
p.stopCycle = GetTimestamp();
p.startCycle = startCycle;
GetHwId(p.hwId);
Expand Down
Loading