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.
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
Torch compile: libcuda.so cannot found #107960
Comments
I don't believe this is a When you're Also keep in mind that |
Hey @msaroufim - many thanks for the detailed response! However, I'm not sure this originates from having multiple versions of PyTorch installed. If I uninstall the base version of PyTorch, and then install PyTorch nightly, I get the If you're confident this is a Colab issue, I can raise it with the Google team! But it would be great to rule out a PyTorch compile issue before doing this! |
Trying your Colab notebook, I could verify that the issue isn't from PyTorch, but have found the workaround: After installing the nightly build, by running
No
The
Verify inference both without compile and with compile:
The first call with compilation takes a long time (
We see a big speed increase even on the T4 GPU (although initial compilation does take quite a while). Keep in mind all the shell commands are to be run with a preceding |
What is the problem and how can solve it? I don't understand. |
@poly-mer 's solution worked for me, but there was no speedup for me, these are the commands I ran step by step: !export LC_ALL="en_US.UTF-8" |
Same issue on the official pytorch cuda 2.1.0 docker image |
I'm exausted. |
this works, thanks |
I have the same problem, but also with the preinstalled torch version. import torch
torch.set_default_device("cuda:0")
@torch.compile
def test(x):
return torch.sin(x)
a = torch.zeros(100)
test(a) After a long backtrace it reports:
Calling
best wishes |
I'm using Singularity on
|
Do you guys know how to solve "libcuda.so cannot found" if I don't have root access? Cuda drivers work with no problem (checked with nvidia-smi and torch). |
Downgrading |
Thanks, it works. |
Closing, feel free to reopen if needed |
🐛 Describe the bug
Using
torch.compile
with a Colab T4 GPU fails and gives a very cryptic error running on nightly 2.1Error logs
Minified repro
https://colab.research.google.com/drive/1XwD2UpPoi6RFLHA9tcXL7BdbOgkKvQ_7?usp=sharing
Versions
PyTorch version: 2.1.0.dev20230825+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A
OS: Ubuntu 22.04.2 LTS (x86_64)
GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Clang version: 14.0.0-1ubuntu1.1
CMake version: version 3.27.2
Libc version: glibc-2.35
Python version: 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] (64-bit runtime)
Python platform: Linux-5.15.109+-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: 11.8.89
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: Tesla T4
Nvidia driver version: 525.105.17
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.0
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 46 bits physical, 48 bits virtual
Byte Order: Little Endian
CPU(s): 2
On-line CPU(s) list: 0,1
Vendor ID: GenuineIntel
Model name: Intel(R) Xeon(R) CPU @ 2.20GHz
CPU family: 6
Model: 79
Thread(s) per core: 2
Core(s) per socket: 1
Socket(s): 1
Stepping: 0
BogoMIPS: 4399.99
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx smap xsaveopt arat md_clear arch_capabilities
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 32 KiB (1 instance)
L1i cache: 32 KiB (1 instance)
L2 cache: 256 KiB (1 instance)
L3 cache: 55 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0,1
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Mitigation; PTE Inversion
Vulnerability Mds: Vulnerable; SMT Host state unknown
Vulnerability Meltdown: Vulnerable
Vulnerability Mmio stale data: Vulnerable
Vulnerability Retbleed: Vulnerable
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers
Vulnerability Spectre v2: Vulnerable, IBPB: disabled, STIBP: disabled, PBRSB-eIBRS: Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Vulnerable
Versions of relevant libraries:
[pip3] numpy==1.23.5
[pip3] pytorch-triton==2.1.0+e6216047b8
[pip3] torch==2.1.0.dev20230825+cu121
[pip3] torchaudio==2.1.0.dev20230825+cu121
[pip3] torchdata==0.6.1
[pip3] torchsummary==1.5.1
[pip3] torchtext==0.15.2
[pip3] torchvision==0.15.2+cu118
[pip3] triton==2.0.0
[conda] Could not collect
cc @ezyang @msaroufim @wconstab @bdhirsh @anijain2305
The text was updated successfully, but these errors were encountered: