Adapt test_misc.py to HPUs #147609

amathewc · 2025-02-21T09:08:18Z

This PR is related to #145476 . That PR had two files (test_functions.py and test_misc.py) . test_functions was causing CI/rebase/merge issues and hence removed for now. This PR contains only test_misc.py.

This is a continuation of #144387 .

MOTIVATION

We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at #126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices.

Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu )

CHANGES

Create a separate class for test functions running on CUDA devices
Extend the functionality of these tests to include HPUs
Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes
Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices

cc: @ankurneog , @EikanWang , @yanboliang , @guangyey

cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames

pytorch-bot · 2025-02-21T09:08:23Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/147609

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

❌ 12 New Failures

As of commit dba2f05 with merge base bd019c0 ():

NEW FAILURES - The following jobs have failed:

pull / linux-focal-cuda12.4-py3.10-gcc9 / test (default, 2, 5, linux.4xlarge.nvidia.gpu) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-focal-cuda12.4-py3.10-gcc9 / test (default, 3, 5, linux.4xlarge.nvidia.gpu) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes
pull / linux-focal-cuda12.4-py3.10-gcc9-sm89 / test (default, 2, 5, linux.g6.4xlarge.experimental.nvidia.gpu) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-focal-cuda12.4-py3.10-gcc9-sm89 / test (default, 3, 5, linux.g6.4xlarge.experimental.nvidia.gpu) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes
pull / linux-focal-py3.13-clang10 / test (default, 2, 5, linux.4xlarge) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-focal-py3.13-clang10 / test (default, 3, 5, linux.4xlarge) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes
pull / linux-focal-py3.9-clang10 / test (default, 2, 5, linux.4xlarge) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-focal-py3.9-clang10 / test (default, 3, 5, linux.4xlarge) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes
pull / linux-jammy-py3.10-clang15-asan / test (default, 2, 6, linux.4xlarge) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-jammy-py3.10-clang15-asan / test (default, 3, 6, linux.4xlarge) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes
pull / linux-jammy-py3.9-gcc11 / test (default, 2, 5, linux.2xlarge) (gh)
dynamo/test_misc.py::MiscTests::test_iterator_limit
pull / linux-jammy-py3.9-gcc11 / test (default, 3, 5, linux.2xlarge) (gh)
dynamo/test_dynamic_shapes.py::DynamicShapesMiscTests::test_iterator_limit_dynamic_shapes

This comment was automatically generated by Dr. CI and updates every 15 minutes.

amathewc · 2025-02-24T11:15:03Z

@ankurneog , @EikanWang , @yanboliang , @guangyey : please review this PR.

EikanWang · 2025-02-24T11:54:26Z

test/dynamo/test_misc.py

+            res = opt_func(a)
+            self.assertIsInstance(res, torch.Tensor)
+
+    # @unittest.skipIf(not TEST_CUDA, "requires cuda")


Suggested change

# @unittest.skipIf(not TEST_CUDA, "requires cuda")

EikanWang · 2025-02-24T11:57:30Z

test/dynamo/test_misc.py

+            self.assertEqual(res.device.index, 0)
+            self.assertEqual(counter.frame_count, 2)
+
+    def test_torch_device_python_type(self):


Suggested change

def test_torch_device_python_type(self):

def test_torch_device_python_type(self, device):

EikanWang · 2025-02-24T11:59:49Z

test/dynamo/test_misc.py

+    @unittest.skipIf(not torch.cuda.is_available(), "Test requires CUDA.")
+    def test_symint_as_device_kwarg_non_strict_export(self):


Does this case cannot work well on HPU?

EikanWang · 2025-02-24T12:02:30Z

test/dynamo/test_misc.py

+            else:
+                return x - 1
+
+        x = torch.rand(4)


The x is a CPU tensor. How does it verify the CUDA and HPU path?

EikanWang · 2025-02-24T12:07:30Z

test/dynamo/test_misc.py

+        # FIXME(XuehaiPan): do not inline infinite generator if it does not raise errors in eager mode
+        def fn(x):
+            def gen():
+                while True:
+                    yield x
+
+            return list(zip(range(10), gen()))
+
+        x = torch.randn([0, 1, 2, 3, 4, 5])
+        compiled_fn = torch.compile(fn, backend="eager", fullgraph=True)
+        with self.assertRaisesRegex(
+            torch._dynamo.exc.Unsupported, "infinite generator"
+        ):
+            compiled_fn(x)
+


Duplicated code block.

EikanWang · 2025-02-24T12:18:47Z

test/dynamo/test_misc.py


        def write_state(state):
-            torch.set_grad_enabled(state[0])
+            torch.set_grad_enabled(state[0]),


Suggested change

torch.set_grad_enabled(state[0]),

torch.set_grad_enabled(state[0])

EikanWang · 2025-02-24T12:18:55Z

test/dynamo/test_misc.py

+            torch.set_grad_enabled(state[0]),
            torch.use_deterministic_algorithms(state[1])
-            torch._C._set_cublas_allow_tf32(state[2])
+            torch._C._set_cublas_allow_tf32(state[2]),


Suggested change

torch._C._set_cublas_allow_tf32(state[2]),

torch._C._set_cublas_allow_tf32(state[2])

EikanWang · 2025-02-24T12:19:17Z

test/dynamo/test_misc.py

+            torch.cuda.set_device(1)
+            return a + 1
+
+        with torch.cuda.device(0):


This case is device-specific. It should be decorated by requires_cuda.

EikanWang · 2025-02-24T12:19:59Z

test/dynamo/test_misc.py

        _, ne = run(torch.ones(1))
        self.assertFalse(ne)

-    def test_ne_operator_with_custom_ne(self):


EikanWang · 2025-02-24T12:21:52Z

test/dynamo/test_misc.py


        torch.allclose(inp1_custom.grad, inp1_usual.grad)

-    def test_retain_grad(self):


amathewc · 2025-03-19T10:14:45Z

The review comments and merge conflicts will be addressed in a separate PR.

This PR is related to #145476 . That PR had two files (test_functions.py and test_misc.py) . test_functions was causing CI/rebase/merge issues and hence removed for now. This PR contains only test_misc.py. This is a continuation of #144387 . ## MOTIVATION We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at #126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices. Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu ) ## CHANGES Create a separate class for test functions running on CUDA devices Extend the functionality of these tests to include HPUs Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices PS: Most of these changes were initially part of #147609 , but closed that PR due to merge conflicts. The review comments were handled in this PR. Pull Request resolved: #149499 Approved by: https://github.com/EikanWang, https://github.com/desertfire, https://github.com/cyyever

This PR is related to pytorch#145476 . That PR had two files (test_functions.py and test_misc.py) . test_functions was causing CI/rebase/merge issues and hence removed for now. This PR contains only test_misc.py. This is a continuation of pytorch#144387 . ## MOTIVATION We recently integrated support for Intel Gaudi devices (identified as 'hpu') into the common_device_type framework via the pull request at pytorch#126970. This integration allows tests to be automatically instantiated for Gaudi devices upon loading the relevant library. Building on this development, the current pull request extends the utility of these hooks by adapting selected CUDA tests to operate on Gaudi devices. Additionally, we have confirmed that these modifications do not interfere with the existing tests on CUDA devices. Other accelerators can also extend the functionality by adding the device in the devices list. ( For eg: xpu ) ## CHANGES Create a separate class for test functions running on CUDA devices Extend the functionality of these tests to include HPUs Use instantiate_device_type_tests with targeted attributes to generate device-specific test instances within the new classes Apply skipIfHPU decorator to bypass tests that are not yet compatible with HPU devices PS: Most of these changes were initially part of pytorch#147609 , but closed that PR due to merge conflicts. The review comments were handled in this PR. Pull Request resolved: pytorch#149499 Approved by: https://github.com/EikanWang, https://github.com/desertfire, https://github.com/cyyever

Adapt test_misc.py to HPUs

dba2f05

pytorch-bot bot added module: dynamo topic: not user facing topic category labels Feb 21, 2025

pytorchbot added the open source label Feb 21, 2025

EikanWang requested a review from yanboliang February 24, 2025 11:53

EikanWang requested changes Feb 24, 2025

View reviewed changes

mikaylagawarecki added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Feb 25, 2025

amathewc closed this Mar 19, 2025

amathewc deleted the dynamo_changes2 branch March 19, 2025 10:14

amathewc mentioned this pull request Mar 19, 2025

Adapt test_misc.py for HPUs #149499

Closed

	def test_torch_device_python_type(self):
	def test_torch_device_python_type(self, device):

		@unittest.skipIf(not torch.cuda.is_available(), "Test requires CUDA.")
		def test_symint_as_device_kwarg_non_strict_export(self):

	torch.set_grad_enabled(state[0]),
	torch.set_grad_enabled(state[0])

	torch._C._set_cublas_allow_tf32(state[2]),
	torch._C._set_cublas_allow_tf32(state[2])


		torch.allclose(inp1_custom.grad, inp1_usual.grad)

		def test_retain_grad(self):

Adapt test_misc.py to HPUs #147609

Adapt test_misc.py to HPUs #147609

Uh oh!

Conversation

amathewc commented Feb 21, 2025 • edited by pytorch-bot bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

MOTIVATION

CHANGES

Uh oh!

pytorch-bot bot commented Feb 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/147609

❌ 12 New Failures

Uh oh!

amathewc commented Feb 24, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

amathewc commented Mar 19, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

amathewc commented Feb 21, 2025 •

edited by pytorch-bot bot

Loading

pytorch-bot bot commented Feb 21, 2025 •

edited

Loading