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

rpc : make RPC servers come first in the device list (#9296)

Browse source 
* rpc : make RPC servers come first in the device list

* rpc : disable options for non-RPC builds

* rpc : rpc_count always zero for non-RPC builds
This commit is contained in:
Radoslav Gerganov 2024-09-04 11:08:32 +03:00 • committed by GitHub
parent 9379d3cc17
commit 82e3b03c11
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 50 additions and 33 deletions
Download patch file Download diff file Expand all files Collapse all files

4
common/common.cpp
View file

@ -1234,11 +1234,13 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
#endif // GGML_USE_CUDA_SYCL_VULKAN
return true;
}
#ifdef GGML_USE_RPC
if (arg == "--rpc") {
CHECK_ARG
params.rpc_servers = argv[i];
return true;
}
#endif
if (arg == "--no-mmap") {
params.use_mmap = false;
return true;
@ -1929,7 +1931,9 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "*", " --image FILE", "path to an image file. use with multimodal models. Specify multiple times for batching" });
options.push_back({ "backend" });
#ifdef GGML_USE_RPC
options.push_back({ "*", " --rpc SERVERS", "comma separated list of RPC servers" });
#endif
if (llama_supports_mlock()) {
options.push_back({ "*", " --mlock", "force system to keep model in RAM rather than swapping or compressing" });