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

[PR #15065] [MERGED] mlx: fix KV cache snapshot memory leak #77292

New Issue
Closed
opened 2026-05-05 09:57:41 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-05-05 09:57:41 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/15065
Author: @jessegross
Created: 3/25/2026
Status: ✅ Merged
Merged: 3/26/2026
Merged by: @jessegross

Base: main ← Head: jessegross/memleak

📝 Commits (1)

  • d432aab mlx: fix KV cache snapshot memory leak

📊 Changes

3 files changed (+9 additions, -16 deletions)

View changed files

📝 x/mlxrunner/cache/cache.go (+6 -6)
📝 x/mlxrunner/mlx/array.go (+3 -1)
📝 x/mlxrunner/mlx/ops_extra.go (+0 -9)

📄 Description

mlx.Copy shares the backing buffer with its source (via copy_shared_buffer) rather than allocating independent storage. When used to snapshot a slice of the KV cache, the snapshot array holds the entire original cache buffer alive through the shared data pointer — even after eval detaches the computation graph.

Replace Copy with Contiguous in Snapshot and Split. Contiguous allocates a compact buffer when the source buffer is significantly larger than the logical slice (Contiguous::eval checks buffer_size > nbytes + 16384), which is always the case for KV cache slices.

🔄 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/15065 **Author:** [@jessegross](https://github.com/jessegross) **Created:** 3/25/2026 **Status:** ✅ Merged **Merged:** 3/26/2026 **Merged by:** [@jessegross](https://github.com/jessegross) **Base:** `main` ← **Head:** `jessegross/memleak` --- ### 📝 Commits (1) - [`d432aab`](https://github.com/ollama/ollama/commit/d432aabe6953f7345bbd73f97f1b9f260b6953e1) mlx: fix KV cache snapshot memory leak ### 📊 Changes **3 files changed** (+9 additions, -16 deletions) <details> <summary>View changed files</summary> 📝 `x/mlxrunner/cache/cache.go` (+6 -6) 📝 `x/mlxrunner/mlx/array.go` (+3 -1) 📝 `x/mlxrunner/mlx/ops_extra.go` (+0 -9) </details> ### 📄 Description mlx.Copy shares the backing buffer with its source (via copy_shared_buffer) rather than allocating independent storage. When used to snapshot a slice of the KV cache, the snapshot array holds the entire original cache buffer alive through the shared data pointer — even after eval detaches the computation graph. Replace Copy with Contiguous in Snapshot and Split. Contiguous allocates a compact buffer when the source buffer is significantly larger than the logical slice (Contiguous::eval checks buffer_size > nbytes + 16384), which is always the case for KV cache slices. --- <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 2026-05-05 09:57:41 -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#77292
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?