llama.cpp
(Bugfix, ggml-cuda) Pool alloc count fix + small size computation type adjustment
#18559
Merged
Merged
(Bugfix, ggml-cuda) Pool alloc count fix + small size computation type adjustment #18559
JohannesGaessler
merged
2 commits into
ggml-org:master
from
pl752:pool_alloc_count_fix
CUDA: Fixed obj byte size instead of obj count being passed to pool a…
cca7749a
CUDA: Explicitly casted some of the int alloc counts before multiplic…
55096144
JohannesGaessler
approved these changes
on 2026-01-03
Login to write a write a comment.
Login via GitHub