github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2065] Any ollama command results in CORE DUMPED (ollama not using GPU) #26955

New Issue
Closed
opened 2026-04-22 03:45:03 -05:00 by GiteaMirror · 8 comments
GiteaMirror commented 2026-04-22 03:45:03 -05:00
Owner
Copy Link

Originally created by @Rushmore75 on GitHub (Jan 19, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2065

Trying to interact with the command at all just returns Illegal instruction (core dumped). The journalctl logs just show

Started Ollama Service
ollama.service: Main process exited, code=dumped, status=4/ILL
ollama.service: Failed with result 'core-dump;

System:
Kernel: 5.15.0-91-generic
Distro: Ubuntu 22.04.3 LTS
Hardware: (Proxmox 8.1.3)

  • CPU: x86-64-v2-AES
  • GPU: (Passthru) Nvidia 1070
  • BIOS: SeaBIOS
  • Machine: i440fx

I would imagine it is linked to #2000 - perhaps something to so with VMs?

Originally created by @Rushmore75 on GitHub (Jan 19, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2065 Trying to interact with the command at all just returns `Illegal instruction (core dumped)`. The journalctl logs just show ``` Started Ollama Service ollama.service: Main process exited, code=dumped, status=4/ILL ollama.service: Failed with result 'core-dump; ``` System: Kernel: 5.15.0-91-generic Distro: Ubuntu 22.04.3 LTS Hardware: (Proxmox 8.1.3) * CPU: x86-64-v2-AES * GPU: (Passthru) Nvidia 1070 * BIOS: SeaBIOS * Machine: i440fx I would imagine it is linked to #2000 - perhaps something to so with VMs?
GiteaMirror added the linuxbug labels 2026-04-22 03:45:03 -05:00
GiteaMirror commented 2026-04-22 03:45:05 -05:00
Author
Owner
Copy Link

@Rushmore75 commented on GitHub (Jan 19, 2024):

Also tried with the q35 machine, still crashes

<!-- gh-comment-id:1899726078 --> @Rushmore75 commented on GitHub (Jan 19, 2024): Also tried with the `q35` machine, still crashes
GiteaMirror commented 2026-04-22 03:45:05 -05:00
Author
Owner
Copy Link

@Rushmore75 commented on GitHub (Jan 19, 2024):

As far as i can tell (building & running the program locally with gdb) the error is coming from https://github.com/gabriel-vasile/mimetype/blob/master/internal/magic/ftyp.go line 11...

<!-- gh-comment-id:1899795279 --> @Rushmore75 commented on GitHub (Jan 19, 2024): As far as i can tell (building & running the program locally with gdb) the error is coming from https://github.com/gabriel-vasile/mimetype/blob/master/internal/magic/ftyp.go line 11...
GiteaMirror commented 2026-04-22 03:45:06 -05:00
Author
Owner
Copy Link

@Rushmore75 commented on GitHub (Jan 19, 2024):

updating mimetype in go.mod seemed to fix the issue. It required the addition of a second package however, I forgot the name but it was listed in the output message from the addition

<!-- gh-comment-id:1899822090 --> @Rushmore75 commented on GitHub (Jan 19, 2024): updating mimetype in [go.mod](./go.mod) seemed to fix the issue. It required the addition of a second package however, I forgot the name but it was listed in the output message from the addition
GiteaMirror commented 2026-04-22 03:45:07 -05:00
Author
Owner
Copy Link

@Rushmore75 commented on GitHub (Jan 20, 2024):

Was already installed:

cuda-drivers-545
cuda-drivers
cuda-keyring

Fixed my PPAs

commands source

sudo apt-get --purge remove "*cublas*" "cuda*" "*nvidia*"
sudo apt-get clean
sudo apt-get autoremove
sudo apt-get update
sudo apt-get upgrade

Then installed

nvidia-cuda-toolkit
nvidia-driver-535

Rebuild ollama

(I just removed the whole repo and re-cloned it)

  1. Add to go.mod:
    Change github.com/gabriel-vasile/mimetype v1.4.3 to github.com/gabriel-vasile/mimetype v1.4.3
    This was causing core dump to happen before, idk why, but updating it fixes it.

  2. go get github.com/go-playground/validator/v10@v10.14.0

  3. go generate ./...

  4. go build -buildmode=pie -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external -ldflags=-buildid=''
    These flags are from the ollama-cuda AUR package, idk really what they do lol

How ever, still no gpu accelration... I'm using the llama2 model now.
nvtop shows no programs using the gpu and nvidia-smi doesn't either.
When I run the program it shows "INFO CUDA Compute Capability detected: 6.1"

<!-- gh-comment-id:1901782229 --> @Rushmore75 commented on GitHub (Jan 20, 2024): # Was already installed: cuda-drivers-545 cuda-drivers cuda-keyring # Fixed my PPAs [commands source](https://askubuntu.com/questions/1289811/cant-install-nvidia-driver-toolkit-on-ubuntu-20-04-lts-needs-uninstallable-pa) ```bash sudo apt-get --purge remove "*cublas*" "cuda*" "*nvidia*" sudo apt-get clean sudo apt-get autoremove sudo apt-get update sudo apt-get upgrade ``` # Then installed nvidia-cuda-toolkit nvidia-driver-535 # Rebuild ollama (I just removed the whole repo and re-cloned it) 1) Add to `go.mod`: Change `github.com/gabriel-vasile/mimetype v1.4.3` to `github.com/gabriel-vasile/mimetype v1.4.3` This was causing **core dump** to happen before, idk why, but updating it fixes it. 2) `go get github.com/go-playground/validator/v10@v10.14.0` 3) `go generate ./...` 4) `go build -buildmode=pie -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external -ldflags=-buildid=''` These flags are from the [ollama-cuda AUR](https://gitlab.archlinux.org/archlinux/packaging/packages/ollama-cuda/-/blob/main/PKGBUILD?ref_type=heads) package, idk really what they do lol How ever, still no gpu accelration... I'm using the llama2 model now. `nvtop` shows no programs using the gpu and `nvidia-smi` doesn't either. When I run the program it shows "INFO CUDA Compute Capability detected: 6.1"
GiteaMirror commented 2026-04-22 03:45:07 -05:00
Author
Owner
Copy Link

@nkeilar commented on GitHub (Jan 21, 2024):

I'm also running into similar issues, Ubuntu 22.04, using the 545 drivers... Lots of stability issues. But was hard to get Ubuntu to be happy with a single consistent set of drivers.

<!-- gh-comment-id:1902611076 --> @nkeilar commented on GitHub (Jan 21, 2024): I'm also running into similar issues, Ubuntu 22.04, using the 545 drivers... Lots of stability issues. But was hard to get Ubuntu to be happy with a single consistent set of drivers.
GiteaMirror commented 2026-04-22 03:45:08 -05:00
Author
Owner
Copy Link

@pwgit-create commented on GitHub (Jan 21, 2024):

I have the same issue on Ubuntu 22.04.3. Just got my Nvidia 4070 TI passed to my VM and Ollama installed with GPU enabled for the first time 😃 Has a fix been integrated into the latest release of Ollama or is the problem on my side? Awesome work with Ollama by the way, I Love it!

EDIT: Running the binary from pre-release v0.1.21 has resulted in it now working :)

<!-- gh-comment-id:1902646444 --> @pwgit-create commented on GitHub (Jan 21, 2024): I have the same issue on Ubuntu 22.04.3. Just got my Nvidia 4070 TI passed to my VM and Ollama installed with GPU enabled for the first time :smiley: Has a fix been integrated into the latest release of Ollama or is the problem on my side? Awesome work with Ollama by the way, I Love it! EDIT: Running the binary from pre-release v0.1.21 has resulted in it now working :)
GiteaMirror commented 2026-04-22 03:45:09 -05:00
Author
Owner
Copy Link

@Rushmore75 commented on GitHub (Jan 24, 2024):

nice, it would seem that the original issue that this was opened for - you can get around by just building from source or using a pre-release

<!-- gh-comment-id:1907392534 --> @Rushmore75 commented on GitHub (Jan 24, 2024): nice, it would seem that the original issue that this was opened for - you can get around by just building from source or using a pre-release
GiteaMirror commented 2026-04-22 03:45:10 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 11, 2024):

Going to go ahead and close this since it should now be working.

<!-- gh-comment-id:1989101917 --> @pdevine commented on GitHub (Mar 11, 2024): Going to go ahead and close this since it should now be working.
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 linux
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#26955
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?