github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 21:55:43 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #8710] amdgpu is not supported #5646

New Issue
Closed
opened 2026-04-12 16:56:13 -05:00 by GiteaMirror · 12 comments
GiteaMirror commented 2026-04-12 16:56:13 -05:00
Owner
Copy Link

Originally created by @c1tt1 on GitHub (Jan 30, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8710

What is the issue?

Here is the output:

2025/01/30 20:47:41 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\Users\myuser\.ollama\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]"
time=2025-01-30T20:47:41.671Z level=INFO source=images.go:432 msg="total blobs: 5"
time=2025-01-30T20:47:41.671Z level=INFO source=images.go:439 msg="total unused blobs removed: 0"
time=2025-01-30T20:47:41.672Z level=INFO source=routes.go:1238 msg="Listening on 127.0.0.1:11434 (version 0.5.7)"
time=2025-01-30T20:47:41.674Z level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]"
time=2025-01-30T20:47:41.674Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs"
time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:167 msg=packages count=1
time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=16
time=2025-01-30T20:47:42.301Z level=WARN source=amd_windows.go:140 msg="amdgpu is not supported (supported types:[gfx1030 gfx1100 gfx1101 gfx1102 gfx906])" gpu_type=gfx1103 gpu=0 library=C:\Users\myuser\AppData\Local\Programs\Ollama\lib\ollama
time=2025-01-30T20:47:42.303Z level=INFO source=types.go:131 msg="inference compute" id=GPU-92c3d56e-9c9f-ce93-24da-a319b3f4fb47 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4050 Laptop GPU" total="6.0 GiB" available="5.0 GiB"

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.5.7

Originally created by @c1tt1 on GitHub (Jan 30, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8710 ### What is the issue? Here is the output: 2025/01/30 20:47:41 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\myuser\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]" time=2025-01-30T20:47:41.671Z level=INFO source=images.go:432 msg="total blobs: 5" time=2025-01-30T20:47:41.671Z level=INFO source=images.go:439 msg="total unused blobs removed: 0" time=2025-01-30T20:47:41.672Z level=INFO source=routes.go:1238 msg="Listening on 127.0.0.1:11434 (version 0.5.7)" time=2025-01-30T20:47:41.674Z level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]" time=2025-01-30T20:47:41.674Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs" time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:167 msg=packages count=1 time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=16 time=2025-01-30T20:47:42.301Z level=WARN source=amd_windows.go:140 msg="amdgpu is not supported (supported types:[gfx1030 gfx1100 gfx1101 gfx1102 gfx906])" gpu_type=gfx1103 gpu=0 library=C:\Users\myuser\AppData\Local\Programs\Ollama\lib\ollama time=2025-01-30T20:47:42.303Z level=INFO source=types.go:131 msg="inference compute" id=GPU-92c3d56e-9c9f-ce93-24da-a319b3f4fb47 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4050 Laptop GPU" total="6.0 GiB" available="5.0 GiB" ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.5.7
GiteaMirror added the bug label 2026-04-12 16:56:13 -05:00
GiteaMirror commented 2026-04-12 16:56:14 -05:00
Author
Owner
Copy Link

@philippstoboy commented on GitHub (Jan 30, 2025):

It looks like Ollama detected an AMD GPU (gfx1103), but this architecture is not supported. The currently supported AMD architectures are gfx1030, gfx1100, gfx1101, gfx1102, and gfx906 (List of Supported AMD GPUs).

However, the logs confirm that a NVIDIA GeForce RTX 4050 Laptop GPU was detected and initialized with CUDA (v12). If Ollama is defaulting to the CPU instead of using the NVIDIA GPU, you may need to explicitly force CUDA usage.

Possible Fix:

Try setting the environment variable to ensure Ollama uses CUDA:

set OLLAMA_LLM_LIBRARY=cuda
<!-- gh-comment-id:2625649802 --> @philippstoboy commented on GitHub (Jan 30, 2025): It looks like Ollama detected an **AMD GPU (gfx1103)**, but this architecture is **not supported**. The currently supported AMD architectures are `gfx1030`, `gfx1100`, `gfx1101`, `gfx1102`, and `gfx906` ([List of Supported AMD GPUs](https://github.com/ollama/ollama/blob/main/docs/gpu.md#amd-radeon)). However, the logs confirm that a **NVIDIA GeForce RTX 4050 Laptop GPU** was detected and initialized with CUDA (`v12`). If Ollama is defaulting to the CPU instead of using the NVIDIA GPU, you may need to explicitly force CUDA usage. ### Possible Fix: Try setting the environment variable to ensure Ollama uses CUDA: ```sh set OLLAMA_LLM_LIBRARY=cuda ```
GiteaMirror commented 2026-04-12 16:56:15 -05:00
Author
Owner
Copy Link

@mxyng commented on GitHub (Jan 30, 2025):

What exactly is the issue? Are you not able to run Ollama with hardware acceleration?

The warning about amdgpu likely relates to the integrated GPU found on your CPU. It does detect the NVIDIA GPU which will be used for inference.

Your logs is incomplete so it's unclear where your issue lies.

<!-- gh-comment-id:2625656391 --> @mxyng commented on GitHub (Jan 30, 2025): What exactly is the issue? Are you not able to run Ollama with hardware acceleration? The warning about amdgpu likely relates to the integrated GPU found on your CPU. It does detect the NVIDIA GPU which will be used for inference. Your logs is incomplete so it's unclear where your issue lies.
GiteaMirror commented 2026-04-12 16:56:15 -05:00
Author
Owner
Copy Link

@waltercool commented on GitHub (Jan 30, 2025):

It is supported, but what's your GPU?

Your bug report doesn't say anything useful about your system.

Like @mxyng says, if you are using a iGPU, then you will have issues unless you force it. This is by design of AMD.

<!-- gh-comment-id:2625792172 --> @waltercool commented on GitHub (Jan 30, 2025): It is supported, but what's your GPU? Your bug report doesn't say anything useful about your system. Like @mxyng says, if you are using a iGPU, then you will have issues unless you force it. This is by design of AMD.
GiteaMirror commented 2026-04-12 16:56:16 -05:00
Author
Owner
Copy Link

@c1tt1 commented on GitHub (Jan 31, 2025):

@philippstoboy

I set the following before running but still encounter the same issue:

set OLLAMA_LLM_LIBRARY=cuda
<!-- gh-comment-id:2627046007 --> @c1tt1 commented on GitHub (Jan 31, 2025): @philippstoboy I set the following before running but still encounter the same issue: ```sh set OLLAMA_LLM_LIBRARY=cuda ```
GiteaMirror commented 2026-04-12 16:56:17 -05:00
Author
Owner
Copy Link

@c1tt1 commented on GitHub (Jan 31, 2025):

This is my GPU

AMD Radeon(TM) 780M
NVIDIA GeForce RTX 4050 Laptop GPU

Image

<!-- gh-comment-id:2627058871 --> @c1tt1 commented on GitHub (Jan 31, 2025): This is my GPU AMD Radeon(TM) 780M NVIDIA GeForce RTX 4050 Laptop GPU ![Image](https://github.com/user-attachments/assets/7e215bd6-5cf0-4af4-aca5-245a57bb9eba)
GiteaMirror commented 2026-04-12 16:56:17 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 31, 2025):

I set the following before running but still encounter the same issue:

What exactly is the issue?

AMD Radeon(TM) 780M is unsupported and will not be used.
NVIDIA GeForce RTX 4050 Laptop GPU is supported, is detected, and will be used.

If you run:

ollama run llama3.2 hello

the model should be downloaded and will respond with a greeting. Does that not work? If it doesn't, please include a full log.

<!-- gh-comment-id:2627127812 --> @rick-github commented on GitHub (Jan 31, 2025): > I set the following before running but still encounter the same issue: What exactly is the issue? AMD Radeon(TM) 780M is unsupported and will not be used. NVIDIA GeForce RTX 4050 Laptop GPU is supported, is detected, and will be used. If you run: ``` ollama run llama3.2 hello ``` the model should be downloaded and will respond with a greeting. Does that not work? If it doesn't, please include a full log.
GiteaMirror commented 2026-04-12 16:56:18 -05:00
Author
Owner
Copy Link

@c1tt1 commented on GitHub (Jan 31, 2025):

It's all good if run:

ollama run llama3.2 hello

But the issue stems from:

ollama serve
<!-- gh-comment-id:2627342941 --> @c1tt1 commented on GitHub (Jan 31, 2025): It's all good if run: ```sh ollama run llama3.2 hello ``` But the issue stems from: ```sh ollama serve ```
GiteaMirror commented 2026-04-12 16:56:18 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 31, 2025):

But the issue stems from:

ollama serve

What issue?

<!-- gh-comment-id:2627511029 --> @rick-github commented on GitHub (Jan 31, 2025): > But the issue stems from: > > ollama serve **What issue?**
GiteaMirror commented 2026-04-12 16:56:19 -05:00
Author
Owner
Copy Link

@c1tt1 commented on GitHub (Jan 31, 2025):

But the issue stems from:
ollama serve

What issue?

The ollama serve returns the output in the issue like the following

2025/01/30 20:47:41 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\Users\myuser.ollama\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]"
time=2025-01-30T20:47:41.671Z level=INFO source=images.go:432 msg="total blobs: 5"
time=2025-01-30T20:47:41.671Z level=INFO source=images.go:439 msg="total unused blobs removed: 0"
time=2025-01-30T20:47:41.672Z level=INFO source=routes.go:1238 msg="Listening on 127.0.0.1:11434 (version 0.5.7)"
time=2025-01-30T20:47:41.674Z level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]"
time=2025-01-30T20:47:41.674Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs"
time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:167 msg=packages count=1
time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=16
time=2025-01-30T20:47:42.301Z level=WARN source=amd_windows.go:140 msg="amdgpu is not supported (supported types:[gfx1030 gfx1100 gfx1101 gfx1102 gfx906])" gpu_type=gfx1103 gpu=0 library=C:\Users\myuser\AppData\Local\Programs\Ollama\lib\ollama
time=2025-01-30T20:47:42.303Z level=INFO source=types.go:131 msg="inference compute" id=GPU-92c3d56e-9c9f-ce93-24da-a319b3f4fb47 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4050 Laptop GPU" total="6.0 GiB" available="5.0 GiB"

<!-- gh-comment-id:2627764856 --> @c1tt1 commented on GitHub (Jan 31, 2025): > > But the issue stems from: > > ollama serve > > **What issue?** The `ollama serve` returns the output in the issue like the following 2025/01/30 20:47:41 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\Users\myuser\.ollama\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]" time=2025-01-30T20:47:41.671Z level=INFO source=images.go:432 msg="total blobs: 5" time=2025-01-30T20:47:41.671Z level=INFO source=images.go:439 msg="total unused blobs removed: 0" time=2025-01-30T20:47:41.672Z level=INFO source=routes.go:1238 msg="Listening on 127.0.0.1:11434 (version 0.5.7)" time=2025-01-30T20:47:41.674Z level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]" time=2025-01-30T20:47:41.674Z level=INFO source=gpu.go:226 msg="looking for compatible GPUs" time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:167 msg=packages count=1 time=2025-01-30T20:47:41.674Z level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=16 time=2025-01-30T20:47:42.301Z level=WARN source=amd_windows.go:140 msg="amdgpu is not supported (supported types:[gfx1030 gfx1100 gfx1101 gfx1102 gfx906])" gpu_type=gfx1103 gpu=0 library=C:\Users\myuser\AppData\Local\Programs\Ollama\lib\ollama time=2025-01-30T20:47:42.303Z level=INFO source=types.go:131 msg="inference compute" id=GPU-92c3d56e-9c9f-ce93-24da-a319b3f4fb47 library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4050 Laptop GPU" total="6.0 GiB" available="5.0 GiB"
GiteaMirror commented 2026-04-12 16:56:19 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 31, 2025):

These are logs. Your issue is that the server is telling you what it is doing?

<!-- gh-comment-id:2627770530 --> @rick-github commented on GitHub (Jan 31, 2025): These are logs. Your issue is that the server is telling you what it is doing?
GiteaMirror commented 2026-04-12 16:56:20 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 31, 2025):

You are running Windows, you shouldn't need to start the server manually with ollama serve. Go to the Windows task bar (whatever it's called, it probably has something "Type here to search" in it), type "ollama", click on "Open". The server logs will be hidden.

<!-- gh-comment-id:2627787053 --> @rick-github commented on GitHub (Jan 31, 2025): You are running Windows, you shouldn't need to start the server manually with `ollama serve`. Go to the Windows task bar (whatever it's called, it probably has something "Type here to search" in it), type "ollama", click on "Open". The server logs will be hidden.
GiteaMirror commented 2026-04-12 16:56:21 -05:00
Author
Owner
Copy Link

@c1tt1 commented on GitHub (Jan 31, 2025):

All resolved.

This issue can be closed.

<!-- gh-comment-id:2627790675 --> @c1tt1 commented on GitHub (Jan 31, 2025): All resolved. This issue can be closed.
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#5646
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?