github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10867] qwen3:30b-a3b has poor performance on WSL2 #32900

New Issue
Closed
opened 2026-04-22 14:49:44 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-04-22 14:49:44 -05:00
Owner
Copy Link

Originally created by @junzhang-bjtu on GitHub (May 26, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10867

What is the issue?

qwen3:30b-a3b can't even run at about 1 token/s on WSL2, but it can run at about 15 tokens/s on Windows.

I just use CPU.

Moreover, after loading qwen3:30b-a3b on WSL2, it consumed more than 40GB of memory, while it only consumed about 20GB of memory on Windows.

Not only qwen3:30b-a3b, but also qwen3:8b has poor performance.

OS

Windows 11 24H2

WSL2 Ubuntu-24.04

GPU

Intel(R) Arc(TM) 140T

CPU

Intel(R) Core(TM) Ultra 9 285H

Memory

2 x 48G 5600MT/s

Ollama version

0.7.1

Originally created by @junzhang-bjtu on GitHub (May 26, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10867 ### What is the issue? qwen3:30b-a3b can't even run at about 1 token/s on WSL2, but it can run at about 15 tokens/s on Windows. I just use CPU. Moreover, after loading qwen3:30b-a3b on WSL2, it consumed more than 40GB of memory, while it only consumed about 20GB of memory on Windows. Not only qwen3:30b-a3b, but also qwen3:8b has poor performance. ### OS Windows 11 24H2 WSL2 Ubuntu-24.04 ### GPU Intel(R) Arc(TM) 140T ### CPU Intel(R) Core(TM) Ultra 9 285H ### Memory 2 x 48G 5600MT/s ### Ollama version 0.7.1
GiteaMirror added the bug label 2026-04-22 14:49:44 -05:00
GiteaMirror commented 2026-04-22 14:49:49 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 26, 2025):

Have you made the GPU available inside WSL?

https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2023-0/configure-wsl-2-for-gpu-workflows.html

<!-- gh-comment-id:2910158920 --> @rick-github commented on GitHub (May 26, 2025): Have you made the GPU available inside WSL? https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2023-0/configure-wsl-2-for-gpu-workflows.html
GiteaMirror commented 2026-04-22 14:49:50 -05:00
Author
Owner
Copy Link

@junzhang-bjtu commented on GitHub (May 26, 2025):

Have you made the GPU available inside WSL?

https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2023-0/configure-wsl-2-for-gpu-workflows.html

I don't have a graphics card, so I can only use the CPU.

<!-- gh-comment-id:2910163346 --> @junzhang-bjtu commented on GitHub (May 26, 2025): > Have you made the GPU available inside WSL? > > https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2023-0/configure-wsl-2-for-gpu-workflows.html I don't have a graphics card, so I can only use the CPU.
GiteaMirror commented 2026-04-22 14:49:51 -05:00
Author
Owner
Copy Link

@scscgit commented on GitHub (May 26, 2025):

@MXS-Jun From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T

<!-- gh-comment-id:2910713165 --> @scscgit commented on GitHub (May 26, 2025): @MXS-Jun From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T
GiteaMirror commented 2026-04-22 14:49:52 -05:00
Author
Owner
Copy Link

@junzhang-bjtu commented on GitHub (May 26, 2025):

@MXS-Jun From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T

OK, I will try it later. Thanks.

<!-- gh-comment-id:2910718448 --> @junzhang-bjtu commented on GitHub (May 26, 2025): > [@MXS-Jun](https://github.com/MXS-Jun) From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T OK, I will try it later. Thanks.
GiteaMirror commented 2026-04-22 14:49:54 -05:00
Author
Owner
Copy Link

@junzhang-bjtu commented on GitHub (May 27, 2025):

@MXS-Jun From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T

I have installed the things below in WSL2, but it's useless.

oneAPI

Intel GPU Driver

And I found that both dense models and MoE models have poor performance.

It's weird. Because it doesn't have problems in Ollama 0.6.6.

<!-- gh-comment-id:2910891760 --> @junzhang-bjtu commented on GitHub (May 27, 2025): > [@MXS-Jun](https://github.com/MXS-Jun) From the context it seems he meant the integrated GPU Intel(R) Arc(TM) 140T I have installed the things below in WSL2, but it's useless. [oneAPI](https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2025-1/base-apt.html#BASE-[APT](https://www.intel.com/content/www/us/en/docs/oneapi/installation-guide-linux/2025-1/base-apt.html#BASE-APT)) [Intel GPU Driver](https://dgpu-docs.intel.com/driver/client/overview.html) And I found that both dense models and MoE models have poor performance. It's weird. Because it doesn't have problems in Ollama 0.6.6.
GiteaMirror commented 2026-04-22 14:49:55 -05:00
Author
Owner
Copy Link

@junzhang-bjtu commented on GitHub (May 27, 2025):

I have changed to ollama-ipex-llm whose version is 2.2.0.

It bases on Ollama 0.6.2.

Everything is fine.

<!-- gh-comment-id:2910929807 --> @junzhang-bjtu commented on GitHub (May 27, 2025): I have changed to ollama-ipex-llm whose version is 2.2.0. It bases on Ollama 0.6.2. Everything is fine.
GiteaMirror commented 2026-04-22 14:49:56 -05:00
Author
Owner
Copy Link

@junzhang-bjtu commented on GitHub (Jun 6, 2025):

If you don't have a dedicated graphics card and only have Intel's integrated graphics, please prioritize using ollama-ipex-llm instead of the official ollama.

ipex-llm

<!-- gh-comment-id:2947810053 --> @junzhang-bjtu commented on GitHub (Jun 6, 2025): If you don't have a dedicated graphics card and only have Intel's integrated graphics, please prioritize using ollama-ipex-llm instead of the official ollama. [ipex-llm](https://github.com/intel/ipex-llm)
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#32900
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?