Home Explore Help
Sign In
cturan
/
llama.cpp
mirror of https://github.com/cturan/llama.cpp
1
0
Fork 0
Files Issues 0 Wiki
Browse Source

ggml-cpu: reduce asm calls for hsum (#14037)

Signed-off-by: Aaron Teo <aaron.teo1@ibm.com>
Aaron Teo 7 months ago
parent
6231c5cd6d
commit
50d2227953
1 changed files with 2 additions and 4 deletions
Unified View Show Diff Stats
  1. 2 4
      ggml/src/ggml-cpu/simd-mappings.h

+ 2 - 4
ggml/src/ggml-cpu/simd-mappings.h
View File 

@@ -944,10 +944,8 @@ static inline void __lsx_f16x4_store(ggml_fp16_t * x, __m128 y) {
 
     for (int i = 0; i < offset; ++i) {              \
 
     for (int i = 0; i < offset; ++i) {              \
 
         x[i] = vec_add(x[i], x[offset + i]);        \
 
         x[i] = vec_add(x[i], x[offset + i]);        \
 
     }                                               \
 
     }                                               \
 
-    res = vec_extract(x[0], 0) +                    \
 
-          vec_extract(x[0], 1) +                    \
 
-          vec_extract(x[0], 2) +                    \
 
-          vec_extract(x[0], 3);                     \
 
+    float32x4_t tmp = x[0] + vec_reve(x[0]);        \
 
+    res = tmp[0] + tmp[1];                          \
 
 }
 
 }
 
 
 
 #define GGML_F32_VEC        GGML_F32x4
 
 #define GGML_F32_VEC        GGML_F32x4