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 #8762] AMD APU support degraded on Docker with new build scripts. #67744

New Issue
Closed
opened 2026-05-04 11:32:49 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-04 11:32:49 -05:00
Owner
Copy Link

Originally created by @rjmalagon on GitHub (Feb 1, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8762

What is the issue?

Somehow, the new cmake build innovations affect ROCM on AMD APUs. Before this change, I was capable of docker building for AMD APUS with few changes (ROCM version to 6.2.4), just before this commit, it works fine.

It's built now, but ROCM fails during execution.
I acknowledge that I have some trouble eye parsing this new dockerfile and cmake files, and maybe I am using the wrong docker build arguments.

Docker build command:
docker build -f Dockerfile --no-cache --build-arg TARGETARCH=amd64 --build-arg FLAVOR=rocm -t localhost/ollama-rocm:latest

About my hardware.
AMD Ryzen 7 7735HS with AMD RAdeon 680M APU, 96GB shared RAM (configured 48GB up to 80GB for APU via kernel options), 6.12 Linux kernel with GTT enable for ROCM.

Ollama log

time=2025-02-01T15:28:25.660Z level=INFO source=memory.go:356 msg="offload to rocm" layers.requested=-1 layers.model=29 layers.offload=29 layers.split="" memory.available="[48.0 GiB]" memory.gpu_overhead="0 B" memory.required.full="28.6 GiB" memory.required.partial="28.6 GiB" memory.required.kv="7.0 GiB" memory.required.allocations="[28.6 GiB]" memory.weights.total="19.2 GiB" memory.weights.repeating="18.1 GiB" memory.weights.nonrepeating="1.0 GiB" memory.graph.full="7.3 GiB" memory.graph.partial="9.0 GiB"
time=2025-02-01T15:28:25.660Z level=INFO source=server.go:186 msg="enabling flash attention"
time=2025-02-01T15:28:25.661Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/bin/ollama runner --model /root/.ollama/models/blobs/sha256-e70f8fe961784414047b588f056d081199e03d9c45aab3e5aee2502f4aebfca4 --ctx-size 131072 --batch-size 512 --n-gpu-layers 29 --threads 6 --flash-attn --kv-cache-type f16 --parallel 1 --port 45687"
time=2025-02-01T15:28:25.661Z level=INFO source=sched.go:449 msg="loaded runners" count=1
time=2025-02-01T15:28:25.661Z level=INFO source=server.go:564 msg="waiting for llama runner to start responding"
time=2025-02-01T15:28:25.661Z level=INFO source=server.go:598 msg="waiting for server to become available" status="llm server error"
time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:936 msg="starting go runner"
time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:937 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | cgo(gcc)" threads=6
time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:995 msg="Server listening on 127.0.0.1:45687"
load_backend: loaded CPU backend from /usr/lib/ollama/libggml-cpu-haswell.so

OS

Linux, Docker

GPU

AMD

CPU

AMD

Ollama version

0.5.7 @ 475333d commit

Originally created by @rjmalagon on GitHub (Feb 1, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8762 ### What is the issue? Somehow, the new cmake build innovations affect ROCM on AMD APUs. Before this change, I was capable of docker building for AMD APUS with few changes (ROCM version to 6.2.4), just before this commit, it works fine. It's built now, but ROCM fails during execution. I acknowledge that I have some trouble eye parsing this new dockerfile and cmake files, and maybe I am using the wrong docker build arguments. Docker build command: `docker build -f Dockerfile --no-cache --build-arg TARGETARCH=amd64 --build-arg FLAVOR=rocm -t localhost/ollama-rocm:latest ` About my hardware. AMD Ryzen 7 7735HS with AMD RAdeon 680M APU, 96GB shared RAM (configured 48GB up to 80GB for APU via kernel options), 6.12 Linux kernel with GTT enable for ROCM. Ollama log ``` time=2025-02-01T15:28:25.660Z level=INFO source=memory.go:356 msg="offload to rocm" layers.requested=-1 layers.model=29 layers.offload=29 layers.split="" memory.available="[48.0 GiB]" memory.gpu_overhead="0 B" memory.required.full="28.6 GiB" memory.required.partial="28.6 GiB" memory.required.kv="7.0 GiB" memory.required.allocations="[28.6 GiB]" memory.weights.total="19.2 GiB" memory.weights.repeating="18.1 GiB" memory.weights.nonrepeating="1.0 GiB" memory.graph.full="7.3 GiB" memory.graph.partial="9.0 GiB" time=2025-02-01T15:28:25.660Z level=INFO source=server.go:186 msg="enabling flash attention" time=2025-02-01T15:28:25.661Z level=INFO source=server.go:388 msg="starting llama server" cmd="/usr/bin/ollama runner --model /root/.ollama/models/blobs/sha256-e70f8fe961784414047b588f056d081199e03d9c45aab3e5aee2502f4aebfca4 --ctx-size 131072 --batch-size 512 --n-gpu-layers 29 --threads 6 --flash-attn --kv-cache-type f16 --parallel 1 --port 45687" time=2025-02-01T15:28:25.661Z level=INFO source=sched.go:449 msg="loaded runners" count=1 time=2025-02-01T15:28:25.661Z level=INFO source=server.go:564 msg="waiting for llama runner to start responding" time=2025-02-01T15:28:25.661Z level=INFO source=server.go:598 msg="waiting for server to become available" status="llm server error" time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:936 msg="starting go runner" time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:937 msg=system info="CPU : LLAMAFILE = 1 | CPU : LLAMAFILE = 1 | cgo(gcc)" threads=6 time=2025-02-01T15:28:25.671Z level=INFO source=runner.go:995 msg="Server listening on 127.0.0.1:45687" load_backend: loaded CPU backend from /usr/lib/ollama/libggml-cpu-haswell.so ``` ### OS Linux, Docker ### GPU AMD ### CPU AMD ### Ollama version 0.5.7 @ 475333d commit
GiteaMirror added the bug label 2026-05-04 11:32:49 -05:00
GiteaMirror commented 2026-05-04 11:32:57 -05:00
Author
Owner
Copy Link

@rjmalagon commented on GitHub (Feb 17, 2025):

I will close it. It was an issue with the differences between buildah and buildkit. I was able to correct my build scripts to use buildkit correctly with the current dockerfile.

If someone needs some info about how to use buildkit to build Ollama in non-conventional environments, don't doubt to reach me.

<!-- gh-comment-id:2662133052 --> @rjmalagon commented on GitHub (Feb 17, 2025): I will close it. It was an issue with the differences between buildah and buildkit. I was able to correct my build scripts to use buildkit correctly with the current dockerfile. If someone needs some info about how to use buildkit to build Ollama in non-conventional environments, don't doubt to reach me.
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#67744
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?