Skip to content

fix segfault in cat on CPU with tensors that can't be indexed with 32-bit ints. #21530

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
umanwizard wants to merge 2 commits into from
Closed

fix segfault in cat on CPU with tensors that can't be indexed with 32-bit ints. #21530

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3adf33c
fix cat on CPU with big tensors
Jun 7, 2019
8b1a6b1
slowTest
Jun 7, 2019
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
2 changes: 1 addition & 1 deletion aten/src/TH/generic/THTensor.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -799,7 +799,7 @@ void THTensor_(catArray)(THTensor *result, THTensor **inputs, int numInputs, int
if (!should_skip(inputs[j])) {
THTensor* input0 = inputs[j];
scalar_t* input0_data = THStorage_(data)(THTensor_getStoragePtr(input0)) + input0->storage_offset();
int local_inner = inner * input0->size(dimension);
int64_t local_inner = inner * input0->size(dimension);
if (local_inner != 0) {
memcpy(result_data + offset, input0_data + o*local_inner, local_inner*sizeof(scalar_t));
} // input0_size != 0
Expand Down
10 changes: 10 additions & 0 deletions test/test_torch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4948,6 +4948,16 @@ def _test_cat_empty(self, use_cuda=False):
def test_cat_empty(self):
self._test_cat_empty(self)

@slowTest
def test_cat_big(self):
Copy link
Member

@fmassa fmassa Jun 7, 2019

Choose a reason for hiding this comment

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

Should this test be marked as slow or something like that? Also, how much memory does it need, > 2GB?

Copy link
Contributor Author

@umanwizard umanwizard Jun 7, 2019 •
edited
Loading

Choose a reason for hiding this comment

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

@fmassa The test needs 5 GB and takes about 8 seconds on my machine. Do you know if there is a good way to mark tests as not running in memory-constrained environments?

Alternatively we can just land this without the test. What do you think?

Copy link
Member

@fmassa fmassa Jun 7, 2019

Choose a reason for hiding this comment

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

I think we should have the test. Maybe at some point in the future we might consider adding a flag for memory intensive tests, both on the cpu and you

SIZE1 = 6500
SIZE2 = 4500
concat_list = []
concat_list.append(torch.ones((SIZE1, 1024 * 512), dtype=torch.uint8))
concat_list.append(torch.ones((SIZE2, 1024 * 512), dtype=torch.uint8))
result = torch.cat(concat_list)
self.assertEqual(result.size(0), SIZE1 + SIZE2)

def test_narrow(self):
x = torch.Tensor([[0, 1, 2], [3, 4, 5], [6, 7, 8]])
self.assertEqual(x.narrow(0, 0, 1), torch.Tensor([[0, 1, 2]]))
Expand Down