CUDA not available in nightly, but is for 2.0.1

How to reproduce

> mamba create --name pytorch-test python=3.9
> mamba activate pytorch-test
> mamba install -c "nvidia/label/cuda-11.8.0" cuda-toolkit
> mamba install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
> python
>>> import torch
>>> torch.cuda.is_available()

> mamba create --name pytorch-test-2 python=3.9
> mamba activate pytorch-test-2
> mamba install -c "nvidia/label/cuda-11.8.0" cuda-toolkit
> mamba install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch-nightly -c nvidia
> python
>>> import torch
>>> torch.cuda.is_available()
False # should be True, no??

Output from python

Collecting environment information...
PyTorch version: 2.1.0.dev20230905
Is debug build: False
CUDA used to build PyTorch: Could not collect
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.3 LTS (x86_64)
GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.35

Python version: 3.9.18 | packaged by conda-forge | (main, Aug 30 2023, 03:49:32)  [GCC 12.3.0] (64-bit runtime)
Python platform: Linux-6.2.0-32-generic-x86_64-with-glibc2.35
Is CUDA available: False
CUDA runtime version: 11.8.89
GPU models and configuration: GPU 0: NVIDIA GeForce GTX 1060 6GB
Nvidia driver version: 525.125.06
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

Architecture:                       x86_64
CPU op-mode(s):                     32-bit, 64-bit
Address sizes:                      43 bits physical, 48 bits virtual
Byte Order:                         Little Endian
CPU(s):                             12
On-line CPU(s) list:                0-11
Vendor ID:                          AuthenticAMD
Model name:                         AMD Ryzen 5 2600 Six-Core Processor
CPU family:                         23
Model:                              8
Thread(s) per core:                 2
Core(s) per socket:                 6
Socket(s):                          1
Stepping:                           2
Frequency boost:                    enabled
CPU max MHz:                        3400.0000
CPU min MHz:                        1550.0000
BogoMIPS:                           6786.17
Flags:                              fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb hw_pstate ssbd ibpb vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 clzero irperf xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif overflow_recov succor smca sev sev_es
Virtualisation:                     AMD-V
L1d cache:                          192 KiB (6 instances)
L1i cache:                          384 KiB (6 instances)
L2 cache:                           3 MiB (6 instances)
L3 cache:                           16 MiB (2 instances)
NUMA node(s):                       1
NUMA node0 CPU(s):                  0-11
Vulnerability Gather data sampling: Not affected
Vulnerability Itlb multihit:        Not affected
Vulnerability L1tf:                 Not affected
Vulnerability Mds:                  Not affected
Vulnerability Meltdown:             Not affected
Vulnerability Mmio stale data:      Not affected
Vulnerability Retbleed:             Mitigation; untrained return thunk; SMT vulnerable
Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:           Mitigation; Retpolines, IBPB conditional, STIBP disabled, RSB filling, PBRSB-eIBRS Not affected
Vulnerability Srbds:                Not affected
Vulnerability Tsx async abort:      Not affected

Versions of relevant libraries:
[pip3] numpy==1.25.2
[pip3] torch==2.1.0.dev20230905
[pip3] torchaudio==2.2.0.dev20230905
[pip3] torchvision==0.16.0.dev20230905
[conda] blas                      2.116                       mkl    conda-forge
[conda] blas-devel                3.9.0            16_linux64_mkl    conda-forge
[conda] libblas                   3.9.0            16_linux64_mkl    conda-forge
[conda] libcblas                  3.9.0            16_linux64_mkl    conda-forge
[conda] liblapack                 3.9.0            16_linux64_mkl    conda-forge
[conda] liblapacke                3.9.0            16_linux64_mkl    conda-forge
[conda] mkl                       2022.1.0           h84fe81f_915    conda-forge
[conda] mkl-devel                 2022.1.0           ha770c72_916    conda-forge
[conda] mkl-include               2022.1.0           h84fe81f_915    conda-forge
[conda] numpy                     1.25.2           py39h6183b62_0    conda-forge
[conda] pytorch                   2.1.0.dev20230905     py3.9_cpu_0    pytorch-nightly
[conda] pytorch-cuda              11.8                 h7e8668a_5    pytorch-nightly
[conda] pytorch-mutex             1.0                         cpu    pytorch-nightly
[conda] torchaudio                2.2.0.dev20230905        py39_cpu    pytorch-nightly
[conda] torchvision               0.16.0.dev20230905        py39_cpu    pytorch-nightly


I’m trying to test whether `torch.asarray` does not respect `set_default_device` · Issue #106773 · pytorch/pytorch · GitHub has affected behaviour when the input to torch.asarray is a numpy array.

Hi Lucas!

I can reproduce this (using conda on linux):

>>> import torch
>>> torch.__version__
>>> torch.version.cuda
>>> torch.cuda.is_available()

I get this with the install command copied from the home page:

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch-nightly -c nvidia

and also with cuda 12.1:

conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch-nightly -c nvidia

In both cases no part of conda install seems to fail and various cuda stuff
does seem to be getting installed. (2.0.1 does work with cuda for me.)

(It’s been months since I’ve installed a nightly, so I have no idea when this broke.)


K. Frank

Thanks for reproducing @KFrank ! Should I open an issue on the repo?

Hi Lucas!

I would think so – can’t hurt.

I’m pretty sure that the nightlies are built automatically. They do get broken
from time to time and letting a human know would likely be appreciated.


K. Frank

Thanks for reporting this issue @lucascolley!
I can also reproduce it and indeed it seems no CUDA builds were uploaded in the last few days as seen here. You could work around this issue by either installing the last successful build from conda via:

conda install pytorch==2.1.0.dev20230901 pytorch-cuda=11.8 -c pytorch-nightly -c nvidia

which should then show:

The following packages will be downloaded:

    package                    |            build
    pytorch-2.1.0.dev20230901  |py3.10_cuda11.8_cudnn8.7.0_0        1.46 GB  pytorch-nightly
    pytorch-mutex-1.0          |             cuda           3 KB  pytorch-nightly

or you could also use the nightly pip wheels. I’ll forward it, but you can also create an issue in the pytorch/pytorch or pytorch/builder repository.

Also, thanks @KFrank for checking it so quickly and verifying!

1 Like

Thanks @ptrblck ! I’ll let you know if I have any problems with using the latest available version. I’ve opened CUDA not available in nightly · Issue #108607 · pytorch/pytorch · GitHub.

1 Like

Thank you @lucascolley we are looking into this issue now. We will try to resolve this asap.

1 Like