github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #14357] ollama's NVML queries fail with the 580.xx driver #71390

New Issue
Closed
opened 2026-05-05 01:28:09 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-05 01:28:09 -05:00
Owner
Copy Link

Originally created by @hyperu on GitHub (Feb 22, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14357

The issue is that ollama's NVML queries fail with the 580.xx driver, causing total_vram="0 B" during GPU discovery. This prevents any GPU layer offloading.

What we tried:

  1. ✅ Added LD_LIBRARY_PATH=/usr/lib/ollama to service override
  2. ✅ Added ollama user to video and render groups
  3. ✅ Rebuilt ollama-cuda against CUDA 13.0
  4. ❌ None fixed the VRAM detection

Root cause: Ollama's GPU discovery code uses NVML library calls that don't work properly with the new 580.xx driver (CUDA 13.0). The CUDA backend loads correctly later, but the initial GPU detection fails.

Originally created by @hyperu on GitHub (Feb 22, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14357 The issue is that ollama's NVML queries fail with the 580.xx driver, causing total_vram="0 B" during GPU discovery. This prevents any GPU layer offloading. What we tried: 1. ✅ Added LD_LIBRARY_PATH=/usr/lib/ollama to service override 2. ✅ Added ollama user to video and render groups 3. ✅ Rebuilt ollama-cuda against CUDA 13.0 4. ❌ None fixed the VRAM detection Root cause: Ollama's GPU discovery code uses NVML library calls that don't work properly with the new 580.xx driver (CUDA 13.0). The CUDA backend loads correctly later, but the initial GPU detection fails.
GiteaMirror added the bugnvidianeeds more info labels 2026-05-05 01:28:10 -05:00
GiteaMirror commented 2026-05-05 01:28:16 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 22, 2026):

Works fine on my systems:

| NVIDIA-SMI 580.95.05              Driver Version: 580.95.05      CUDA Version: 13.0     |
level=INFO source=routes.go:1739 msg="vram-based default context" total_vram="31.9 GiB" default_num_ctx=32768

Set OLLAMA_DEBUG=2 in the server environment and post server logs to help in debugging.

<!-- gh-comment-id:3940530298 --> @rick-github commented on GitHub (Feb 22, 2026): Works fine on my systems: ``` | NVIDIA-SMI 580.95.05 Driver Version: 580.95.05 CUDA Version: 13.0 | level=INFO source=routes.go:1739 msg="vram-based default context" total_vram="31.9 GiB" default_num_ctx=32768 ``` Set `OLLAMA_DEBUG=2` in the server environment and post [server logs](https://docs.ollama.com/troubleshooting) to help in debugging.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug needs more info nvidia
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#71390
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?