Skip to content

[0.6.0-UT] Fix Test Shard #519

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
wants to merge 1 commit into
base: rocm-jaxlib-v0.6.0
Choose a base branch
from
Open

[0.6.0-UT] Fix Test Shard #519

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
6 changes: 4 additions & 2 deletions tests/ffi_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -276,10 +276,12 @@ def test_invalid_result_type(self):

@jtu.run_on_devices("gpu", "cpu")
def test_shard_map(self):
if jtu.is_device_rocm:
self.skipTest("Skip on ROCm: tests/ffi_test.py::FfiTest::test_shard_map")
# if jtu.is_device_rocm:
Copy link

@gulsumgudukbay gulsumgudukbay Sep 10, 2025

Choose a reason for hiding this comment

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

Please remove commented lines.

# self.skipTest("Skip on ROCm: tests/ffi_test.py::FfiTest::test_shard_map")
mesh = jtu.create_mesh((len(jax.devices()),), ("i",))
x = self.rng().randn(8, 4, 5).astype(np.float32)
n = len(jax.devices())
x = x[:(x.shape[0] // n) * n]
Copy link

@gulsumgudukbay gulsumgudukbay Sep 10, 2025

Choose a reason for hiding this comment

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

This trims the array, however trimming the array could hinder the mathematical correction of this test. Instead of trimming we can also pad the array like I did in #509 to prevent information loss. Could you please explain how this is mathematically correct and we are not losing info?


@partial(shard_map, mesh=mesh, in_specs=P("i"), out_specs=P("i"))
def f(x):
Expand Down