# How to check which cuda version my pytorch is using **URL:** \ **Category:** Uncategorized\ **Created:** [March 31, 2021, 3:57pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622 "2021-03-31T15:57:46Z")\ **Posts on this page:** 8\ **Page:** 1
**Author:** ![JaeJin\_Cho](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/jaejin_cho/32/1042_2.png) [@JaeJin\_Cho](https://discuss.pytorch.org/u/JaeJin_Cho)\ **Post date:** [March 31, 2021, 3:57pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/1 "2021-03-31T15:57:46Z")
Hello! I have multiple CUDA versions installed on the server, e.g., /opt/NVIDIA/cuda-9.1 and /opt/NVIDIA/cuda-10, and /usr/local/cuda is linked to the latter one. I believe I installed my pytorch with cuda 10.2 based on what I get from running torch.version.cuda. How can I check which version of CUDA that the installed pytorch actually uses in running? I set my CUDA\_PATH=/opt/NVIDIA/cuda-9.1 but it still seems to run without any problem on a gpu. Thanks, Jaejin Cho ---
**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\ **Post date:** [April 1, 2021, 6:36am UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/3 "2021-04-01T06:36:44Z")
You could check the linked CUDA version via `print(torch.version.cuda)`. ---
**Author:** ![JaeJin\_Cho](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/jaejin_cho/32/1042_2.png) [@JaeJin\_Cho](https://discuss.pytorch.org/u/JaeJin_Cho)\ **Post date:** [April 1, 2021, 1:42pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/5 "2021-04-01T13:42:43Z")
Thank you for the quick answer. I actually used “torch.version.cuda” and checked it as in my original post (and it was 10.2). However, I was wondering why it still works with the setting “CUDA\_PATH=/opt/NVIDIA/cuda-9.1”. In other words, how it finds the correct cuda version to use (or which environment variable it uses to do so) in run-time? ---
**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\ **Post date:** [April 1, 2021, 4:15pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/6 "2021-04-01T16:15:05Z")
`CUDA_PATH` might be the wrong env var. I’m using `CUDA_HOME`, which works fine. ---
**Author:** ![JaeJin\_Cho](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/jaejin_cho/32/1042_2.png) [@JaeJin\_Cho](https://discuss.pytorch.org/u/JaeJin_Cho)\ **Post date:** [April 4, 2021, 1:17am UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/7 "2021-04-04T01:17:52Z")
Thank you for the answer! Added this code in my script, logging.warning(‘cuda version: {}’.format(torch.version.cuda)) import os logging.warning(‘CUDA\_PATH: {}’.format(os.environ[“CUDA\_PATH”])) logging.warning(‘CUDA\_HOME: {}’.format(os.environ[“CUDA\_HOME”])) Then, the outputs are like below: WARNING: cuda version: 10.2 WARNING: CUDA\_PATH: /opt/NVIDIA/cuda-9.1 WARNING: CUDA\_HOME: /opt/NVIDIA/cuda-9.1 It seems working fine but I am still curious what pytorch refers to for the correct CUDA version under the hood. ---
**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\ **Post date:** [April 4, 2021, 1:20am UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/8 "2021-04-04T01:20:33Z")
The `CUDA_HOME` env var is used during the build of PyTorch in order to compile the CUDA kernels, so you would need to rebuild it. ---
**Author:** ![doem97](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/doem97/32/39380_2.png) [@doem97](https://discuss.pytorch.org/u/doem97)\ **Post date:** [September 10, 2023, 3:35pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/9 "2023-09-10T15:35:31Z")
@ptrblck I am a little confused by cuda version, torch version. My CUDA\_HOME currently points to `cuda v11.1` While my existing torch installation reads `torch.version.cuda=11.7` Which cuda executive lib it is calling? I am confused because this wrong-version torch seems running quite fine. ---
**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\ **Post date:** [September 10, 2023, 6:30pm UTC](https://discuss.pytorch.org/t/how-to-check-which-cuda-version-my-pytorch-is-using/116622/10 "2023-09-10T18:30:53Z")
PyTorch ships with its own CUDA dependencies and your locally installed CUDA toolkit will be used if you build PyTorch from source or a custom CUDA extension.