Skip to content

Commit ebfb137

Browse files
committed
Update on "[Executorch][llama] Enable quantized sdpa"
Enable leveraging quantized sdpa op when quantized kv cache is used. Instead of adding yet another arg, at the moment I have chosen to leverage quantize_kv_cache option. Differential Revision: [D71833064](https://our.internmc.facebook.com/intern/diff/D71833064/) [ghstack-poisoned]
2 parents f744daf + 1d6c70c commit ebfb137

1 file changed

Lines changed: 1 addition & 0 deletions

File tree

extension/llm/custom_ops/CMakeLists.txt

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -21,6 +21,7 @@ if(NOT EXECUTORCH_ROOT)
2121
endif()
2222

2323
set(_common_compile_options -Wno-deprecated-declarations -fPIC)
24+
list(APPEND _common_compile_options "-march=armv8.2-a+dotprod")
2425

2526
include(${EXECUTORCH_ROOT}/tools/cmake/Utils.cmake)
2627
include(${EXECUTORCH_ROOT}/tools/cmake/Codegen.cmake)

0 commit comments

Comments
 (0)