github-starred/ollama-ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-03-11 20:23:55 -05:00
Code Issues 2.2k Packages Projects Releases 100 Wiki Activity

[PR #8792] Addded environment-controlled number of threads for ollama server ($OMP_NUM_THREADS). #11509

New Issue
Open
opened 2025-11-12 16:15:49 -06:00 by GiteaMirror · 0 comments
GiteaMirror commented 2025-11-12 16:15:49 -06:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/8792
Author: @james-irwin
Created: 2/3/2025
Status: 🔄 Open

Base: main ← Head: add_num_threads

📝 Commits (2)

  • b568646 Addded environment-controlled number of threads for ollama server ().
  • da9f78a Merge branch 'main' into add_num_threads

📊 Changes

1 file changed (+2 additions, -1 deletions)

View changed files

📝 llm/server.go (+2 -1)

📄 Description

OMP_NUM_THREADS is a common (if dated) controlling environment variable to assert the number of desired workers used at runtime for an application. While ollama server isn't necessarily using OMP for its threading model, this is consistent with other systems that obey OMP_NUM_THREADS while also having a native controlling environment variable (for example. those that also entertain exclusive/tailored controlling environment variables MKL_NUM_THREADS and GOTO_NUM_THREADS).

Motivation: Using systems with high core counts and needing to pin to NUMA zones (via numactl) as well as heterogeneous cores within the single operating system image. Without this, the following scenario presents:

#cores in system = N
#cores desired = M (M<<N)
$> numactl -C 0-(M-1) go run ollama serve

Will launch N threads bound to M cores negatively affecting performance.

With this change:

$> OMP_NUM_THREADS=M numactl -C 0-(M-1) go run ollama serve

Will launch M threads and place them on the M assigned cores delivering expected performance.

🔄 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/8792 **Author:** [@james-irwin](https://github.com/james-irwin) **Created:** 2/3/2025 **Status:** 🔄 Open **Base:** `main` ← **Head:** `add_num_threads` --- ### 📝 Commits (2) - [`b568646`](https://github.com/ollama/ollama/commit/b568646af44340ac636fa8e39f7f2514a502c82b) Addded environment-controlled number of threads for ollama server (). - [`da9f78a`](https://github.com/ollama/ollama/commit/da9f78a86c4e44dc784372e2cec137339f093a7a) Merge branch 'main' into add_num_threads ### 📊 Changes **1 file changed** (+2 additions, -1 deletions) <details> <summary>View changed files</summary> 📝 `llm/server.go` (+2 -1) </details> ### 📄 Description OMP_NUM_THREADS is a common (if dated) controlling environment variable to assert the number of desired workers used at runtime for an application. While ollama server isn't necessarily using OMP for its threading model, this is consistent with other systems that obey OMP_NUM_THREADS while also having a native controlling environment variable (for example. those that also entertain exclusive/tailored controlling environment variables MKL_NUM_THREADS and GOTO_NUM_THREADS). Motivation: Using systems with high core counts and needing to pin to NUMA zones (via numactl) as well as heterogeneous cores within the single operating system image. Without this, the following scenario presents: #cores in system = N #cores desired = M (M<<N) $> numactl -C 0-(M-1) go run ollama serve Will launch N threads bound to M cores negatively affecting performance. With this change: $> OMP_NUM_THREADS=M numactl -C 0-(M-1) go run ollama serve Will launch M threads and place them on the M assigned cores delivering expected performance. --- <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 2025-11-12 16:15:49 -06:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
client2
cloud
compatibility
context-length
create
docker
documentation
embeddings
engine
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
install
integration
intel
js
linux
macos
memory
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
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-ollama#11509
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?