root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 03:14:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

ggml : fix redefinition of vaddvq_f32 for 32-bit ARM (#6906)

Browse Source
This commit is contained in:
Georgi Gerganov 2024-04-25 15:48:25 +03:00 committed by GitHub
parent 4ab99d8d47
commit 51543729ff
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 0 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
ggml.c
View File

@ -858,18 +858,6 @@ ggml_type_traits_t ggml_internal_get_type_traits(enum ggml_type type) {
// simd mappings // simd mappings
// //
#if defined(__ARM_NEON)
#if !defined(__aarch64__)
// 64-bit compatibility
inline static float vaddvq_f32(float32x4_t v) {
return vgetq_lane_f32(v, 0) + vgetq_lane_f32(v, 1) + vgetq_lane_f32(v, 2) + vgetq_lane_f32(v, 3);
}
#endif
#endif
// we define a common set of C macros which map to specific intrinsics based on the current architecture // we define a common set of C macros which map to specific intrinsics based on the current architecture
// we then implement the fundamental computation operations below using only these macros // we then implement the fundamental computation operations below using only these macros
// adding support for new architectures requires to define the corresponding SIMD macros // adding support for new architectures requires to define the corresponding SIMD macros