llama.cpp/scripts
Max Krasnyansky 3eb2be1ca5
Hexagon Op queue & dispatch optimizations (#16820)
* hexagon: remove dspqueue callbacks and do all read processing inplace

* hexagon: there is no need to ref/deref the buffers at this point

We're not going to release the buffers without flushing the session queue.
So there is no need to inc/dec the refcounts for every request.
We also don't need to include those bufs in the response.

* hexagon: bump the thread count in the adb wrapper scripts

We can use more CPU cores now that the dedicated dspqueue polling threads are not used (ie no contention).
Also enable more agressive polling for now since we still map Flash Attention (and a few other kernels) to
the CPU and those dspqueue threads were keeping the CPU cores are higher clock freqs.

* hexagon: add lhez as the second code owner
2025-10-29 06:29:12 -07:00
..
apple scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
jinja scripts : add Jinja tester PySide6 simple app (#15756) 2025-09-05 01:05:12 +02:00
snapdragon Hexagon Op queue & dispatch optimizations (#16820) 2025-10-29 06:29:12 -07:00
build-info.sh llama : reorganize source code + improve CMake (#8006) 2024-06-26 18:33:02 +03:00
check-requirements.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
compare-commits.sh scripts: add sqlite3 check for compare-commits.sh (#15633) 2025-08-28 19:23:22 +08:00
compare-llama-bench.py scripts: strip "AMD Instinct" from GPU name (#15668) 2025-08-29 22:04:08 +02:00
create_ops_docs.py Docs: add instructions for adding backends (#14889) 2025-07-27 09:36:43 +08:00
debug-test.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
fetch_server_test_models.py llama : move end-user examples to tools directory (#13249) 2025-05-02 20:27:13 +02:00
gen-authors.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
gen-unicode-data.py py : type-check all Python scripts with Pyright (#8341) 2024-07-07 15:04:39 -04:00
get-flags.mk build : pass all warning flags to nvcc via -Xcompiler (#5570) 2024-02-18 16:21:52 -05:00
get-hellaswag.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
get-pg.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
get-wikitext-2.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
get-wikitext-103.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
get-winogrande.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
get_chat_template.py scripts: corrected encoding when getting chat template (#11866) (#11907) 2025-02-18 10:30:16 +01:00
hf.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
install-oneapi.bat support SYCL backend windows build (#5208) 2024-01-31 08:08:07 +05:30
server-bench.py llama: use FA + max. GPU layers by default (#15434) 2025-08-30 16:32:10 +02:00
sync-ggml-am.sh scripts : update sync scripts 2025-08-18 22:06:44 +03:00
sync-ggml.last ggml : bump version to 0.9.4 (ggml/1363) 2025-09-30 13:53:55 +03:00
sync-ggml.sh scripts : update sync scripts 2025-08-18 22:06:44 +03:00
sync_vendor.py sync : vendor (#13901) 2025-05-30 16:25:45 +03:00
tool_bench.py server : speed up tests (#15836) 2025-09-06 14:45:24 +02:00
tool_bench.sh scripts : make the shell scripts cross-platform (#14341) 2025-06-30 10:17:18 +02:00
verify-checksum-models.py convert.py : add python logging instead of print() (#6511) 2024-05-03 22:36:41 +03:00
xxd.cmake llama : move end-user examples to tools directory (#13249) 2025-05-02 20:27:13 +02:00