github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10737] again, ollama is not using all VRAM with newer models, and I think with older ones too. #7051

New Issue
Closed
opened 2026-04-12 18:58:04 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 18:58:04 -05:00
Owner
Copy Link

Originally created by @Fade78 on GitHub (May 16, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10737

What is the issue?

ollama version: 0.7.0

This is Qwen3:32 with a context of 19000 that should go into the VRAM. Already did this. But as you can see, ollama thinks it's 66GB in size and decides to load it in part outside the GPUs. Those three GPUs are dedicated to LLM and there is only one process per GPU. Meanwhile the VRAM is not used at all. I think this is part of another bug I opened and that was closed without investigation.

So what is wrong:

  • Qwen3:32b shouldn't, if memory serves, take 66GB of VRAM. In fact it should go into 48GB.
  • Ollama decide to split the imaginary 66GB 26/74. It means it believes I have 48GB of VRAM. I do.
  • Ollama put the imaginary 48GB load in the VRAM but only half of it is consumed in reality.

Image

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @Fade78 on GitHub (May 16, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10737 ### What is the issue? ollama version: 0.7.0 This is Qwen3:32 with a context of 19000 that should go into the VRAM. Already did this. But as you can see, ollama thinks it's 66GB in size and decides to load it in part outside the GPUs. Those three GPUs are dedicated to LLM and there is only one process per GPU. Meanwhile the VRAM is not used at all. I think this is part of another bug I opened and that was closed without investigation. So what is wrong: - Qwen3:32b shouldn't, if memory serves, take 66GB of VRAM. In fact it should go into 48GB. - Ollama decide to split the imaginary 66GB 26/74. It means it believes I have 48GB of VRAM. I do. - Ollama put the imaginary 48GB load in the VRAM but only half of it is consumed in reality. ![Image](https://github.com/user-attachments/assets/fbe8f03e-114b-489c-bf8d-cfbe855407d5) ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-12 18:58:04 -05:00
GiteaMirror commented 2026-04-12 18:58:05 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 16, 2025):

Server logs may aid in debugging.

<!-- gh-comment-id:2887090583 --> @rick-github commented on GitHub (May 16, 2025): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may aid in debugging.
GiteaMirror commented 2026-04-12 18:58:05 -05:00
Author
Owner
Copy Link

@Fade78 commented on GitHub (May 16, 2025):

0.7.0 doesn't appear here anymore so I close this issue, I'll be back with the 0.6.8 I use now.

<!-- gh-comment-id:2887091601 --> @Fade78 commented on GitHub (May 16, 2025): 0.7.0 doesn't appear here anymore so I close this issue, I'll be back with the 0.6.8 I use now.
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
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#7051
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?