ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-03-04 19:10:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

Slightly faster TG for split mode "graph" (#1057)

Browse Source 
* Rearrange graph nodes

So that we can do graph portions that are the same on 2 or more
GPUs at the same time.

* Separate graph compute implementation for split mode graph

* This is better

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2025-12-12 07:54:37 +01:00
committed by GitHub
parent bf03f63c34
commit f65fefa36c
4 changed files with 183 additions and 91 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/llama.cpp
View File

@@ -4771,6 +4771,9 @@ struct llama_context * llama_new_context_with_model(
LLAMA_LOG_INFO("XXXXXXXXXXXXXXXXXXXXX Setting only active experts offload\n");
ggml_backend_sched_set_only_active_experts(ctx->sched, true);
}
if (model->split_mode == LLAMA_SPLIT_MODE_GRAPH) {
ggml_backend_sched_set_split_mode_graph(ctx->sched, true);
}
return ctx;
}