github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2797] Please consider supporting Intel GPU ARC A770 (16G) #48203

New Issue
Closed
opened 2026-04-28 07:09:48 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-28 07:09:48 -05:00
Owner
Copy Link

Originally created by @HelloMorningStar on GitHub (Feb 28, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2797

Originally assigned to: @dhiltgen on GitHub.

Here is a demo of ARC A770 running llama2:

https://www.reddit.com/r/LocalLLaMA/comments/1b0c6u8/llama_2_inference_with_pytorch_on_intel_arc/

The Intel Arc A770 is a powerful graphics card that is well-suited for a variety of tasks, including machine learning. It has 16GB of GDDR6 memory, a 256-bit memory interface, and a boost clock of 2.1 GHz. It also supports ray tracing and XeSS, which can improve performance in games and other applications.

The llama2 demo shows how the Intel Arc A770 can be used to accelerate machine learning inference tasks. The demo uses PyTorch, a popular machine learning framework, to run a llama2 model on the Intel Arc A770. The results show that the Intel Arc A770 can achieve significant performance gains over CPUs and other GPUs.

Originally created by @HelloMorningStar on GitHub (Feb 28, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2797 Originally assigned to: @dhiltgen on GitHub. Here is a demo of ARC A770 running llama2: https://www.reddit.com/r/LocalLLaMA/comments/1b0c6u8/llama_2_inference_with_pytorch_on_intel_arc/ The Intel Arc A770 is a powerful graphics card that is well-suited for a variety of tasks, including machine learning. It has 16GB of GDDR6 memory, a 256-bit memory interface, and a boost clock of 2.1 GHz. It also supports ray tracing and XeSS, which can improve performance in games and other applications. The llama2 demo shows how the Intel Arc A770 can be used to accelerate machine learning inference tasks. The demo uses PyTorch, a popular machine learning framework, to run a llama2 model on the Intel Arc A770. The results show that the Intel Arc A770 can achieve significant performance gains over CPUs and other GPUs.
GiteaMirror added the intel label 2026-04-28 07:09:48 -05:00
GiteaMirror commented 2026-04-28 07:09:55 -05:00
Author
Owner
Copy Link

@renyuneyun commented on GitHub (Apr 12, 2024):

Maybe #2033 could solve this?
According to this article, Vulkan has universal acceleration support for lots of GPUs, including Intel's (integrated or discrete) GPUs.

But actually I'm also curious on the status of the new neural engines that Intel introduced to the Ultra series of chips (and similar stuffs AMD introduces to their chips). That's not GPU, and Vulkan cannot support, I believe? Not sure what tools can unify the support of that.

<!-- gh-comment-id:2052105556 --> @renyuneyun commented on GitHub (Apr 12, 2024): Maybe #2033 could solve this? According to [this article](https://blog.nomic.ai/posts/gpt4all-gpu-inference-with-vulkan), Vulkan has universal acceleration support for lots of GPUs, including Intel's (integrated or discrete) GPUs. But actually I'm also curious on the status of the new neural engines that Intel introduced to the Ultra series of chips (and similar stuffs AMD introduces to their chips). That's not GPU, and Vulkan cannot support, I believe? Not sure what tools can unify the support of that.
GiteaMirror commented 2026-04-28 07:09:57 -05:00
Author
Owner
Copy Link

@DurianyDoriana commented on GitHub (Apr 15, 2024):

Maybe #2033 could solve this?

But actually I'm also curious on the status of the new neural engines that Intel introduced to the Ultra series of chips (and similar stuffs AMD introduces to their chips). That's not GPU, and Vulkan cannot support, I believe? Not sure what tools can unify the support of that.
Yes, Vulkan works great in Llama.cpp, GPT4All and other ready programs such as Jan.ai on Intel iGPU's and dGPU's.

Intel also supports 50+ LLM models and LangChain through IPEX-LLM. In fact, Ollama is mentioned on the IPEX-LLM github page:
https://github.com/intel-analytics/ipex-llm

Regarding your second question, their NPU are currently only supported through OpenVINO in Linux and Windows, and there's "Limited initial support for DirectML".
https://github.com/openvinotoolkit/openvino/tree/master/src/plugins/intel_npu
https://downloadmirror.intel.com/820222/NPU_Win_Release_Notes_v2267.pdf

<!-- gh-comment-id:2057890976 --> @DurianyDoriana commented on GitHub (Apr 15, 2024): > Maybe #2033 could solve this? > > But actually I'm also curious on the status of the new neural engines that Intel introduced to the Ultra series of chips (and similar stuffs AMD introduces to their chips). That's not GPU, and Vulkan cannot support, I believe? Not sure what tools can unify the support of that. Yes, Vulkan works great in Llama.cpp, GPT4All and other ready programs such as Jan.ai on Intel iGPU's and dGPU's. Intel also supports 50+ LLM models and LangChain through IPEX-LLM. In fact, Ollama is mentioned on the IPEX-LLM github page: https://github.com/intel-analytics/ipex-llm Regarding your second question, their NPU are currently only supported through OpenVINO in Linux and Windows, and there's "Limited initial support for DirectML". https://github.com/openvinotoolkit/openvino/tree/master/src/plugins/intel_npu https://downloadmirror.intel.com/820222/NPU_Win_Release_Notes_v2267.pdf
GiteaMirror commented 2026-04-28 07:09:59 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Apr 15, 2024):

Dup of #1590

<!-- gh-comment-id:2057924219 --> @dhiltgen commented on GitHub (Apr 15, 2024): Dup of #1590
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 intel
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#48203
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?