llama.cpp
context : reserve new scheduler when graph topology changes
#18547
Open
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
11
Changes
View On
GitHub
context : reserve new scheduler when graph topology changes
#18547
ggerganov
wants to merge 11 commits into
master
from
gg/llama-reserve
danbev
approved these changes on 2026-01-02
Base automatically changed from
gg/metal-adjust-fa-extra-size
to
master
12 days ago
ggerganov
force pushed
from
400466c0
to
bd5de6ba
12 days ago
ggerganov
force pushed
from
89d19e00
to
c92df391
10 days ago
ggerganov
force pushed
from
c92df391
to
cf2b3cae
9 days ago
ggerganov
force pushed
from
cf2b3cae
to
4b744105
3 days ago
context : reserve new scheduler when graph topology changes
e115c637
cont : fix
7b526420
cont : fix reserve
94426b2e
cont : reserve only when changes occur + timing
03e9d66c
context : add comments
5260bb79
llama : reserve on sampler changes
0c0d0fdc
common : allow null common_sampler
b579b970
ggerganov
force pushed
from
4b744105
to
b579b970
2 days ago
ggerganov
requested a review
from
ngxson
2 days ago
github-actions
added
examples
github-actions
added
server
ngxson
commented on 2026-01-12
server : task declares needs (embd, logits, sampling)
ffa0d15e
server : do not init sampler if not needed
be9e6ef2
llama : fix need_reserve when unsetting a sampler
3084bfe6
server : consolidate slot reset/clear logic
d9146ed2
Login to write a write a comment.
Login via GitHub
Reviewers
danbev
ngxson
Assignees
No one assigned
Labels
examples
server
Milestone
No milestone
Login to write a write a comment.
Login via GitHub