github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 13:19:40 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #11534] why ollama use high CPU when all model already in GPU 100% ? #54124

New Issue
Closed
opened 2026-04-29 05:15:07 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-29 05:15:07 -05:00
Owner
Copy Link

Originally created by @indogood1 on GitHub (Jul 26, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11534

What is the issue?

ollama version is 0.9.4-rc2

Image [server.log](https://github.com/user-attachments/files/21442235/server.log)

Thank you very much,

Relevant log output

Sat Jul 26 14:32:52 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 572.42                 Driver Version: 572.42         CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                  Driver-Model | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA CMP 70HX              WDDM  |   00000000:10:00.0 Off |                  N/A |
| 67%   62C    P0             81W /  220W |    5216MiB /   8192MiB |     38%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
|   1  NVIDIA P104-100              WDDM  |   00000000:1C:00.0 Off |                  N/A |
| 56%   56C    P0             41W /   90W |    2660MiB /   8192MiB |     38%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
|   2  NVIDIA P104-100              WDDM  |   00000000:2E:00.0 Off |                  N/A |
| 64%   63C    P0             37W /   90W |    2818MiB /   8192MiB |     23%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A            7596      C   ...breOffice\program\soffice.bin      N/A      |
|    0   N/A  N/A            7772      C   ...al\Programs\Ollama\ollama.exe      N/A      |
|    1   N/A  N/A            7772      C   ...al\Programs\Ollama\ollama.exe      N/A      |
|    1   N/A  N/A           18544      C   ...al\Programs\Ollama\ollama.exe      N/A      |
|    2   N/A  N/A            7772      C   ...al\Programs\Ollama\ollama.exe      N/A      |
|    2   N/A  N/A           18544      C   ...al\Programs\Ollama\ollama.exe      N/A      |
+-----------------------------------------------------------------------------------------+

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.9.4-rc2

Originally created by @indogood1 on GitHub (Jul 26, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11534 ### What is the issue? ollama version is 0.9.4-rc2 <img width="1033" height="95" alt="Image" src="https://github.com/user-attachments/assets/737cac63-47ba-499f-bcb8-4de9c14c49d2" /> [server.log](https://github.com/user-attachments/files/21442235/server.log) Thank you very much, ### Relevant log output ```shell Sat Jul 26 14:32:52 2025 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 572.42 Driver Version: 572.42 CUDA Version: 12.8 | |-----------------------------------------+------------------------+----------------------+ | GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA CMP 70HX WDDM | 00000000:10:00.0 Off | N/A | | 67% 62C P0 81W / 220W | 5216MiB / 8192MiB | 38% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ | 1 NVIDIA P104-100 WDDM | 00000000:1C:00.0 Off | N/A | | 56% 56C P0 41W / 90W | 2660MiB / 8192MiB | 38% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ | 2 NVIDIA P104-100 WDDM | 00000000:2E:00.0 Off | N/A | | 64% 63C P0 37W / 90W | 2818MiB / 8192MiB | 23% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+ +-----------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=========================================================================================| | 0 N/A N/A 7596 C ...breOffice\program\soffice.bin N/A | | 0 N/A N/A 7772 C ...al\Programs\Ollama\ollama.exe N/A | | 1 N/A N/A 7772 C ...al\Programs\Ollama\ollama.exe N/A | | 1 N/A N/A 18544 C ...al\Programs\Ollama\ollama.exe N/A | | 2 N/A N/A 7772 C ...al\Programs\Ollama\ollama.exe N/A | | 2 N/A N/A 18544 C ...al\Programs\Ollama\ollama.exe N/A | +-----------------------------------------------------------------------------------------+ ``` ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.9.4-rc2
GiteaMirror added the bug label 2026-04-29 05:15:07 -05:00
GiteaMirror commented 2026-04-29 05:15:09 -05:00
Author
Owner
Copy Link

@indogood1 commented on GitHub (Jul 26, 2025):

server.log

<!-- gh-comment-id:3121455126 --> @indogood1 commented on GitHub (Jul 26, 2025): [server.log](https://github.com/user-attachments/files/21442239/server.log)
GiteaMirror commented 2026-04-29 05:15:09 -05:00
Author
Owner
Copy Link

@mingoal commented on GitHub (Jul 27, 2025):

same here.
It seems vscode hands over something to ollama and ollama needs to handle that with CPU, then use GPU to handle something.
And it seems no cache between two input messages, ollama always needs to redo it.

If send the message directly to ollama on console, no high CPU on ollama is observed.

<!-- gh-comment-id:3123846757 --> @mingoal commented on GitHub (Jul 27, 2025): same here. It seems vscode hands over something to ollama and ollama needs to handle that with CPU, then use GPU to handle something. And it seems no cache between two input messages, ollama always needs to redo it. If send the message directly to ollama on console, no high CPU on ollama is observed.
GiteaMirror commented 2026-04-29 05:15:10 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 27, 2025):

@indogood1 While the GPUs are doing inference, the CPU is the controller - it is busy sending instructions to the GPUs on what kernels to launch, moving intermediate results between the GPUs, etc. So the CPU will still be busy even when the GPUs are the only devices processing the model weights.

@mingoal VSCode may be setting a large context buffer, causing the model to spill to RAM where CPU does inference.

<!-- gh-comment-id:3124118622 --> @rick-github commented on GitHub (Jul 27, 2025): @indogood1 While the GPUs are doing inference, the CPU is the controller - it is busy sending instructions to the GPUs on what kernels to launch, moving intermediate results between the GPUs, etc. So the CPU will still be busy even when the GPUs are the only devices processing the model weights. @mingoal VSCode may be setting a large context buffer, causing the model to spill to RAM where CPU does inference.
GiteaMirror commented 2026-04-29 05:15:10 -05:00
Author
Owner
Copy Link

@indogood1 commented on GitHub (Jul 31, 2025):

so i think it is not a problem.. and normal.

<!-- gh-comment-id:3139425925 --> @indogood1 commented on GitHub (Jul 31, 2025): so i think it is not a problem.. and normal.
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#54124
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?