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

reorder operations

Browse source
This commit is contained in:
ct-clmsn 2023-12-28 12:53:43 -05:00
parent 2e3d229597
commit c9c4e1f077
1 changed files with 6 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

7
llama.cpp
View file

@ -8936,12 +8936,17 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
return hpx::make_ready_future<void>(); return hpx::make_ready_future<void>();
}; };
hpx::future<void> this_fut = computefn(0, counters[0], thread_local_hist, local_sizes);
for (int it = 1; it < nthread_use - 1; ++it) { for (int it = 1; it < nthread_use - 1; ++it) {
futures.push_back(hpx::run_as_hpx_thread(computefn, it, counters[it], thread_local_hist, local_sizes)); futures.push_back(hpx::run_as_hpx_thread(computefn, it, counters[it], thread_local_hist, local_sizes));
} }
hpx::future<void> this_fut =
computefn(0, counters[0], thread_local_hist, local_sizes);
hpx::wait_all(futures); hpx::wait_all(futures);
this_fut.wait(); this_fut.wait();
for(auto & local_hist : thread_local_hist) { for(auto & local_hist : thread_local_hist) {
for(auto j = 0; j < int(local_hist.size()); ++j) { for(auto j = 0; j < int(local_hist.size()); ++j) {
hist_cur[j] += local_hist[j]; hist_cur[j] += local_hist[j];