Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-08-27 23:50:49 +02:00
Code Issues Packages Projects Releases Wiki Activity

move pipeline parallelism to a --pipelineparallel launch flag

Browse Source
This commit is contained in:
Concedo
2025-12-11 21:03:41 +08:00
parent b7428048fc
commit fd0d0cab03
5 changed files with 20 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-backend.cpp
+2 -2
View File
@@ -662,8 +662,8 @@ static bool ggml_is_view_op(enum ggml_op op) {
#endif
#ifndef GGML_SCHED_MAX_COPIES
//kcpp reduced from 4 to 2 to try make buffer sizes smaller on multigpu
#define GGML_SCHED_MAX_COPIES 2
//kcpp can reduce this if you want to try make buffer sizes smaller on multigpu
#define GGML_SCHED_MAX_COPIES 4
#endif
struct ggml_backend_sched_split {