thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

llama : disable Direct IO by default (#19109)

Browse source 
* llama : disable Direct IO by default

* cont : override mmap if supported
This commit is contained in:
Georgi Gerganov 2026-01-28 09:11:13 +02:00 • committed by GitHub
parent eef375ce16
commit c5c64f72ac
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 10 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

2
src/llama-quant.cpp
View file

@ -545,7 +545,7 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
}
std::vector<std::string> splits = {};
llama_model_loader ml(fname_inp, splits, use_mmap, /*use_direct_io*/ true, /*check_tensors*/ true, /*no_alloc*/ false, kv_overrides, nullptr);
llama_model_loader ml(fname_inp, splits, use_mmap, /*use_direct_io*/ false, /*check_tensors*/ true, /*no_alloc*/ false, kv_overrides, nullptr);
ml.init_mappings(false); // no prefetching
llama_model model(llama_model_default_params());