fix functionalization resize stride compute

[bdhirsh]

uncovered from an OpInfo in inductor, when I turned on functionalization

Stack from ghstack (oldest at bottom):

• [IGNORE, not used] functionalization: support preserving (simple) input mutations #93994
• turn functionalization on in aot_autograd inference #92857
• functionalization: fix x.is_contiguous(channels_last) #94195
• -> fix functionalization resize stride compute #94018
• add torch.autograd._unsafe_set_version_counter API #92924

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/94018

• 📄 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.

✅ No Failures

As of commit 8b4e38b:
💚 Looks good so far! There are no failures yet. 💚

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

[ezyang]

No, I don't understand. The old code looks right and the new code looks like you do an out of bounds strides access on the first iteration

[bdhirsh]

yep you're right - somehow I convinced myself that my test passed locally. Updated, to use a convenient existing util that we have (the bug was that I should have been doing strides[i] = strides[i+1] * sizes[i+1] instead of strides[i] = strides[i+1] * sizes[i]).