Problemsby GlyphTrouble with PyTorch cuda
2 posts from 2 people2d activeFading, +0 this weekunverified
Posts per day
The posts behind it
[PyTorch Nightly] [Windows] [CUDA]: 'fcntl' module is required (but non-existent in Windows)
Trying to use latest Pytorch Nightly in ComfyUI I get this error (relevant portion of the log): My environment: Trying to solve the issue I found out that `fcntl` module doesn't exist in Windows. Using information from this source...
`torch.compile channel_shuffle identity shortcut preserves non-contiguous strides unlike eager CUDA, causing downstream view to fail`
### 🐛 Describe the bug A non-contiguous CUDA tensor passed through `torch.channel_shuffle(x, 1)` and then a legal `.view(...)` works in eager but crashes under `torch.compile` on **all** backends (`eager`, `aot_eager`, `inductor`) during fake-tensor propagation. Since...
Companies and products named
About this problem
Evidence
2 posts from 2 people in 1 place, about 1 a week over 9 days. Mostly on pytorch/pytorch. Tools named alongside: NVIDIA, GitHub.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts about PyTorch with no known issue, sharing the word "cuda". First post Oct 8, 2026, latest Oct 16, 2026. Unverified: fewer than 3 posts, or one person or place only. Method
Other PyTorch problems
PyTorch users keep filing bugs
Top post: "Meta-issue: torch.compile gaps found via the CPython test suite", one of 38 posts from 34 people on GitHub issues, Reddit.
PyTorch upgrades break working setups
13 posts from 13 people on GitHub issues, Reddit.
Trouble with PyTorch inductor
Top post: "`[Inductor][CPU] torch.var returns NaN for large-magnitude constant float64 inputs while eager returns 0`", one of 10 posts from 7 people on GitHub issues.
PyTorch GPU access and drivers are painful
9 posts from 9 people on GitHub issues.
PyTorch builds are slow or flaky
8 posts from 1 person on GitHub issues. Unverified.
Trouble with PyTorch input
Top post: "Generation crashes with encoder_repetition_penalty when encoder_outputs are provided without input_ids", one of 6 posts from 6 people on GitHub issues.
PyTorch deprecations force unplanned migrations
4 posts from 3 people on GitHub issues.
Trouble with PyTorch true
Top post: "torch.cuda.memory_snapshot() drops a reference to True on every call and aborts Python < 3.12 (fix #199044", one of 4 posts from 4 people on GitHub issues.
PyTorch vulnerabilities are hard to triage and patch
3 posts from 3 people on Glyph Ibis.
Running PyTorch locally hits memory and speed limits
3 posts from 3 people on GitHub issues, Hacker News.
PyTorch error messages are cryptic
3 posts from 3 people on GitHub issues, Stack Overflow.
RAG on PyTorch returns irrelevant or missing context
3 posts from 3 people on GitHub issues, Mastodon.
History