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

[PR #454] [MERGED] first pass at linux gpu support #36031

New Issue
Closed
opened 2026-04-22 20:45:13 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-22 20:45:13 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/454
Author: @BruceMacD
Created: 9/1/2023
Status: ✅ Merged
Merged: 9/12/2023
Merged by: @BruceMacD

Base: main ← Head: brucemacd/linux-gpu

📝 Commits (6)

📊 Changes

9 files changed (+158 additions, -22 deletions)

View changed files

➕ Dockerfile.cuda (+22 -0)
📝 api/types.go (+1 -1)
📝 docs/development.md (+10 -0)
📝 llm/llama.cpp/generate.go (+8 -6)
📝 llm/llama.cpp/generate_darwin_amd64.go (+11 -7)
📝 llm/llama.cpp/generate_darwin_arm64.go (+10 -6)
➕ llm/llama.cpp/generate_linux.go (+14 -0)
📝 llm/llama.go (+74 -2)
📝 server/routes.go (+8 -0)

📄 Description

This is the basic implementation of enabling a linux build with GPU support.

Building for Linux with CPU support is unchanged (generate and build as normal).

Building for Linux with GPU requires generating with the gpu tag. This is to allow non-GPU linux builds to continue to be built locally without issue.

How to build/run:

  • generate the required dependencies: go generate ./...
  • build the binary go build .
    and run as normal:
    ./ollama serve &
    ./ollama run llama2

Follow up:

  • Packaging nvidia drivers or downloading them automatically
  • Better heuristics for determining the number of layers to load into GPU

Part of #259

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/454 **Author:** [@BruceMacD](https://github.com/BruceMacD) **Created:** 9/1/2023 **Status:** ✅ Merged **Merged:** 9/12/2023 **Merged by:** [@BruceMacD](https://github.com/BruceMacD) **Base:** `main` ← **Head:** `brucemacd/linux-gpu` --- ### 📝 Commits (6) - [`63c5115`](https://github.com/ollama/ollama/commit/63c5115f0c4a89d5ee245527781271c90701f850) linux gpu support - [`5e672ce`](https://github.com/ollama/ollama/commit/5e672ce7240acaf7d3f85350f59119417c2329e5) handle multiple gpus - [`86be0f1`](https://github.com/ollama/ollama/commit/86be0f1d110c80607a57c636bd933e7daeac6067) Update generate_linux.sh - [`7709c0c`](https://github.com/ollama/ollama/commit/7709c0c0f9d0acf0e3ecb7a1bc1af218a325404d) add cuda docker image (#488) - [`e6093f7`](https://github.com/ollama/ollama/commit/e6093f7ab56f0e33d91c98b5c33af0fec0f342a8) use total gpu memory - [`f1aab4f`](https://github.com/ollama/ollama/commit/f1aab4f551cfe17cd9c1ca8652765692885ae0bc) Update development.md ### 📊 Changes **9 files changed** (+158 additions, -22 deletions) <details> <summary>View changed files</summary> ➕ `Dockerfile.cuda` (+22 -0) 📝 `api/types.go` (+1 -1) 📝 `docs/development.md` (+10 -0) 📝 `llm/llama.cpp/generate.go` (+8 -6) 📝 `llm/llama.cpp/generate_darwin_amd64.go` (+11 -7) 📝 `llm/llama.cpp/generate_darwin_arm64.go` (+10 -6) ➕ `llm/llama.cpp/generate_linux.go` (+14 -0) 📝 `llm/llama.go` (+74 -2) 📝 `server/routes.go` (+8 -0) </details> ### 📄 Description This is the basic implementation of enabling a linux build with GPU support. Building for Linux with CPU support is unchanged (generate and build as normal). Building for Linux with GPU requires generating with the `gpu` tag. This is to allow non-GPU linux builds to continue to be built locally without issue. How to build/run: - generate the required dependencies: `go generate ./...` - build the binary `go build .` and run as normal: `./ollama serve &` `./ollama run llama2` Follow up: - Packaging nvidia drivers or downloading them automatically - Better heuristics for determining the number of layers to load into GPU Part of #259 --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-22 20:45:13 -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 pull-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#36031
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?