Description
🐛 Describe the bug
Edge-case input leads to what seems to be a persistent infinite loop. This issue was discovered in #134184 and left as follow-up work. There is also a correctness bug in nn.ReplicationPad1d (see #135447).
shape = [65536, 2, 4] in the MRE below, results in 100% GPU utilization, which will persist even if the Python process is killed. Other shapes (eg. shape = [65535, 2, 4] or shape = [65537, 2, 4]) will not produce this behavior.
import torch
device = 'mps'
shape = [65536, 2, 4]
pl, pr = 3, 4
x = torch.randn(shape, device=device, requires_grad=True)
model = torch.nn.ReplicationPad1d((pl, pr))
out = model(x)
g = torch.randn_like(out)
out.backward(g)
print(x.grad[:, :, 1 : -1])Versions
PyTorch version: 2.5.0a0+git042f2f7
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 14.6.1 (arm64)
GCC version: Could not collect
Clang version: 15.0.0 (clang-1500.3.9.4)
CMake version: version 3.30.1
Libc version: N/A
Python version: 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:13:44) [Clang 16.0.6 ] (64-bit runtime)
Python platform: macOS-14.6.1-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Apple M3 Max
Versions of relevant libraries:
[pip3] flake8==6.1.0
[pip3] flake8-bugbear==23.3.23
[pip3] flake8-comprehensions==3.15.0
[pip3] flake8-executable==2.1.3
[pip3] flake8-logging-format==0.9.0
[pip3] flake8-pyi==23.3.1
[pip3] flake8-simplify==0.19.3
[pip3] mypy==1.10.0
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.26.0
[pip3] optree==0.12.1
[pip3] torch==2.5.0a0+git042f2f7
[pip3] torch-tb-profiler==0.4.3
[pip3] torchvision==0.20.0a0+0d80848
[pip3] triton==3.0.0
[conda] numpy 1.26.0 pypi_0 pypi
[conda] optree 0.12.1 pypi_0 pypi
[conda] torch 2.5.0a0+git042f2f7 dev_0
[conda] torch-tb-profiler 0.4.3 pypi_0 pypi
[conda] torchfix 0.4.0 pypi_0 pypi
[conda] torchvision 0.20.0a0+0d80848 dev_0
[conda] triton 3.0.0 pypi_0 pypi
cc @ezyang @gchanan @zou3519 @kadeng @msaroufim @kulinseth @albanD @malfet @DenisVieriu97 @jhavukainen