Skip to content

[TD] Enable test removal on most default configs + distributed CUDA for everyone #125931

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
wants to merge 7 commits into from
Closed

[TD] Enable test removal on most default configs + distributed CUDA for everyone #125931

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
05492d5
yolo
clee2000 May 10, 2024
620ded9
tc
clee2000 May 10, 2024
4cb91e2
tc
clee2000 May 10, 2024
94c3cbf
update
clee2000 May 13, 2024
2f79b3a
update
clee2000 May 13, 2024
f7ebbdf
update
clee2000 May 13, 2024
b9a423e
tc
clee2000 May 13, 2024
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
1 change: 0 additions & 1 deletion .github/pytorch-probot.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
tracking_issue: 24422
ciflow_tracking_issue: 64124
TD_rollout_issue: 123120
ciflow_push_tags:
- ciflow/binaries
- ciflow/binaries_conda
Expand Down
24 changes: 16 additions & 8 deletions test/run_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,6 @@
TEST_WITH_ASAN,
TEST_WITH_CROSSREF,
TEST_WITH_ROCM,
TEST_WITH_SLOW,
TEST_WITH_SLOW_GRADCHECK,
)

Expand Down Expand Up @@ -76,9 +75,11 @@
sys.path.remove(str(REPO_ROOT))

TEST_CONFIG = os.getenv("TEST_CONFIG", "")
BUILD_ENVIRONMENT = os.getenv("BUILD_ENVIRONMENT", "")
RERUN_DISABLED_TESTS = os.getenv("PYTORCH_TEST_RERUN_DISABLED_TESTS", "0") == "1"
DISTRIBUTED_TEST_PREFIX = "distributed"
INDUCTOR_TEST_PREFIX = "inductor"
IS_SLOW = "slow" in TEST_CONFIG or "slow" in BUILD_ENVIRONMENT


# Note [ROCm parallel CI testing]
Expand Down Expand Up @@ -494,7 +495,7 @@ def run_test(
None
if not options.enable_timeout
else THRESHOLD * 6
if TEST_WITH_SLOW
if IS_SLOW
else THRESHOLD * 3
if should_retry
and isinstance(test_module, ShardedTest)
Expand Down Expand Up @@ -1180,18 +1181,25 @@ def parse_args():
and (
TEST_WITH_CROSSREF
or TEST_WITH_ASAN
or (
strtobool(os.environ.get("TD_DISTRIBUTED", "False"))
and TEST_CONFIG == "distributed"
and TEST_CUDA
)
or (TEST_CONFIG == "distributed" and TEST_CUDA)
or (IS_WINDOWS and not TEST_CUDA)
or TEST_CONFIG == "nogpu_AVX512"
or TEST_CONFIG == "nogpu_NO_AVX2"
or (
"sm86" not in BUILD_ENVIRONMENT
Copy link
Contributor

Choose a reason for hiding this comment

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

Ok, the logic to check for "sm86" here looks a bit awkward I think because that can be changed in the future. But this is probably ok because I assume this is only temporary

Copy link
Contributor Author

@clee2000 clee2000 May 14, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

yeah all of the logic here is a bit cursed because of how specific the inclusions and exclusions are

and TEST_CONFIG == "default"
and TEST_CUDA
)
or (not TEST_CUDA and TEST_CONFIG == "default")
)
and get_pr_number() is not None
and not strtobool(os.environ.get("NO_TD", "False"))
and not TEST_WITH_SLOW,
and not IS_SLOW
and not TEST_WITH_ROCM
and not IS_MACOS
and "onnx" not in BUILD_ENVIRONMENT
and "debug" not in BUILD_ENVIRONMENT
and "parallelnative" not in BUILD_ENVIRONMENT,
)
parser.add_argument(
"additional_unittest_args",
Expand Down