github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11809] Low-inference-speed/ Low-level mistakes by you. #7836

New Issue
Closed
opened 2026-04-12 20:00:17 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-12 20:00:17 -05:00
Owner
Copy Link

Originally created by @MasihMoafi on GitHub (Aug 8, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11809

What is the issue?

So I tried very much NOT to leave like a hateful note, but your app sucks! It's using more than 80% of my RAM whenever I'm running any model like Qwen3 even! This is a bug in the new version. Isn't your app supposed to be made on top of llama.cpp? Then why is it so slow and it's not even using my GPU? Based on the documentation I should be able to get like six tokens per second using the newest GPT version, but I'm getting like one every second or two at most. So, to sum up, you suck.

Relevant log output

You suck.

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.11.0

Originally created by @MasihMoafi on GitHub (Aug 8, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11809 ### What is the issue? So I tried very much NOT to leave like a hateful note, but your app sucks! It's using more than 80% of my RAM whenever I'm running any model like Qwen3 even! This is a bug in the new version. Isn't your app supposed to be made on top of llama.cpp? Then why is it so slow and it's not even using my GPU? Based on the documentation I should be able to get like six tokens per second using the newest GPT version, but I'm getting like one every second or two at most. So, to sum up, you suck. ### Relevant log output ```shell You suck. ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.11.0
GiteaMirror added the bug label 2026-04-12 20:00:17 -05:00
GiteaMirror commented 2026-04-12 20:00:18 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 8, 2025):

Server logs will aid in debugging.

<!-- gh-comment-id:3167692468 --> @rick-github commented on GitHub (Aug 8, 2025): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) will aid in debugging.
GiteaMirror commented 2026-04-12 20:00:19 -05:00
Author
Owner
Copy Link

@MasihMoafi commented on GitHub (Aug 8, 2025):

Thank you. It shows that ollama detects my VRAM as low, (8GB), and then it dumps everything on memory! Is there a low-level control anywhere?!

<!-- gh-comment-id:3168923669 --> @MasihMoafi commented on GitHub (Aug 8, 2025): Thank you. It shows that ollama detects my VRAM as low, (8GB), and then it dumps everything on memory! Is there a low-level control anywhere?!
GiteaMirror commented 2026-04-12 20:00:19 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 8, 2025):

Post logs.

<!-- gh-comment-id:3168926927 --> @rick-github commented on GitHub (Aug 8, 2025): Post logs.
GiteaMirror commented 2026-04-12 20:00:19 -05:00
Author
Owner
Copy Link

@MasihMoafi commented on GitHub (Aug 8, 2025):

server-1.log
server-2.log
server-3.log
server-4.log
server-5.log

<!-- gh-comment-id:3169179006 --> @MasihMoafi commented on GitHub (Aug 8, 2025): [server-1.log](https://github.com/user-attachments/files/21691135/server-1.log) [server-2.log](https://github.com/user-attachments/files/21691138/server-2.log) [server-3.log](https://github.com/user-attachments/files/21691137/server-3.log) [server-4.log](https://github.com/user-attachments/files/21691136/server-4.log) [server-5.log](https://github.com/user-attachments/files/21691139/server-5.log)
GiteaMirror commented 2026-04-12 20:00:20 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 8, 2025):

time=2025-08-06T23:02:44.035+03:30 level=INFO source=server.go:175 msg=offload library=cuda layers.requested=-1
 layers.model=25 layers.offload=0 layers.split="" memory.available="[7.0 GiB]" memory.gpu_overhead="0 B"
 memory.required.full="14.8 GiB" memory.required.partial="0 B" memory.required.kv="3.1 GiB" memory.required.allocations="[0 B]"
 memory.weights.total="11.7 GiB" memory.weights.repeating="10.7 GiB" memory.weights.nonrepeating="1.1 GiB"
 memory.graph.full="32.0 GiB" memory.graph.partial="64.0 GiB"

Context of 131072 is too large. It results in a memory graph of 32G which will not fit in the available 7G of VRAM, so the model is loaded in system RAM. You could reduce OLLAMA_CONTEXT_LENGTH, and then set OLLAMA_FLASH_ATTENTION and OLLAMA_KV_CACHE_TYPE to reduce the memory footprint.

<!-- gh-comment-id:3169322851 --> @rick-github commented on GitHub (Aug 8, 2025): ``` time=2025-08-06T23:02:44.035+03:30 level=INFO source=server.go:175 msg=offload library=cuda layers.requested=-1 layers.model=25 layers.offload=0 layers.split="" memory.available="[7.0 GiB]" memory.gpu_overhead="0 B" memory.required.full="14.8 GiB" memory.required.partial="0 B" memory.required.kv="3.1 GiB" memory.required.allocations="[0 B]" memory.weights.total="11.7 GiB" memory.weights.repeating="10.7 GiB" memory.weights.nonrepeating="1.1 GiB" memory.graph.full="32.0 GiB" memory.graph.partial="64.0 GiB" ``` Context of 131072 is too large. It results in a memory graph of 32G which will not fit in the available 7G of VRAM, so the model is loaded in system RAM. You could reduce [`OLLAMA_CONTEXT_LENGTH`](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size), and then set [`OLLAMA_FLASH_ATTENTION`](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-enable-flash-attention) and [`OLLAMA_KV_CACHE_TYPE`](https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-set-the-quantization-type-for-the-kv-cache) to reduce the memory footprint.
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#7836
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?