github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #13821] [Feature Request] Add OpenCL and OpenBLAS support for broader hardware compatibility #71113

New Issue
Closed
opened 2026-05-05 00:23:50 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-05-05 00:23:50 -05:00
Owner
Copy Link

Originally created by @jonarser on GitHub (Jan 21, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13821

Description:
I am a user with older/limited hardware ([GT 710M 1GB, Intel HD 4000]) and would like to run local LLMs efficiently. Currently, Ollama/GPT4All primarily supports acceleration via CUDA (NVIDIA), ROCm (AMD), and Metal (Apple), which excludes many integrated and older discrete GPUs that only have OpenCL drivers.

Proposed Solution:
Integrate the OpenCL and OpenBLAS backends from the upstream llama.cpp project. The OpenCL backend in llama.cpp (-DGGML_OPENCL=ON) already provides good cross-platform support for Adreno, older Intel HD Graphics, and many other GPUs. OpenBLAS (-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS) can significantly accelerate CPU inference, which is crucial when GPU memory is insufficient.

This would make the project accessible to a much wider range of users with non-mainstream or legacy hardware.

My Use Case:

  • Hardware: Hybrid graphics, NVIDIA GT 710M 1GB + Intel HD 4000).
  • Driver/OS: NVIDIA driver 388.xx on Windows 10 and 390xx Linux.
  • Current Limitation: Vulkan support is unstable, CUDA is not optimal for this old GPU. OpenCL is the only viable cross-platform compute API that works.
  • Trying to use Functiongemma 270m

Additional context:

  • The llama.cpp project, which underpins these tools, already has stable OpenCL and BLAS backends.
  • An issue requesting OpenCL support for Ollama already exists (https://github.com/ollama/ollama/issues/7536), indicating community interest.
  • For users with weak or no dedicated GPU, CPU acceleration via BLAS libraries is often the only performant option.
Originally created by @jonarser on GitHub (Jan 21, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13821 Description: I am a user with older/limited hardware ([GT 710M 1GB, Intel HD 4000]) and would like to run local LLMs efficiently. Currently, Ollama/GPT4All primarily supports acceleration via CUDA (NVIDIA), ROCm (AMD), and Metal (Apple), which excludes many integrated and older discrete GPUs that only have OpenCL drivers. Proposed Solution: Integrate the OpenCL and OpenBLAS backends from the upstream llama.cpp project. The OpenCL backend in llama.cpp (-DGGML_OPENCL=ON) already provides good cross-platform support for Adreno, older Intel HD Graphics, and many other GPUs. OpenBLAS (-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS) can significantly accelerate CPU inference, which is crucial when GPU memory is insufficient. This would make the project accessible to a much wider range of users with non-mainstream or legacy hardware. My Use Case: - Hardware: Hybrid graphics, NVIDIA GT 710M 1GB + Intel HD 4000). - Driver/OS: NVIDIA driver 388.xx on Windows 10 and 390xx Linux. - Current Limitation: Vulkan support is unstable, CUDA is not optimal for this old GPU. OpenCL is the only viable cross-platform compute API that works. - Trying to use Functiongemma 270m Additional context: - The llama.cpp project, which underpins these tools, already has stable OpenCL and BLAS backends. - An issue requesting OpenCL support for Ollama already exists (https://github.com/ollama/ollama/issues/7536), indicating community interest. - For users with weak or no dedicated GPU, CPU acceleration via BLAS libraries is often the only performant option.
GiteaMirror added the feature request label 2026-05-05 00:23:50 -05:00
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 feature request
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#71113
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?