github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #1992] CUDA GPU is too old #26907

New Issue
Closed
opened 2026-04-22 03:38:33 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-22 03:38:33 -05:00
Owner
Copy Link

Originally created by @tlaanemaa on GitHub (Jan 14, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/1992

Hello.

First of all, thanks for bringing us this awesome project!

I have a pretty old GPU, Nvidia GTX 970, but it used to work fine with Ollama 0.1.15.
Now I upgraded to 0.1.20 and I get the following error:

2024/01/14 19:50:06 gpu.go:88: Detecting GPU type
2024/01/14 19:50:06 gpu.go:203: Searching for GPU management library libnvidia-ml.so
2024/01/14 19:50:06 gpu.go:248: Discovered GPU libraries: [/usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1]
2024/01/14 19:50:06 gpu.go:94: Nvidia GPU detected
2024/01/14 19:50:06 gpu.go:138: CUDA GPU is too old. Falling back to CPU mode. Compute Capability detected: 5.2
2024/01/14 19:50:06 routes.go:953: no GPU detected

Im running Ollama in docker with GPU pass through and it seems to show up within the container:

root@a84d0bca74d1:/# nvidia-smi
Sun Jan 14 20:03:51 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 545.36                 Driver Version: 546.33       CUDA Version: 12.3     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA GeForce GTX 970         On  | 00000000:01:00.0  On |                  N/A |
| 60%   29C    P8              13W / 151W |    566MiB /  4096MiB |      3%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|  No running processes found                                                           |
+---------------------------------------------------------------------------------------+

I realize my GPU is old, but it used to work.
Do you know if there's a way to make it work again? I'd prefer to not be stuck on 0.1.15, if possible 😅
I'm happy to build the docker image from source, if thats needed.

Thanks in advance!

Originally created by @tlaanemaa on GitHub (Jan 14, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/1992 Hello. First of all, thanks for bringing us this awesome project! I have a pretty old GPU, Nvidia GTX 970, but it used to work fine with Ollama 0.1.15. Now I upgraded to 0.1.20 and I get the following error: ``` 2024/01/14 19:50:06 gpu.go:88: Detecting GPU type 2024/01/14 19:50:06 gpu.go:203: Searching for GPU management library libnvidia-ml.so 2024/01/14 19:50:06 gpu.go:248: Discovered GPU libraries: [/usr/lib/x86_64-linux-gnu/libnvidia-ml.so.1] 2024/01/14 19:50:06 gpu.go:94: Nvidia GPU detected 2024/01/14 19:50:06 gpu.go:138: CUDA GPU is too old. Falling back to CPU mode. Compute Capability detected: 5.2 2024/01/14 19:50:06 routes.go:953: no GPU detected ``` Im running Ollama in docker with GPU pass through and it seems to show up within the container: ``` root@a84d0bca74d1:/# nvidia-smi Sun Jan 14 20:03:51 2024 +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 545.36 Driver Version: 546.33 CUDA Version: 12.3 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce GTX 970 On | 00000000:01:00.0 On | N/A | | 60% 29C P8 13W / 151W | 566MiB / 4096MiB | 3% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ +---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | No running processes found | +---------------------------------------------------------------------------------------+ ``` I realize my GPU is old, but it used to work. Do you know if there's a way to make it work again? I'd prefer to not be stuck on 0.1.15, if possible 😅 I'm happy to build the docker image from source, if thats needed. Thanks in advance!
GiteaMirror commented 2026-04-22 03:38:36 -05:00
Author
Owner
Copy Link

@tlaanemaa commented on GitHub (Jan 14, 2024):

I tried building the docker image locally, it seems to build ollama from source, but still the same :/

<!-- gh-comment-id:1891063755 --> @tlaanemaa commented on GitHub (Jan 14, 2024): I tried building the docker image locally, it seems to build ollama from source, but still the same :/
GiteaMirror commented 2026-04-22 03:38:37 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Jan 14, 2024):

@tlaanemaa sorry about that – we're working on making sure Ollama works with compute capability 5 cards in this issue #1756

<!-- gh-comment-id:1891087649 --> @jmorganca commented on GitHub (Jan 14, 2024): @tlaanemaa sorry about that – we're working on making sure Ollama works with compute capability 5 cards in this issue #1756
GiteaMirror commented 2026-04-22 03:38:37 -05:00
Author
Owner
Copy Link

@zuluion commented on GitHub (May 5, 2024):

GPU is RTX 2060 Super, but still indicates [0] CUDA GPU is too old. Calculated capacity found: 1.0

I installed ollama through scoop, not running it in docker.

GPU driver version: 526.98

<!-- gh-comment-id:2094854311 --> @zuluion commented on GitHub (May 5, 2024): GPU is RTX 2060 Super, but still indicates `[0] CUDA GPU is too old. Calculated capacity found: 1.0` I installed ollama through scoop, not running it in docker. GPU driver version: 526.98
GiteaMirror commented 2026-04-22 03:38:38 -05:00
Author
Owner
Copy Link

@cr1cr1 commented on GitHub (May 6, 2024):

https://github.com/ollama/ollama/issues/4008#issuecomment-2096623526

<!-- gh-comment-id:2096636318 --> @cr1cr1 commented on GitHub (May 6, 2024): https://github.com/ollama/ollama/issues/4008#issuecomment-2096623526
GiteaMirror commented 2026-04-22 03:38:39 -05:00
Author
Owner
Copy Link

@abcbarryn commented on GitHub (Jul 20, 2025):

I am having a similar issue.
INFO source=gpu.go:303 msg="[0] CUDA GPU is too old. Compute Capability detected: 3.5"

<!-- gh-comment-id:3094821686 --> @abcbarryn commented on GitHub (Jul 20, 2025): I am having a similar issue. `INFO source=gpu.go:303 msg="[0] CUDA GPU is too old. Compute Capability detected: 3.5"`
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
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#26907
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?