This website uses Cookies. Click Accept to agree to our website's cookie use as described in our Privacy Policy. Click Preferences to customize your cookie settings.
Accept
Reject
Preferences
Log in to ask a question
Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.
Log in to ask a question

Updating Cloudrun gpu driver

Posted on 01-22-2025 03:39 PM
AliFlowjin
Bronze 2
Bronze 2
Reply posted on --/--/---- --:-- AM

Hi
I am currently using Cloudrun with GPU for video processing with ffmpeg 
I want to update my ffmpeg to support more encoding but the newer ffmpeg require newer nvidia driver version which is not currently offered by google cloud
has anyone tried upgrading the nvidia driver with the docker?

2 5 470
Topic Labels
5 REPLIES 5

Posted on --/--/---- --:-- AM
mcbsalceda
Staff
Reply posted on --/--/---- --:-- AM

Hi @AliFlowjin,

Welcome to Google Cloud Community!

You can use one L4 GPU per Cloud Run instance. An L4 GPU has the following pre-installed drivers:

  • The current NVIDIA driver version: 535.129.03 (CUDA 12.2)

By default, all of the NVIDIA L4 driver libraries are mounted under /usr/local/nvidia/lib64. Cloud Run automatically appends this path to the LD_LIBRARY_PATH environment variable (i.e. ${LD_LIBRARY_PATH}:/usr/local/nvidia/lib64) of the container with the GPU. This allows the dynamic linker to find the NVIDIA driver libraries. The linker searches and resolves paths in the order you list in the LD_LIBRARY_PATH environment variable. Any values you specify in this variable take precedence over the default Cloud Run driver libraries path /usr/local/nvidia/lib64.

If you want to use a CUDA version greater than 12.2, the easiest way is to depend on a newer NVIDIA base image with forward compatibility packages already installed. Another option is to manually install the NVIDIA forward compatibility packages and add them to LD_LIBRARY_PATH. Consult NVIDIA's compatibility matrix to determine which CUDA versions are forward compatible with the provided NVIDIA driver version (535.129.03).

For the complete documentation, you may refer to GPU support for services.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.

Posted on --/--/---- --:-- AM
AliFlowjin
Bronze 2
Bronze 2
In response to mcbsalceda
Reply posted on --/--/---- --:-- AM

Hi mcbsalceda
Yes I read the GPU documentation page fully and I know all the current version
but in the page and your reply there is no mention of how I can update the nvidia driver which is an old driver to a newer version

0 Likes

Posted on --/--/---- --:-- AM
knet
Staff
Reply posted on --/--/---- --:-- AM

Hi, I just came across your question in my inbox. Is there a specific feature you need from the newer driver? You should be able to use newer versions of CUDA even with the older driver because of CUDA forward compatibility.

 
0 Likes

Posted on --/--/---- --:-- AM
AliFlowjin
Bronze 2
Bronze 2
In response to knet
Reply posted on --/--/---- --:-- AM

Hi
Yes we want to use newer version of FFmpeg with gpu and it is dependent on nvidia driver version and cuda version
we want the ffmpeg video features that needs cuda 13 which needs nvidia driver version more than 550 
https://github.com/rigaya/NVEnc/releases

0 Likes

Posted on --/--/---- --:-- AM
jcruel
Bronze 1
Bronze 1
Reply posted on --/--/---- --:-- AM

It seems impossible to upgrade Cuda drivers and Nvidia Container Toolkit in Cloud Run. The only alternative seems to be GKE and maybe GCP Batch.

Top Solution Authors
View all