Adam Caldwell 9015b6c51d Enable AVX-VNNI 256-bit path for Q8_0 R8 dot product (#1459) ... Replace maddubs_epi16 + madd_epi16 with dpbusd_epi32 in the mul_mat_q8_0_r8_q8_2 dot product lambda when HAVE_VNNI256 is defined. Same sign trick operands (abs(x), sign(y,x)), just fewer instructions per sub-block. Co-authored-by: Adam Caldwell <accaldwell@users.noreply.github.com>		2026-03-18 11:14:23 +01:00
..
cmake	Merge mainline llama.cpp (#3)	2024-07-27 07:55:01 +02:00
include	Fix clang warnings on macOS (#1354)	2026-03-03 16:27:16 +01:00
src	Enable AVX-VNNI 256-bit path for Q8_0 R8 dot product (#1459)	2026-03-18 11:14:23 +01:00
.gitignore	Merge mainline llama.cpp (#3)	2024-07-27 07:55:01 +02:00
CMakeLists.txt	Remove llamafile remnants (#1179)	2026-01-22 13:20:23 +02:00