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 #13141] [MERGED] kvcache: Use SetRows to store cache data #24632

New Issue
Closed
opened 2026-04-19 17:41:59 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-19 17:41:59 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/13141
Author: @jessegross
Created: 11/19/2025
Status: ✅ Merged
Merged: 11/19/2025
Merged by: @jessegross

Base: main ← Head: jessegross/set_rows

📝 Commits (2)

  • 07f10e3 ggml: Automatically make tensors contiguous on reshape
  • af5ca95 kvcache: Use SetRows to store cache data

📊 Changes

4 files changed (+168 additions, -235 deletions)

View changed files

📝 kvcache/causal.go (+41 -174)
📝 kvcache/causal_test.go (+115 -61)
📝 ml/backend.go (+1 -0)
📝 ml/backend/ggml/ggml.go (+11 -0)

📄 Description

We currently copy data into the KV cache in contiguous buffers using ggml_cpy(). ggml_set_rows() was introduced to allow scatter operation so that contiguous buffers are no longer required. The direct primary benefit of this is that we no longer need to perform defragmentation.

However, GGML recently removed an optimization for ggml_cpy() and we picked it up in 544b673 "ggml update to b6840 (#12791)". This caused a roughly 40% drop in token generation performance on CUDA due to CUDA graphs no longer being used. By switching to ggml_set_rows(), the original optimization is no longer necessary and CUDA performance is restored.

Fixes #13112

🔄 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/13141 **Author:** [@jessegross](https://github.com/jessegross) **Created:** 11/19/2025 **Status:** ✅ Merged **Merged:** 11/19/2025 **Merged by:** [@jessegross](https://github.com/jessegross) **Base:** `main` ← **Head:** `jessegross/set_rows` --- ### 📝 Commits (2) - [`07f10e3`](https://github.com/ollama/ollama/commit/07f10e3a4922220e1cef09f864961de7c0887694) ggml: Automatically make tensors contiguous on reshape - [`af5ca95`](https://github.com/ollama/ollama/commit/af5ca9515f0d11ed0f2808f0dde3b78415fed235) kvcache: Use SetRows to store cache data ### 📊 Changes **4 files changed** (+168 additions, -235 deletions) <details> <summary>View changed files</summary> 📝 `kvcache/causal.go` (+41 -174) 📝 `kvcache/causal_test.go` (+115 -61) 📝 `ml/backend.go` (+1 -0) 📝 `ml/backend/ggml/ggml.go` (+11 -0) </details> ### 📄 Description We currently copy data into the KV cache in contiguous buffers using ggml_cpy(). ggml_set_rows() was introduced to allow scatter operation so that contiguous buffers are no longer required. The direct primary benefit of this is that we no longer need to perform defragmentation. However, GGML recently removed an optimization for ggml_cpy() and we picked it up in 544b673 "ggml update to b6840 (#12791)". This caused a roughly 40% drop in token generation performance on CUDA due to CUDA graphs no longer being used. By switching to ggml_set_rows(), the original optimization is no longer necessary and CUDA performance is restored. Fixes #13112 --- <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 2026-04-19 17:41:59 -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#24632
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?