ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-03-02 18:10:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

Enable up to 4 GPUs for Mimo2-Flash (#1115)

Browse Source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2026-01-07 09:40:29 +02:00
committed by GitHub
parent a82dcbf3ee
commit ab1616767b
1 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
src/llama.cpp
View File

@@ -1764,11 +1764,12 @@ static bool llm_load_tensors(
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) {
if (model.arch == LLM_ARCH_MIMO2 && model.devices.size() > 4 && (max_gpu == 0 || max_gpu > 4)) {
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("Split mode 'graph' for Mimo2 does not work with more than 4 GPUs\n");
LLAMA_LOG_WARN(" => setting max_gpu to 4\n");
LLAMA_LOG_WARN("================================================================\n\n");
max_gpu = 4;
}
}
}