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

rm redundant clamp

Browse source
This commit is contained in:
Xuan Son Nguyen 2025-01-27 16:36:20 +01:00
parent 226d59270b
commit 9517aee23c
1 changed files with 2 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

10
ggml/src/ggml-cpu/ggml-cpu-quants.c
View file

@ -1722,14 +1722,8 @@ void quantize_row_q8_K(const float * restrict x, void * restrict y, int64_t k) {
// Pack into 16 i8 values
v128_t i8 = wasm_i8x16_narrow_i16x8(
wasm_i16x8_narrow_i32x4(
wasm_i32x4_min(wasm_i32x4_max(i0, wasm_i32x4_splat(-127)), wasm_i32x4_splat(127)),
wasm_i32x4_min(wasm_i32x4_max(i1, wasm_i32x4_splat(-127)), wasm_i32x4_splat(127))
),
wasm_i16x8_narrow_i32x4(
wasm_i32x4_min(wasm_i32x4_max(i2, wasm_i32x4_splat(-127)), wasm_i32x4_splat(127)),
wasm_i32x4_min(wasm_i32x4_max(i3, wasm_i32x4_splat(-127)), wasm_i32x4_splat(127))
)
wasm_i16x8_narrow_i32x4(i0, i1),
wasm_i16x8_narrow_i32x4(i2, i3)
);
wasm_v128_store(yc[i].qs + j, i8);