github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 13:47:08 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #8828] rocm support broken #31484

New Issue
Closed
opened 2026-04-22 11:56:16 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 11:56:16 -05:00
Owner
Copy Link

Originally created by @svalaskevicius on GitHub (Feb 4, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8828

Originally assigned to: @jmorganca on GitHub.

What is the issue?

updated ollama from 0.5.5.git+32bd37ad to 0.5.8.git+65b7ecac (using own built package on arch / using aur pkg as the base)

and lost support for the AMD gpu - it only falls back to CPU now.

possibly interesting lines:

Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.692Z level=INFO source=routes.go:1237 msg="Listening on 127.0.0.1:11434 (version 0.0.0)"
Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.692Z level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.709Z level=WARN source=amd_linux.go:61 msg="ollama recommends running the https://www.amd.com/en/support/linux-drivers" error="amdgpu version file missing: /sys/module/amdgpu/version stat /sys/module/amdgpu/version: no such file or directory"
Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.711Z level=INFO source=amd_linux.go:386 msg="amdgpu is supported" gpu=GPU-e9de5de535a4a488 gpu_type=gfx1100
Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.711Z level=INFO source=types.go:130 msg="inference compute" id=GPU-e9de5de535a4a488 library=rocm variant="" compute=gfx1100 driver=0.0 name=1002:744c total="24.0 GiB" available="23.2 GiB"
Feb 04 21:20:47 nexus ollama[226896]: [GIN] 2025/02/04 - 21:20:47 | 200 |       33.51µs |       127.0.0.1 | HEAD     "/"
Feb 04 21:20:47 nexus ollama[226896]: [GIN] 2025/02/04 - 21:20:47 | 200 |    3.602183ms |       127.0.0.1 | POST     "/api/show"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/ollama/blobs/sha256-28bfdfaeba9f51611c00ed322ba684ce6db076756dbc46643f98a8a748c5199e gpu=GPU-e9de5de535a4a488 parallel=4 available=24927535104 required="2.8 GiB"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=server.go:100 msg="system memory" total="94.0 GiB" free="86.5 GiB" free_swap="64.0 GiB"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=memory.go:356 msg="offload to rocm" layers.requested=-1 layers.model=31 layers.offload=31 layers.split="" memory.available="[23.2 GiB]" memory.gpu_overhead="0 B" memory.required.full="2.8 GiB" memory.required.partial="2.8 GiB" memory.required.kv="240.0 MiB" memory.required.allocations="[2.8 GiB]" memory.weights.total="1.7 GiB" memory.weights.repeating="1.7 GiB" memory.weights.nonrepeating="81.0 MiB" memory.graph.full="480.0 MiB" memory.graph.partial="480.0 MiB"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:376 msg="starting llama server" cmd="/usr/bin/ollama runner --model /home/ollama/blobs/sha256-28bfdfaeba9f51611c00ed322ba684ce6db076756dbc46643f98a8a748c5199e --ctx-size 8192 --batch-size 512 --n-gpu-layers 31 --threads 12 --parallel 4 --port 34327"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=sched.go:449 msg="loaded runners" count=1
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:552 msg="waiting for llama runner to start responding"
Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:586 msg="waiting for server to become available" status="llm server error"

/sys/module/amdgpu/version: no such file or directory is indeed not there,.

but the module is loaded:

❯ lsmod | grep amdgpu
amdgpu              15163392  20

using

❯ uname -a
Linux nexus 6.13.1-arch1-1 #1 SMP PREEMPT_DYNAMIC Sun, 02 Feb 2025 01:02:29 +0000 x86_64 GNU/Linux

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.0.0

Originally created by @svalaskevicius on GitHub (Feb 4, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8828 Originally assigned to: @jmorganca on GitHub. ### What is the issue? updated ollama from `0.5.5.git+32bd37ad` to `0.5.8.git+65b7ecac` (using own built package on arch / using aur pkg as the base) and lost support for the AMD gpu - it only falls back to CPU now. possibly interesting lines: ``` Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.692Z level=INFO source=routes.go:1237 msg="Listening on 127.0.0.1:11434 (version 0.0.0)" Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.692Z level=INFO source=gpu.go:217 msg="looking for compatible GPUs" Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.709Z level=WARN source=amd_linux.go:61 msg="ollama recommends running the https://www.amd.com/en/support/linux-drivers" error="amdgpu version file missing: /sys/module/amdgpu/version stat /sys/module/amdgpu/version: no such file or directory" Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.711Z level=INFO source=amd_linux.go:386 msg="amdgpu is supported" gpu=GPU-e9de5de535a4a488 gpu_type=gfx1100 Feb 04 21:19:03 nexus ollama[226896]: time=2025-02-04T21:19:03.711Z level=INFO source=types.go:130 msg="inference compute" id=GPU-e9de5de535a4a488 library=rocm variant="" compute=gfx1100 driver=0.0 name=1002:744c total="24.0 GiB" available="23.2 GiB" Feb 04 21:20:47 nexus ollama[226896]: [GIN] 2025/02/04 - 21:20:47 | 200 | 33.51µs | 127.0.0.1 | HEAD "/" Feb 04 21:20:47 nexus ollama[226896]: [GIN] 2025/02/04 - 21:20:47 | 200 | 3.602183ms | 127.0.0.1 | POST "/api/show" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=sched.go:714 msg="new model will fit in available VRAM in single GPU, loading" model=/home/ollama/blobs/sha256-28bfdfaeba9f51611c00ed322ba684ce6db076756dbc46643f98a8a748c5199e gpu=GPU-e9de5de535a4a488 parallel=4 available=24927535104 required="2.8 GiB" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=server.go:100 msg="system memory" total="94.0 GiB" free="86.5 GiB" free_swap="64.0 GiB" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.556Z level=INFO source=memory.go:356 msg="offload to rocm" layers.requested=-1 layers.model=31 layers.offload=31 layers.split="" memory.available="[23.2 GiB]" memory.gpu_overhead="0 B" memory.required.full="2.8 GiB" memory.required.partial="2.8 GiB" memory.required.kv="240.0 MiB" memory.required.allocations="[2.8 GiB]" memory.weights.total="1.7 GiB" memory.weights.repeating="1.7 GiB" memory.weights.nonrepeating="81.0 MiB" memory.graph.full="480.0 MiB" memory.graph.partial="480.0 MiB" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:376 msg="starting llama server" cmd="/usr/bin/ollama runner --model /home/ollama/blobs/sha256-28bfdfaeba9f51611c00ed322ba684ce6db076756dbc46643f98a8a748c5199e --ctx-size 8192 --batch-size 512 --n-gpu-layers 31 --threads 12 --parallel 4 --port 34327" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=sched.go:449 msg="loaded runners" count=1 Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:552 msg="waiting for llama runner to start responding" Feb 04 21:20:47 nexus ollama[226896]: time=2025-02-04T21:20:47.559Z level=INFO source=server.go:586 msg="waiting for server to become available" status="llm server error" ``` `/sys/module/amdgpu/version: no such file or directory` is indeed not there,. but the module is loaded: ``` ❯ lsmod | grep amdgpu amdgpu 15163392 20 ``` using ``` ❯ uname -a Linux nexus 6.13.1-arch1-1 #1 SMP PREEMPT_DYNAMIC Sun, 02 Feb 2025 01:02:29 +0000 x86_64 GNU/Linux ``` ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.0.0
GiteaMirror added the bug label 2026-04-22 11:56:16 -05:00
GiteaMirror commented 2026-04-22 11:56:18 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Feb 5, 2025):

@svalaskevicius sorry about this. Did you build the ROCM libraries? https://github.com/ollama/ollama/blob/main/docs/development.md#linux

The ollama binary will look for the libraries under:

  • ../lib/ollama
  • build/lib/ollama

relative to the ollama binary. Let me know if this doesn't help!

<!-- gh-comment-id:2635593152 --> @jmorganca commented on GitHub (Feb 5, 2025): @svalaskevicius sorry about this. Did you build the ROCM libraries? https://github.com/ollama/ollama/blob/main/docs/development.md#linux The `ollama` binary will look for the libraries under: * `../lib/ollama` * `build/lib/ollama` relative to the `ollama binary`. Let me know if this doesn't help!
GiteaMirror commented 2026-04-22 11:56:18 -05:00
Author
Owner
Copy Link

@svalaskevicius commented on GitHub (Feb 5, 2025):

hi, thanks,

right, I made a package and didn't copy those libs!

it is working ok now :)

<!-- gh-comment-id:2636095666 --> @svalaskevicius commented on GitHub (Feb 5, 2025): hi, thanks, right, I made a package and didn't copy those libs! it is working ok 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#31484
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?