vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

server : add --split-mode parameter

Browse source
This commit is contained in:
slaren 2024-01-11 14:46:53 +01:00
parent c3681af783
commit c4867196b4
1 changed files with 35 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

40
examples/server/server.cpp
View file

@ -2004,12 +2004,15 @@ static void server_print_usage(const char *argv0, const gpt_params &params,
#ifdef LLAMA_SUPPORTS_GPU_OFFLOAD #ifdef LLAMA_SUPPORTS_GPU_OFFLOAD
printf(" -ngl N, --n-gpu-layers N\n"); printf(" -ngl N, --n-gpu-layers N\n");
printf(" number of layers to store in VRAM\n"); printf(" number of layers to store in VRAM\n");
printf(" -sm SPLIT_MODE, --split-mode SPLIT_MODE\n");
printf(" how to split the model across multiple GPUs, one of:\n");
printf(" - none: use one GPU only\n");
printf(" - layer (default): split layers and KV across GPUs\n");
printf(" - row: split rows across GPUs\n");
printf(" -ts SPLIT --tensor-split SPLIT\n"); printf(" -ts SPLIT --tensor-split SPLIT\n");
printf(" how to split tensors across multiple GPUs, comma-separated list of proportions, e.g. 3,1\n"); printf(" fraction of the model to offload to each GPU, comma-separated list of proportions, e.g. 3,1\n");
printf(" -mg i, --main-gpu i the GPU to use for scratch and small tensors\n"); printf(" -mg i, --main-gpu i the GPU to use for the model (with split-mode = none),\n");
printf(" -nommq, --no-mul-mat-q\n"); printf(" or for intermediate results and KV (with split-mode = row)\n");
printf(" use cuBLAS instead of custom mul_mat_q CUDA kernels.\n");
printf(" Not recommended since this is both slower and uses more VRAM.\n");
#endif #endif
printf(" -m FNAME, --model FNAME\n"); printf(" -m FNAME, --model FNAME\n");
printf(" model path (default: %s)\n", params.model.c_str()); printf(" model path (default: %s)\n", params.model.c_str());
@ -2230,6 +2233,33 @@ static void server_params_parse(int argc, char **argv, server_params &sparams,
"See main README.md for information on enabling GPU BLAS support", "See main README.md for information on enabling GPU BLAS support",
{{"n_gpu_layers", params.n_gpu_layers}}); {{"n_gpu_layers", params.n_gpu_layers}});
#endif #endif
}
else if (arg == "--split-mode" || arg == "-sm")
{
if (++i >= argc) {
invalid_param = true;
break;
}
std::string arg_next = argv[i];
if (arg_next == "none")
{
params.split_mode = LLAMA_SPLIT_NONE;
}
else if (arg_next == "layer")
{
params.split_mode = LLAMA_SPLIT_LAYER;
}
else if (arg_next == "row")
{
params.split_mode = LLAMA_SPLIT_ROW;
}
else {
invalid_param = true;
break;
}
#ifndef GGML_USE_CUBLAS
fprintf(stderr, "warning: llama.cpp was compiled without cuBLAS. Setting the split mode has no effect.\n");
#endif // GGML_USE_CUBLAS
} }
else if (arg == "--tensor-split" || arg == "-ts") else if (arg == "--tensor-split" || arg == "-ts")
{ {