YourWishes/cglm
1
0
Fork 0
mirror of https://github.com/recp/cglm.git synced 2026-02-17 03:39:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

simd128: handle both sse2 and simd128 enabled by Emscripten

Browse Source 
https://github.com/recp/cglm/pull/286#issuecomment-1492985403
This commit is contained in:
myfreeer
2023-04-02 12:39:20 +08:00
parent 7ca9a64ecf
commit 3a9e4df393
8 changed files with 171 additions and 170 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
include/cglm/mat2.h
View File

@@ -136,10 +136,10 @@ glm_mat2_zero(mat2 mat) {
CGLM_INLINE
void
glm_mat2_mul(mat2 m1, mat2 m2, mat2 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mat2_mul_sse2(m1, m2, dest);
#elif defined(__wasm__) && defined(__wasm_simd128__)
#if defined(__wasm__) && defined(__wasm_simd128__)
glm_mat2_mul_wasm(m1, m2, dest);
#elif defined( __SSE__ ) || defined( __SSE2__ )
glm_mat2_mul_sse2(m1, m2, dest);
#elif defined(CGLM_NEON_FP)
glm_mat2_mul_neon(m1, m2, dest);
#else
@@ -166,10 +166,10 @@ glm_mat2_mul(mat2 m1, mat2 m2, mat2 dest) {
CGLM_INLINE
void
glm_mat2_transpose_to(mat2 m, mat2 dest) {
#if defined( __SSE__ ) || defined( __SSE2__ )
glm_mat2_transp_sse2(m, dest);
#elif defined(__wasm__) && defined(__wasm_simd128__)
#if defined(__wasm__) && defined(__wasm_simd128__)
glm_mat2_transp_wasm(m, dest);
#elif defined( __SSE__ ) || defined( __SSE2__ )
glm_mat2_transp_sse2(m, dest);
#else
dest[0][0] = m[0][0];
dest[0][1] = m[1][0];
@@ -230,10 +230,11 @@ glm_mat2_trace(mat2 m) {
CGLM_INLINE
void
glm_mat2_scale(mat2 m, float s) {
#if defined( __SSE__ ) || defined( __SSE2__ )
#if defined(__wasm__) && defined(__wasm_simd128__)
glmm_store(m[0], wasm_f32x4_mul(wasm_v128_load(m[0]),
wasm_f32x4_splat(s)));
#elif defined( __SSE__ ) || defined( __SSE2__ )
glmm_store(m[0], _mm_mul_ps(_mm_loadu_ps(m[0]), _mm_set1_ps(s)));
#elif defined(__wasm__) && defined(__wasm_simd128__)
glmm_store(m[0], wasm_f32x4_mul(wasm_v128_load(m[0]), wasm_f32x4_splat(s)));
#elif defined(CGLM_NEON_FP)
vst1q_f32(m[0], vmulq_f32(vld1q_f32(m[0]), vdupq_n_f32(s)));
#else