Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ui: reduce per-token render cost when streaming (#26053)

Browse Source 
* performance harness - the empirical root

Assisted-by: Claude Opus 4.8

* 210.36ms -> 2.67ms per streamed token

Assisted-by: Claude Opus 4.8

* 11.58ms -> 0.62ms per streamed token

Assisted-by: Claude Opus 4.8

* 22.02ms -> 3.33ms per streamed token

Assisted-by: Claude Opus 4.8

* 3.07ms -> 1.36ms per streamed token at 40 messages

Assisted-by: Claude Opus 4.8

---------

Co-authored-by: Zach Winter <dmtommy@icloud.com>
This commit is contained in:
Zach Winter
2026-07-24 22:09:46 +02:00
committed by GitHub
co-authored by Zach Winter
parent 96013c5112
commit 555881ebc8
18 changed files with 1037 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/ui/src/lib/stores/conversations.svelte.ts
+15 -2
View File
@@ -169,8 +169,21 @@ class ConversationsStore {
* Updates a message at a specific index in active messages
*/
updateMessageAtIndex(index: number, updates: Partial<DatabaseMessage>): void {
if (index !== -1 && this.activeMessages[index]) {
this.activeMessages[index] = { ...this.activeMessages[index], ...updates };
const message = index === -1 ? undefined : this.activeMessages[index];
if (!message) return;
// Assign field by field rather than replacing the object. Replacing it
// changes the array slot, which invalidates every consumer that merely
// walks the list - notably ChatMessages.displayMessages, which rebuilds
// entries for every message in the conversation. Deep $state proxies make
// per-field writes fine-grained, so only readers of the changed field wake.
const target = message as unknown as Record<string, unknown>;
for (const [key, value] of Object.entries(updates)) {
if (target[key] !== value) {
target[key] = value;
}
}
}