# Container: CUDA **URL:** https://forums.developer.nvidia.com/c/accelerated-computing/nvidia-gpu-cloud-ngc-users/container-cuda/29.md [Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md) --- ## [Official rocky8 Container HPC SDK 26.5](https://forums.developer.nvidia.com/t/official-rocky8-container-hpc-sdk-26-5/383522)
**Author:** [@Josh\_Perske](https://forums.developer.nvidia.com/u/Josh_Perske)\ **Replies:** 0\ **Last updated:** [September 17, 2026, 9:25am UTC](https://forums.developer.nvidia.com/t/official-rocky8-container-hpc-sdk-26-5/383522 "2026-09-17T09:25:24Z")
Hi Nvidia Team, I’m not sure if I’m in the correct section with my question, if not, feel free to move this thread somewhere else. In the past we used the containers provided on NVIDIA HPC SDK | NVIDIA NGC as base cont… --- ## [CUDA RHEL8 RHEL9 repository metadata references missing RPM causing mirror synchronization failures](https://forums.developer.nvidia.com/t/cuda-rhel8-rhel9-repository-metadata-references-missing-rpm-causing-mirror-synchronization-failures/376216)
**Author:** [@mariusz.pradun](https://forums.developer.nvidia.com/u/mariusz.pradun)\ **Replies:** 4\ **Last updated:** [July 24, 2026, 7:09am UTC](https://forums.developer.nvidia.com/t/cuda-rhel8-rhel9-repository-metadata-references-missing-rpm-causing-mirror-synchronization-failures/376216 "2026-07-24T07:09:11Z")
You have metadata and package inconsistence in globally available repositories https://developer.download.nvidia.com/compute/cuda/repos/rhel8/x86\_64/ https://developer.download.nvidia.com/compute/cuda/repos/rhel9/x86\_6… --- ## [Missing post-processing for NGC torch image](https://forums.developer.nvidia.com/t/missing-post-processing-for-ngc-torch-image/376191)
**Author:** [@yangpeihao](https://forums.developer.nvidia.com/u/yangpeihao)\ **Replies:** 0\ **Last updated:** [July 9, 2026, 9:12am UTC](https://forums.developer.nvidia.com/t/missing-post-processing-for-ngc-torch-image/376191 "2026-07-09T09:12:10Z")
Thanks for your efforts on building NGC torch image for the community. But recently we found torch in NGC image is different from the torch official distribution. Because it loses the post-processing procedure in the tor… --- ## [I’m having issues accessing the VM and course content for “Fundamentals of Accelerated Computing with CUDA Python”](https://forums.developer.nvidia.com/t/i-m-having-issues-accessing-the-vm-and-course-content-for-fundamentals-of-accelerated-computing-with-cuda-python/346241)
**Author:** [@bilalraza5961](https://forums.developer.nvidia.com/u/bilalraza5961)\ **Replies:** 0\ **Last updated:** [September 28, 2025, 7:18am UTC](https://forums.developer.nvidia.com/t/i-m-having-issues-accessing-the-vm-and-course-content-for-fundamentals-of-accelerated-computing-with-cuda-python/346241 "2025-09-28T07:18:40Z")
I’m having issues accessing the VM and course content for “Fundamentals of Accelerated Computing with CUDA Python” via NVIDIA Training Solutions | NVIDIA When I click “start task” to start the VM and access the sections… --- ## [Lab coursework submission not working](https://forums.developer.nvidia.com/t/lab-coursework-submission-not-working/345371)
**Author:** [@pranavhprivate](https://forums.developer.nvidia.com/u/pranavhprivate)\ **Replies:** 1\ **Last updated:** [September 18, 2025, 1:41pm UTC](https://forums.developer.nvidia.com/t/lab-coursework-submission-not-working/345371 "2025-09-18T13:41:14Z")
Hello, I am doing the course “Accelerated Computing with CUDA Python” and one of the lab coursework isn’t working. i have finished all the tasks mentioned and when i assess this task an error occurs, i have tried it agai… --- ## [Unknown error reading config file: /usr/local/etc/opae/opae.cfg](https://forums.developer.nvidia.com/t/unknown-error-reading-config-file-usr-local-etc-opae-opae-cfg/343145)
**Author:** [@y-shida-tg](https://forums.developer.nvidia.com/u/y-shida-tg)\ **Replies:** 0\ **Last updated:** [August 26, 2025, 12:21am UTC](https://forums.developer.nvidia.com/t/unknown-error-reading-config-file-usr-local-etc-opae-opae-cfg/343145 "2025-08-26T00:21:19Z")
We are developing a container application using OPAE-SDK. We are using nvidia/cuda:12.9.0-base-ubi9 as the base image, but the following error is output when starting the container. Below is the log output by the contain… --- ## [Nvidia-container-toolkit RPMs won't install on FIPS enable RHEL/Rocky systems](https://forums.developer.nvidia.com/t/nvidia-container-toolkit-rpms-wont-install-on-fips-enable-rhel-rocky-systems/337383)
**Author:** [@aspeagle](https://forums.developer.nvidia.com/u/aspeagle)\ **Replies:** 2\ **Last updated:** [August 25, 2025, 10:20pm UTC](https://forums.developer.nvidia.com/t/nvidia-container-toolkit-rpms-wont-install-on-fips-enable-rhel-rocky-systems/337383 "2025-08-25T22:20:18Z")
We’re running into an issue getting nvidia-container-toolkit RPMs to install on RHEL/Rocky hosts that are setup in FIPS mode. We see this: Upgrading: libnvidia-container-tools x86\_64 1.17.8-1 cuda-rhel8-x… --- ## [Nvidia/cuda:13.0.0-tensorrt-runtime-ubuntu24.04 arm64 only?](https://forums.developer.nvidia.com/t/nvidia-cuda-13-0-0-tensorrt-runtime-ubuntu24-04-arm64-only/342265)
**Author:** [@sheckman](https://forums.developer.nvidia.com/u/sheckman)\ **Replies:** 1\ **Last updated:** [August 16, 2025, 6:12pm UTC](https://forums.developer.nvidia.com/t/nvidia-cuda-13-0-0-tensorrt-runtime-ubuntu24-04-arm64-only/342265 "2025-08-16T18:12:11Z")
On my Ubuntu 25.04 x86 system, when I pull nvidia/cuda:13.0.0-tensorrt-runtime-ubuntu24.04 and run it, I get: $ podman run --rm --arch amd64 --device nvidia.com/gpu=all nvidia/cuda:13.0.0-tensorrt-runtime-ubuntu24.04 … --- ## [Jetson Jetpack 6.1 -\>Nvidia Driver 540.4.0](https://forums.developer.nvidia.com/t/jetson-jetpack-6-1-nvidia-driver-540-4-0/319771)
**Author:** [@erik.ostrov](https://forums.developer.nvidia.com/u/erik.ostrov)\ **Replies:** 6\ **Last updated:** [July 30, 2025, 8:16pm UTC](https://forums.developer.nvidia.com/t/jetson-jetpack-6-1-nvidia-driver-540-4-0/319771 "2025-07-30T20:16:09Z")
HI , does any body know why the latest Jetpack comes with nvidia driver 540.4.0 and the container provided by nvidia nvcr.io/nvidia/deepstream:7.1-triton-multiarch shows this massage when i run the container \[ERROR: This… --- ## [A 4 trillion dollar company shouldn't set low limits on nvcr.io registry pulls](https://forums.developer.nvidia.com/t/a-4-trillion-dollar-company-shouldnt-set-low-limits-on-nvcr-io-registry-pulls/340631)
**Author:** [@ben312](https://forums.developer.nvidia.com/u/ben312)\ **Replies:** 0\ **Last updated:** [July 29, 2025, 3:29pm UTC](https://forums.developer.nvidia.com/t/a-4-trillion-dollar-company-shouldnt-set-low-limits-on-nvcr-io-registry-pulls/340631 "2025-07-29T15:29:46Z")
There’s no reason my once-per-day GitHub Actions building on top nvcr.io registry images should error with ERROR: ... unexpected status ... nvcr.io/.../25.03-py3: 401 Unauthorized after working for a while without auth… --- ## [GPUDirect Async KI ( GDA-KI) QP Creation query](https://forums.developer.nvidia.com/t/gpudirect-async-ki-gda-ki-qp-creation-query/339717)
**Author:** [@alok.prasad7](https://forums.developer.nvidia.com/u/alok.prasad7)\ **Replies:** 0\ **Last updated:** [July 21, 2025, 8:20am UTC](https://forums.developer.nvidia.com/t/gpudirect-async-ki-gda-ki-qp-creation-query/339717 "2025-07-21T08:20:49Z")
I have basic queries regarding QPs in GDA-KI, Diagram in below links shows that QP is created by CPU But looking at the pics at ( Magnum IO GPUDirect, NCCL, NVSHMEM, and GDA-KI on Grace Hopper and Hopper systems -S61… --- ## [Best Practices for Running Multiple Pixel Streaming Instances (Unreal/Unity) on a Single GPU – Real Partitioning with Vulkan/DirectX?](https://forums.developer.nvidia.com/t/best-practices-for-running-multiple-pixel-streaming-instances-unreal-unity-on-a-single-gpu-real-partitioning-with-vulkan-directx/335851)
**Author:** [@daniel505](https://forums.developer.nvidia.com/u/daniel505)\ **Replies:** 0\ **Last updated:** [June 11, 2025, 12:12pm UTC](https://forums.developer.nvidia.com/t/best-practices-for-running-multiple-pixel-streaming-instances-unreal-unity-on-a-single-gpu-real-partitioning-with-vulkan-directx/335851 "2025-06-11T12:12:46Z")
Hello everyone, I’ve spent quite some time researching this topic but would really appreciate some insights from the experts here. Use Case: We have an Unreal Engine project that uses Pixel Streaming to deliver real-t… --- ## [Argmax from CUDA without enable CUDA language](https://forums.developer.nvidia.com/t/argmax-from-cuda-without-enable-cuda-language/335667)
**Author:** [@alexandro.esparza.cortes](https://forums.developer.nvidia.com/u/alexandro.esparza.cortes)\ **Replies:** 0\ **Last updated:** [June 9, 2025, 2:10pm UTC](https://forums.developer.nvidia.com/t/argmax-from-cuda-without-enable-cuda-language/335667 "2025-06-09T14:10:45Z")
I am facing a significant issue in my project. I have been trying to enable CUDA Language support in my project for cross-compilation (host Linux 64 bits target Orin Nano aarch), but after numerous unsuccessful attempts,… --- ## [Cuda container's drivers compatibility fails - System has unsupported display driver / cuda driver combination](https://forums.developer.nvidia.com/t/cuda-containers-drivers-compatibility-fails-system-has-unsupported-display-driver-cuda-driver-combination/335665)
**Author:** [@eshukrun](https://forums.developer.nvidia.com/u/eshukrun)\ **Replies:** 0\ **Last updated:** [June 9, 2025, 1:52pm UTC](https://forums.developer.nvidia.com/t/cuda-containers-drivers-compatibility-fails-system-has-unsupported-display-driver-cuda-driver-combination/335665 "2025-06-09T13:52:06Z")
Hey, I have the following error when starting nvcr.io/nvidia/cuda-dl-base:25.03-cuda12.8-devel-ubuntu24.04 container, but nvidia-smi on the host shows CUDA version is 12.8 What causes that ? How CUDA/display drivers wo… --- ## [NGC Authentication in CI issue](https://forums.developer.nvidia.com/t/ngc-authentication-in-ci-issue/333114)
**Author:** [@not\_a\_bot](https://forums.developer.nvidia.com/u/not_a_bot)\ **Replies:** 2\ **Last updated:** [May 16, 2025, 7:13pm UTC](https://forums.developer.nvidia.com/t/ngc-authentication-in-ci-issue/333114 "2025-05-16T19:13:40Z")
Recently the NGC/Docker interface changed. Previously we could pull Docker images from NGC without authentication; now our CI pipeline fails with an authentication error when running: docker pull nvcr.io/nvidia/base/ubu… --- ## [Is it possible to have one MIG device recognized as one GPU in a docker container?](https://forums.developer.nvidia.com/t/is-it-possible-to-have-one-mig-device-recognized-as-one-gpu-in-a-docker-container/333467)
**Author:** [@koya.kamada](https://forums.developer.nvidia.com/u/koya.kamada)\ **Replies:** 0\ **Last updated:** [May 16, 2025, 8:54am UTC](https://forums.developer.nvidia.com/t/is-it-possible-to-have-one-mig-device-recognized-as-one-gpu-in-a-docker-container/333467 "2025-05-16T08:54:01Z")
I want to have multiple MIG devices be recognized as 1 MIG device = 1 GPU in a mounted container. For example, if we split 1 GPU into 4 MIG devices and mount them in docker, we want to access them as GPU0,1,2,3 respecti… --- ## [How to enable CUDA Minor Version Compatibility for/in a nvidia/cuda docker image](https://forums.developer.nvidia.com/t/how-to-enable-cuda-minor-version-compatibility-for-in-a-nvidia-cuda-docker-image/330205)
**Author:** [@dinhanhx](https://forums.developer.nvidia.com/u/dinhanhx)\ **Replies:** 0\ **Last updated:** [April 14, 2025, 2:46am UTC](https://forums.developer.nvidia.com/t/how-to-enable-cuda-minor-version-compatibility-for-in-a-nvidia-cuda-docker-image/330205 "2025-04-14T02:46:16Z")
I want to run this nvidia/cuda:12.8 image on a cuda 12.4 machine (rtx 3060). docker run -it --gpus all --rm nvidia/cuda:12.8.1-runtime-ubuntu24.04 docker: Error response from daemon: failed to create task for containe… --- ## [Need clarification on nvidia container toolkit 'rpm' install on host versus inside container image](https://forums.developer.nvidia.com/t/need-clarification-on-nvidia-container-toolkit-rpm-install-on-host-versus-inside-container-image/328823)
**Author:** [@mithunms333](https://forums.developer.nvidia.com/u/mithunms333)\ **Replies:** 0\ **Last updated:** [March 31, 2025, 4:04pm UTC](https://forums.developer.nvidia.com/t/need-clarification-on-nvidia-container-toolkit-rpm-install-on-host-versus-inside-container-image/328823 "2025-03-31T16:04:42Z")
Hello, I need one more info clarification regarding the setup of ‘nvidia container toolkit’. I am not sure if this would be the right forum to ask this. Kindly guide appropriate forum: My production runtime is on a Lin… --- ## [Upgrade jetson orin nano cuda 12.6 to 12.8 broke camera operation](https://forums.developer.nvidia.com/t/upgrade-jetson-orin-nano-cuda-12-6-to-12-8-broke-camera-operation/327357)
**Author:** [@nbeser1](https://forums.developer.nvidia.com/u/nbeser1)\ **Replies:** 0\ **Last updated:** [March 17, 2025, 2:10pm UTC](https://forums.developer.nvidia.com/t/upgrade-jetson-orin-nano-cuda-12-6-to-12-8-broke-camera-operation/327357 "2025-03-17T14:10:26Z")
Per Veraj request I am moving this to a new thread. To recap the problem: I teach an advanced computer architecture class were we loan out jetson orin nano’s to the students. We built the system using SDK Manager (jetpa… --- ## [How does the GPUs in WSL2 get surfaced for something like SLURM (Gres)](https://forums.developer.nvidia.com/t/how-does-the-gpus-in-wsl2-get-surfaced-for-something-like-slurm-gres/265973)
**Author:** [@hicotton02](https://forums.developer.nvidia.com/u/hicotton02)\ **Replies:** 1\ **Last updated:** [March 11, 2025, 4:00pm UTC](https://forums.developer.nvidia.com/t/how-does-the-gpus-in-wsl2-get-surfaced-for-something-like-slurm-gres/265973 "2025-03-11T16:00:13Z")
in a traditional Ubuntu install, you can see the nvidia devices in /dev/nvidiaX. In a WSL2 implementation, they do not exist there, and I cant seem to find out how to surface them. as part of a SLURM cluster, if I want t… --- ## [Does 24-2.1 cuBB support VF as FH port?](https://forums.developer.nvidia.com/t/does-24-2-1-cubb-support-vf-as-fh-port/324832)
**Author:** [@pengzhang39](https://forums.developer.nvidia.com/u/pengzhang39)\ **Replies:** 2\ **Last updated:** [February 24, 2025, 6:51am UTC](https://forums.developer.nvidia.com/t/does-24-2-1-cubb-support-vf-as-fh-port/324832 "2025-02-24T06:51:38Z")
Firstly, enable sriov in BIOS. secondly, create two VF port as fronthaul port. but cant found the vf port in RU console log. Details info pls refer to below link: Network card info: root@alpha:~# ethtool -i enp94s0f… --- ## [I9 4070 on Win 11 WSL2 Morpheus Container to ingest PCAPS](https://forums.developer.nvidia.com/t/i9-4070-on-win-11-wsl2-morpheus-container-to-ingest-pcaps/322374)
**Author:** [@bill.alderson](https://forums.developer.nvidia.com/u/bill.alderson)\ **Replies:** 0\ **Last updated:** [February 3, 2025, 8:39pm UTC](https://forums.developer.nvidia.com/t/i9-4070-on-win-11-wsl2-morpheus-container-to-ingest-pcaps/322374 "2025-02-03T20:39:16Z")
Wanting to ingest pcaps into morpheus for cyber analysis. Can’t get morpheus to work inside ubuntu wsl2 with the morpheus runtime container. i have put errors into chat 01 and o3 high and grok2 and it cannot resolve the… --- ## [UBLAS\_STATUS\_NOT\_SUPPORTED when calling cublasGemmStridedBatchedEx" During Grounding DINO Training](https://forums.developer.nvidia.com/t/ublas-status-not-supported-when-calling-cublasgemmstridedbatchedex-during-grounding-dino-training/319946)
**Author:** [@sameer.afzal](https://forums.developer.nvidia.com/u/sameer.afzal)\ **Replies:** 0\ **Last updated:** [January 14, 2025, 2:33pm UTC](https://forums.developer.nvidia.com/t/ublas-status-not-supported-when-calling-cublasgemmstridedbatchedex-during-grounding-dino-training/319946 "2025-01-14T14:33:36Z")
I am encountering an issue while using the NVIDIA TAO Toolkit version 5.5. I followed the instructions in the Grounding DINO sample notebook available at the official NVIDIA GitHub repository: \[Grounding DINO Notebook.\]… --- ## [Looking for torchaudio compatible NGC container](https://forums.developer.nvidia.com/t/looking-for-torchaudio-compatible-ngc-container/317707)
**Author:** [@jmyyri](https://forums.developer.nvidia.com/u/jmyyri)\ **Replies:** 0\ **Last updated:** [December 22, 2024, 6:41pm UTC](https://forums.developer.nvidia.com/t/looking-for-torchaudio-compatible-ngc-container/317707 "2024-12-22T18:41:08Z")
Dear community, I’ve been working on LLM finetuning on the NGC container and look to integrate TTS tools in the setup, unfortunately torchaudio runs into errors if I try to build it from source. ERROR: No matching dis… --- ## [AttributeError: '\_OpNamespace' 'nvidia' object has no attribute 'MultiscaleDeformableAttnPlugin\_TRT'](https://forums.developer.nvidia.com/t/attributeerror-opnamespace-nvidia-object-has-no-attribute-multiscaledeformableattnplugin-trt/317684)
**Author:** [@yonatan.mirsky](https://forums.developer.nvidia.com/u/yonatan.mirsky)\ **Replies:** 0\ **Last updated:** [December 22, 2024, 7:24am UTC](https://forums.developer.nvidia.com/t/attributeerror-opnamespace-nvidia-object-has-no-attribute-multiscaledeformableattnplugin-trt/317684 "2024-12-22T07:24:41Z")
I am trying to train the Mask2Former model using the container from the tao\_pytorch\_backend repository: nvcr.io/nvidia/tao/tao-toolkit:5.5.0-pyt-base. The training runs on a multi-instance GPU setup in an AWS EC2 instanc… --- ## [How many CUDA Cores Present in A100 GPU](https://forums.developer.nvidia.com/t/how-many-cuda-cores-present-in-a100-gpu/317478)
**Author:** [@aravind.chakravarti](https://forums.developer.nvidia.com/u/aravind.chakravarti)\ **Replies:** 0\ **Last updated:** [December 19, 2024, 1:55pm UTC](https://forums.developer.nvidia.com/t/how-many-cuda-cores-present-in-a100-gpu/317478 "2024-12-19T13:55:19Z")
Hello, We currently have RTX A5000 GPU with us, which has 8192 cuda cores. We were investigating purchasing a A100 GPU (80GB) for a NLP inference purpose. Datasheet of A100 specifies only Tensor-cores and didn’t tell a… --- ## [How to launch Nsight Compute/System GUI in a container](https://forums.developer.nvidia.com/t/how-to-launch-nsight-compute-system-gui-in-a-container/312401)
**Author:** [@tnwilly](https://forums.developer.nvidia.com/u/tnwilly)\ **Replies:** 1\ **Last updated:** [November 27, 2024, 9:17pm UTC](https://forums.developer.nvidia.com/t/how-to-launch-nsight-compute-system-gui-in-a-container/312401 "2024-11-27T21:17:58Z")
I want to profile my CUDA kernels in a docker container. It is okay if I do it with non-gui version of nsight compute and system. But if I want to use GUI, they can not be launched. This is my Dockerfile: FROM nvcr.i… --- ## [Container's for Tensorflow and pytorch JP 6.0 R36.3.0](https://forums.developer.nvidia.com/t/containers-for-tensorflow-and-pytorch-jp-6-0-r36-3-0/312658)
**Author:** [@vinayakk28](https://forums.developer.nvidia.com/u/vinayakk28)\ **Replies:** 2\ **Last updated:** [November 14, 2024, 1:28pm UTC](https://forums.developer.nvidia.com/t/containers-for-tensorflow-and-pytorch-jp-6-0-r36-3-0/312658 "2024-11-14T13:28:32Z")
Hi, Do we have containers available for Tensorflow and pytorch or a combined container (Tensorflow+pytorch) which runs on JP 6.0 R36.3.0 Jetson orin-nano with Python3.10 or later? Please do let us know your feedback. T… --- ## [MPS on kubernetes device plugin](https://forums.developer.nvidia.com/t/mps-on-kubernetes-device-plugin/313071)
**Author:** [@vivin1](https://forums.developer.nvidia.com/u/vivin1)\ **Replies:** 0\ **Last updated:** [November 12, 2024, 7:33pm UTC](https://forums.developer.nvidia.com/t/mps-on-kubernetes-device-plugin/313071 "2024-11-12T19:33:25Z")
When using the Kubernetes device plugin, how is the MPS controller daemon managed ? What happens to existing clients if the connectivity to the MPS controller daemon is disrupted ? --- ## [Run containers using multiple CUDA versions (12.4.x, 12.5.x and 12.6.x) using GPU operator in a node with CUDA v12.6](https://forums.developer.nvidia.com/t/run-containers-using-multiple-cuda-versions-12-4-x-12-5-x-and-12-6-x-using-gpu-operator-in-a-node-with-cuda-v12-6/308997)
**Author:** [@rangasamudrala](https://forums.developer.nvidia.com/u/rangasamudrala)\ **Replies:** 1\ **Last updated:** [October 16, 2024, 3:33pm UTC](https://forums.developer.nvidia.com/t/run-containers-using-multiple-cuda-versions-12-4-x-12-5-x-and-12-6-x-using-gpu-operator-in-a-node-with-cuda-v12-6/308997 "2024-10-16T15:33:20Z")
Display card: TU102 \[TITAN RTX\] OS: RHEL 9.4 nvidia-smi output: NVIDIA-SMI 550.90.12 Driver Version: 550.90.12 CUDA Version: 12.4 GPU Operator: v24.6.1 If run a container that has CUDA v12.4 it work… [Next page](https://forums.developer.nvidia.com/c/accelerated-computing/nvidia-gpu-cloud-ngc-users/container-cuda/29.md?page=1)