github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7546] libggml linked to wrong cuda version #30561

New Issue
Closed
opened 2026-04-22 10:18:05 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-22 10:18:05 -05:00
Owner
Copy Link

Originally created by @jsurloppe on GitHub (Nov 7, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7546

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

Hi,

I just upgraded to ollama 0.4.0 and loading a model fail with the following error:

/tmp/ollama2415219728/runners/cuda_v12/ollama_llama_server: error while loading shared libraries: libcublas.so.11: cannot open shared object file: No such file or directory
time=2024-11-07T08:55:32.986+01:00 level=ERROR source=sched.go:455 msg="error loading llama server" error="llama runner process has terminated: exit status 127"

It seems that the libggml_cuda_v12.so from the binary distribution is linked to cuda 11

$ ldd libggml_cuda_v12.so 
	linux-vdso.so.1 (0x00007ffd315f8000)
	libcuda.so.1 => /usr/lib64/libcuda.so.1 (0x00007f7bae400000)
	libcublas.so.11 => not found
	libcudart.so.11.0 => not found
	libcublasLt.so.11 => not found
	librt.so.1 => /lib64/librt.so.1 (0x00007f7bf9b83000)
	libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f7bf9b7c000)
	libdl.so.2 => /lib64/libdl.so.2 (0x00007f7bf9b77000)
	libstdc++.so.6 => /usr/lib/gcc/x86_64-pc-linux-gnu/13/libstdc++.so.6 (0x00007f7bae000000)
	libm.so.6 => /lib64/libm.so.6 (0x00007f7bae322000)
	libgcc_s.so.1 => /usr/lib/gcc/x86_64-pc-linux-gnu/13/libgcc_s.so.1 (0x00007f7bf9b52000)
	libc.so.6 => /lib64/libc.so.6 (0x00007f7bade1d000)
	/lib64/ld-linux-x86-64.so.2 (0x00007f7bf9ba9000)

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.4.0

Originally created by @jsurloppe on GitHub (Nov 7, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7546 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? Hi, I just upgraded to ollama 0.4.0 and loading a model fail with the following error: ``` /tmp/ollama2415219728/runners/cuda_v12/ollama_llama_server: error while loading shared libraries: libcublas.so.11: cannot open shared object file: No such file or directory time=2024-11-07T08:55:32.986+01:00 level=ERROR source=sched.go:455 msg="error loading llama server" error="llama runner process has terminated: exit status 127" ``` It seems that the `libggml_cuda_v12.so` from the binary distribution is linked to cuda 11 ``` $ ldd libggml_cuda_v12.so linux-vdso.so.1 (0x00007ffd315f8000) libcuda.so.1 => /usr/lib64/libcuda.so.1 (0x00007f7bae400000) libcublas.so.11 => not found libcudart.so.11.0 => not found libcublasLt.so.11 => not found librt.so.1 => /lib64/librt.so.1 (0x00007f7bf9b83000) libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f7bf9b7c000) libdl.so.2 => /lib64/libdl.so.2 (0x00007f7bf9b77000) libstdc++.so.6 => /usr/lib/gcc/x86_64-pc-linux-gnu/13/libstdc++.so.6 (0x00007f7bae000000) libm.so.6 => /lib64/libm.so.6 (0x00007f7bae322000) libgcc_s.so.1 => /usr/lib/gcc/x86_64-pc-linux-gnu/13/libgcc_s.so.1 (0x00007f7bf9b52000) libc.so.6 => /lib64/libc.so.6 (0x00007f7bade1d000) /lib64/ld-linux-x86-64.so.2 (0x00007f7bf9ba9000) ``` ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.4.0
GiteaMirror added the bug label 2026-04-22 10:18:05 -05:00
GiteaMirror commented 2026-04-22 10:18:09 -05:00
Author
Owner
Copy Link

@jsurloppe commented on GitHub (Nov 8, 2024):

Hey,

Yes, I figured out how to build it while waiting for a fix to be released,
it works without issue.

Thanks

--
Julien Surloppe @.***>

Le jeu. 7 nov. 2024 à 18:12, Simone Scanzoni @.***> a
écrit :

You can work around it building your own libggml_cuda_v12.so, if you have
the libraries and build tools needed, with a command like this (assuming
CUDA in /opt/cuda and an RTX40xx) from the root of the cloned repository:

make OS=linux CUDA_12=1 CUDA_PATH=/opt/cuda/ CMAKE_CUDA_ARCHITECTURES="89"
GPU_PATH_ROOT_LINUX=/opt/cuda/ -f llama/make/Makefile.cuda_v12

—
Reply to this email directly, view it on GitHub
https://github.com/ollama/ollama/issues/7546#issuecomment-2462795967,
or unsubscribe
https://github.com/notifications/unsubscribe-auth/AE5RQGS62EQT2XN64LWV6LDZ7ONQXAVCNFSM6AAAAABRKTKOQWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRSG44TKOJWG4
.
You are receiving this because you authored the thread.Message ID:
@.***>

<!-- gh-comment-id:2464227325 --> @jsurloppe commented on GitHub (Nov 8, 2024): Hey, Yes, I figured out how to build it while waiting for a fix to be released, it works without issue. Thanks -- Julien Surloppe ***@***.***> Le jeu. 7 nov. 2024 à 18:12, Simone Scanzoni ***@***.***> a écrit : > You can work around it building your own libggml_cuda_v12.so, if you have > the libraries and build tools needed, with a command like this (assuming > CUDA in /opt/cuda and an RTX40xx) from the root of the cloned repository: > > make OS=linux CUDA_12=1 CUDA_PATH=/opt/cuda/ CMAKE_CUDA_ARCHITECTURES="89" > GPU_PATH_ROOT_LINUX=/opt/cuda/ -f llama/make/Makefile.cuda_v12 > > — > Reply to this email directly, view it on GitHub > <https://github.com/ollama/ollama/issues/7546#issuecomment-2462795967>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AE5RQGS62EQT2XN64LWV6LDZ7ONQXAVCNFSM6AAAAABRKTKOQWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRSG44TKOJWG4> > . > You are receiving this because you authored the thread.Message ID: > ***@***.***> >
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#30561
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?