commit c4321e139663ff2f3ae162043180a8eaba899fe3
author Michael Melesse <micmelesse@gmail.com> Thu Mar 31 18:41:02 2022 +0000
committer PyTorch MergeBot <pytorchmergebot@users.noreply.github.com> Thu Mar 31 18:41:02 2022 +0000
tree 457cbf3ea7f8c3623b1c95c580e5f1fcee979d5a
parent 60729d02f11c2f78b2182d5f005a056dd2ac6488

[ROCm] unskip FFT tests

This pr enables a lot of fft tests on rocm, see the full list here https://github.com/ROCmSoftwarePlatform/pytorch/issues/924.

After enabling the tests we found that 3 tests , test_reference_1d, test_reference_nd and test_fn_grad have issue.  We skip those tests on ROCM in this pr as well. We will address those skipped tests in a subsequent pr.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/74713
Approved by: https://github.com/malfet

test/test_ops_gradients.py

diff --git a/test/test_ops_gradients.py b/test/test_ops_gradients.py
index 71bb243..d50b0d3 100644
--- a/test/test_ops_gradients.py
+++ b/test/test_ops_gradients.py
@@ -7,7 +7,7 @@
     (TestCase, is_iterable_of_tensors, run_tests, gradcheck, gradgradcheck, first_sample)
 from torch.testing._internal.common_methods_invocations import op_db
 from torch.testing._internal.common_device_type import \
-    (instantiate_device_type_tests, ops, OpDTypes)
+    (instantiate_device_type_tests, ops, OpDTypes, skipCUDAIfRocm)
 
 # TODO: fixme https://github.com/pytorch/pytorch/issues/68972
 torch.set_default_dtype(torch.float32)
@@ -113,6 +113,7 @@
             self.skipTest("Skipped! Complex autograd not supported.")
 
     # Tests that gradients are computed correctly
+    @skipCUDAIfRocm
     @_gradcheck_ops(op_db)
     def test_fn_grad(self, device, dtype, op):
         self._skip_helper(op, device, dtype)

test/test_spectral_ops.py

diff --git a/test/test_spectral_ops.py b/test/test_spectral_ops.py
index c11b87b..344c810 100644
--- a/test/test_spectral_ops.py
+++ b/test/test_spectral_ops.py
@@ -13,7 +13,7 @@
     (TestCase, run_tests, TEST_NUMPY, TEST_LIBROSA, TEST_MKL)
 from torch.testing._internal.common_device_type import \
     (instantiate_device_type_tests, ops, dtypes, onlyNativeDeviceTypes,
-     skipCPUIfNoFFT, deviceCountAtLeast, onlyCUDA, OpDTypes, skipIf)
+     skipCPUIfNoFFT, skipCUDAIfRocm, deviceCountAtLeast, onlyCUDA, OpDTypes, skipIf)
 from torch.testing._internal.common_methods_invocations import (
     spectral_funcs, SpectralFuncInfo, SpectralFuncType)
 
@@ -204,6 +204,7 @@
         else:
             return (input, s, dim, norm)
 
+    @skipCUDAIfRocm
     @onlyNativeDeviceTypes
     @ops([op for op in spectral_funcs if op.ndimensional == SpectralFuncType.OneD])
     def test_reference_1d(self, device, dtype, op):
@@ -367,6 +368,7 @@
             op(x)
 
     # nd-fft tests
+    @skipCUDAIfRocm
     @onlyNativeDeviceTypes
     @unittest.skipIf(not TEST_NUMPY, 'NumPy not found')
     @ops([op for op in spectral_funcs if op.ndimensional == SpectralFuncType.ND])

torch/testing/_internal/common_methods_invocations.py

diff --git a/torch/testing/_internal/common_methods_invocations.py b/torch/testing/_internal/common_methods_invocations.py
index c0c4ec8..c873b1e 100644
--- a/torch/testing/_internal/common_methods_invocations.py
+++ b/torch/testing/_internal/common_methods_invocations.py
@@ -5330,7 +5330,6 @@
         decorators = list(decorators) if decorators is not None else []
         decorators += [
             skipCPUIfNoFFT,
-            skipCUDAIfRocm,
         ]
 
         super().__init__(name=name,