# PyTorch CUDA incompatibility with NVIDIA RTX 5070 Ti
**URL:** https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233
**Category:** Support
**Created:** [June 5, 2025, 2:39pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233 "2025-06-05T14:39:06Z")
**Posts on this page:** 6
**Page:** 1
### Author:  [@rodrigot](https://discourse.slicer.org/u/rodrigot)
#### Post date: [June 5, 2025, 2:39pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/1 "2025-06-05T14:39:06Z")
Hello everyone,
I’m currently using a 3D Slicer (version 5.9.0-2025-06-03) on Ubuntu Linux. I have an NVIDIA GeForce RTX 5070 Ti GPU installed.
When launching Slicer, I’m consistently seeing the following `UserWarning` in the Python Interactor:
* * *
NVIDIA GeForce RTX 5070 Ti with CUDA capability sm\_120 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm\_50 sm\_60 sm\_70 sm\_75 sm\_80 sm\_86 sm\_90.
If you want to use the NVIDIA GeForce RTX 5070 Ti GPU with PyTorch, please check the instructions at [https://pytorch.org/get-started/locally/](https://pytorch.org/get-started/locally/)
* * *
My understanding is that while Slicer detects my GPU, the PyTorch version bundled does not support the sm\_120 (compute capability 12.0) architecture of my RTX 5070 Ti.
This means that Slicer modules are unable to access or utilize my GPU for accelerated processing, and consequently, will fall back to CPU execution, which is slower.
Is there anyone who can help resolve this issue so that the Slicer modules can run on this GPU?
Thank you so much.
---
### Author:  [@muratmaga](https://discourse.slicer.org/u/muratmaga)
#### Post date: [June 5, 2025, 3:22pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/2 "2025-06-05T15:22:36Z")
As far as I can tell only 12.8 series support 5070Ti. So uninstall your pytorch (I assume you are using PyTorch utils?) restart slicer, then instead of automatic search for compatible versions, choose specifically cu12.8 and try again.
---
### Author:  [@Mark\_Ryan](https://discourse.slicer.org/u/Mark_Ryan)
#### Post date: [June 7, 2025, 10:51pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/3 "2025-06-07T22:51:05Z")
Running on a 5090 - had to update Pytorch to 12.8 from within slicer python console. Around a 4 GB download so takes a while. SM\_120 refers to the 50 series GPUs I believe, so 12.8 was the only supported version.
---
### Author:  [@jamesobutler](https://discourse.slicer.org/u/jamesobutler)
#### Post date: [June 9, 2025, 3:18pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/4 "2025-06-09T15:18:45Z")
@Mark_Ryan You were successful with this on Windows? Or were you using a Linux distro? Rodrigo appears to be using Ubuntu 24.04 per:
>
>
> Request for PyTorch 2.7 Support in 3D Slicer for NVIDIA Blackwell GPUs
>
> System I…nformation:
> \- GPU: NVIDIA GeForce RTX 5070 Ti
> \- Architecture: NVIDIA Blackwell
> \- OS: Linux Unbutu 24.04.2 LTS
>
> PyTorch version 2.7 introduces support for the Blackwell architecture, which is required for compatibility with this GPU. Earlier versions of PyTorch (≤2.6) do not support it, resulting in the following error:
> \- “The current PyTorch install supports CUDA capabilities sm\_50 sm\_60 sm\_70 sm\_75 sm\_80 sm\_86 sm\_90.”
>
> Unfortunately, the PyTorch version currently available in 3D Slicer is limited to 2.6
---
### Author:  [@Mark\_Ryan](https://discourse.slicer.org/u/Mark_Ryan)
#### Post date: [June 9, 2025, 6:26pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/5 "2025-06-09T18:26:24Z")
This was from within windows, but everything downloaded within python shell in Slicer. But I needed ChatGPT to get me that far, so can’t speak for the Linux thing.
---
### Author:  [@jamesobutler](https://discourse.slicer.org/u/jamesobutler)
#### Post date: [June 9, 2025, 7:57pm UTC](https://discourse.slicer.org/t/pytorch-cuda-incompatibility-with-nvidia-rtx-5070-ti/43233/6 "2025-06-09T19:57:22Z")
Ok thanks for the update. Yes, Rodrigo’s issues exists because of using the Slicer linux package which is currently built based on CentOS 7 and glibc 2.17 and there is not a Pytorch 2.7.x whl that is compatible with it.
Pytorch 2.6.0 for example has manylinux1 (CentOS 5 based, glibc 2.5) based whls which can be run by the Slicer linux package, however Pytorch 2.7.0 has manylinux\_2\_28 (glibc 2.28) based whls which is too new of a glibc to match what Slicer provides for the linux package.
> **[GitHub - pypa/manylinux: Python wheels that work on any linux (almost)](https://github.com/pypa/manylinux?tab=readme-ov-file#manylinux)**
>
> Python wheels that work on any linux (almost)