github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #6735] [MERGED] runner.go: Prompt caching #38083

New Issue
Closed
opened 2026-04-22 22:45:24 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-22 22:45:24 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/6735
Author: @jessegross
Created: 9/10/2024
Status: ✅ Merged
Merged: 9/11/2024
Merged by: @jessegross

Base: jmorganca/llama ← Head: jessegross/prompt_cache

📝 Commits (1)

  • 0273dfe runner.go: Prompt caching

📊 Changes

4 files changed (+378 additions, -24 deletions)

View changed files

📝 llama/llama.go (+4 -0)
➕ llama/runner/cache.go (+156 -0)
➕ llama/runner/cache_test.go (+192 -0)
📝 llama/runner/runner.go (+26 -24)

📄 Description

Currently, KV cache entries from a sequence are discarded at the end of each processing run. In a typical chat conversation, this results in each message taking longer and longer to process as the entire history of the conversation needs to be replayed.

Prompt caching retains the KV entries as long as possible so that we only need to process the newest message in the converation, at least until there are too many simultaneous conversations and something needs to be evicted.

🔄 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/6735 **Author:** [@jessegross](https://github.com/jessegross) **Created:** 9/10/2024 **Status:** ✅ Merged **Merged:** 9/11/2024 **Merged by:** [@jessegross](https://github.com/jessegross) **Base:** `jmorganca/llama` ← **Head:** `jessegross/prompt_cache` --- ### 📝 Commits (1) - [`0273dfe`](https://github.com/ollama/ollama/commit/0273dfe0e7710c2d84f625bc14de5438d4b33c1d) runner.go: Prompt caching ### 📊 Changes **4 files changed** (+378 additions, -24 deletions) <details> <summary>View changed files</summary> 📝 `llama/llama.go` (+4 -0) ➕ `llama/runner/cache.go` (+156 -0) ➕ `llama/runner/cache_test.go` (+192 -0) 📝 `llama/runner/runner.go` (+26 -24) </details> ### 📄 Description Currently, KV cache entries from a sequence are discarded at the end of each processing run. In a typical chat conversation, this results in each message taking longer and longer to process as the entire history of the conversation needs to be replayed. Prompt caching retains the KV entries as long as possible so that we only need to process the newest message in the converation, at least until there are too many simultaneous conversations and something needs to be evicted. --- <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 22:45:24 -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#38083
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?