vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ggml : add loongarch lsx and lasx support (#6454)

Browse source 
* add loongarch lsx and lasx optimize code

* Add loongarch compilation support to makefile

* revert stb_image.h

* opt bytes_from_nibbles_32 and sum_i16_pairs_float

* fix undeclared

* format code

* update

* update 2

---------

Co-authored-by: Jinyang He <hejinyang@loongson.cn>
This commit is contained in:
junchao-loongson 2024-05-20 15:19:21 +08:00 • committed by GitHub
parent 1cc0155d04
commit 65c58207ec
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 2316 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

28
ggml-impl.h
View file

@ -455,6 +455,34 @@ static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) {
#include <riscv_vector.h>
#endif
#if defined(__loongarch64)
#if defined(__loongarch_asx)
#include <lasxintrin.h>
#endif
#if defined(__loongarch_sx)
#include <lsxintrin.h>
#endif
#endif
#if defined(__loongarch_asx)
typedef union {
int32_t i;
float f;
} ft_union;
/* float type data load instructions */
static __m128 __lsx_vreplfr2vr_s(float val) {
ft_union fi_tmpval = {.f = val};
return (__m128)__lsx_vreplgr2vr_w(fi_tmpval.i);
}
static __m256 __lasx_xvreplfr2vr_s(float val) {
ft_union fi_tmpval = {.f = val};
return (__m256)__lasx_xvreplgr2vr_w(fi_tmpval.i);
}
#endif
#ifdef __F16C__
#ifdef _MSC_VER