CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-03 09:22:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

parallel : add disabled experimental batch chunking in powers of two

Browse Source
This commit is contained in:
Georgi Gerganov
2023-09-20 20:14:05 +03:00
parent ded9b43cad
commit b2debf65f2
1 changed files with 7 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
examples/parallel/parallel.cpp
View File

@@ -253,6 +253,13 @@ int main(int argc, char ** argv) {
int32_t n_batch = params.n_batch; int32_t n_batch = params.n_batch;
for (int32_t i = 0; i < (int32_t) batch.n_tokens; i += n_batch) { for (int32_t i = 0; i < (int32_t) batch.n_tokens; i += n_batch) {
// experiment: process in powers of 2
//if (i + n_batch > (int32_t) batch.n_tokens && n_batch > 32) {
// n_batch /= 2;
// i -= n_batch;
// continue;
//}
const int32_t n_tokens = std::min(n_batch, (int32_t) (batch.n_tokens - i)); const int32_t n_tokens = std::min(n_batch, (int32_t) (batch.n_tokens - i));
llama_batch batch_view = { llama_batch batch_view = {