Skip to content

[SYCL] Optimize OOO profiling tag when profiling is enabled #16116

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

Draft
wants to merge 3 commits into
base: sycl
Choose a base branch
from
Draft

[SYCL] Optimize OOO profiling tag when profiling is enabled #16116

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
89f1a3f
[SYCL] Optimize OOO profiling tag when profiling is enabled
steffenlarsen Nov 19, 2024
4512bab
Merge remote-tracking branch 'intel/sycl' into steffen/optimize_profi…
steffenlarsen Sep 29, 2025
2413018
Merge remote-tracking branch 'intel/sycl' into steffen/optimize_profi…
steffenlarsen Oct 22, 2025
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
8 changes: 8 additions & 0 deletions sycl/include/sycl/ext/oneapi/experimental/profiling_tag.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,14 @@ inline event submit_profiling_tag(queue &Queue,
const sycl::detail::code_location &CodeLoc =
sycl::detail::code_location::current()) {
if (Queue.get_device().has(aspect::ext_oneapi_queue_profiling_tag)) {
// If the queue is out-of-order and profiling is enabled, the implementation
// can save some operations by just using the required barrier event
// directly.
if (!Queue.is_in_order() &&
Queue.has_property<sycl::property::queue::enable_profiling>())
return Queue.ext_oneapi_submit_barrier();

// Otherwise, we use the internal implementation of the profiling tag.
return Queue.submit(
[=](handler &CGH) {
sycl::detail::HandlerAccess::internalProfilingTagImpl(CGH);
Expand Down
7 changes: 6 additions & 1 deletion sycl/unittests/Extensions/ProfilingTag.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -129,6 +129,8 @@ TEST_F(ProfilingTagTest, ProfilingTagSupportedProfilingQueue) {
"urEnqueueTimestampRecordingExp", &after_urEnqueueTimestampRecordingExp);
mock::getCallbacks().set_after_callback("urEventGetProfilingInfo",
&after_urEventGetProfilingInfo);
mock::getCallbacks().set_after_callback(
"urEnqueueEventsWaitWithBarrier", &after_urEnqueueEventsWaitWithBarrier);

sycl::context Ctx{sycl::platform()};
sycl::queue Queue{Ctx,
Expand All @@ -138,8 +140,11 @@ TEST_F(ProfilingTagTest, ProfilingTagSupportedProfilingQueue) {

ASSERT_TRUE(Dev.has(sycl::aspect::ext_oneapi_queue_profiling_tag));

// As an optimization, the implementation will use a single barrier when
// submitting a profiling tag on an out-of-order queue with profiling enabled.
sycl::event E = sycl::ext::oneapi::experimental::submit_profiling_tag(Queue);
ASSERT_EQ(size_t{1}, counter_urEnqueueTimestampRecordingExp);
ASSERT_EQ(size_t{0}, counter_urEnqueueTimestampRecordingExp);
ASSERT_EQ(size_t{1}, counter_urEnqueueEventsWaitWithBarrier);

E.get_profiling_info<sycl::info::event_profiling::command_start>();
ASSERT_TRUE(LatestProfilingQuery.has_value());
Expand Down
Loading