ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-09 07:50:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

Graph parallel for Mimo-V2-Flash (#1105)

Browse Source 
* WIP

* Cleanup

* Set max_gpu to 2 for Mimo2

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2026-01-05 09:58:54 +02:00
committed by GitHub
parent 385fc14110
commit 419a397ce0
5 changed files with 45 additions and 40 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llama.cpp
View File

@@ -1730,6 +1730,7 @@ static bool is_model_split_supported(const llama_model & model) {
LLM_ARCH_GLM4_MOE,
LLM_ARCH_MISTRAL3,
LLM_ARCH_COHERE2,
LLM_ARCH_MIMO2,
};
auto it = k_supported.find(model.arch);
return it != k_supported.end();
@@ -1760,6 +1761,13 @@ static bool llm_load_tensors(
LLAMA_LOG_WARN(" => changing split mode to 'layer'\n");
LLAMA_LOG_WARN("=======================================================\n\n");
split_mode = LLAMA_SPLIT_MODE_LAYER;
} else {
if (model.arch == LLM_ARCH_MIMO2 && model.devices.size() > 2 && max_gpu != 2) {
LLAMA_LOG_WARN("\n================================================================\n");
LLAMA_LOG_WARN("Split mode 'graph' for Mimo2 does not work with more than 2 GPUs\n");
LLAMA_LOG_WARN(" => setting max_gpu to 2\n");
LLAMA_LOG_WARN("================================================================\n\n");
}
}
}