guezoloic/t3dsr
1
0
Fork 0
mirror of https://github.com/guezoloic/t3dsr.git synced 2026-01-25 06:34:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: vec4f_dot for x64

Browse Source
This commit is contained in:
Loïc GUEZO
2025-06-22 13:51:55 +02:00
parent 7a1fb0daa3
commit 1c41293832
1 changed files with 10 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
src/math/vec4.c
View File

@@ -136,23 +136,18 @@ Vec4f_t vec4f_scale(Vec4f_t a, float scalar)
float vec4f_dot(Vec4f_t a, Vec4f_t b) float vec4f_dot(Vec4f_t a, Vec4f_t b)
{ {
float result;
#if defined (SIMD_X86) #if defined (SIMD_X86)
__m128 va = _mm_load_ps(a.data); __m128 va = _mm_load_ps(a.data);
__m128 vb = _mm_load_ps(b.data); __m128 vb = _mm_load_ps(b.data);
__m128 vmul = _mm_mul_ps(va, vb); __m128 vmul = _mm_mul_ps(va, vb);
// [y*y y*y, w*w, w*w] __m128 shuf = _mm_shuffle_ps(vmul, vmul, _MM_SHUFFLE(2, 3, 0, 1)); // [y, y, w, w]
__m128 shuf = _mm_movehdup_ps(vmul); __m128 sum = _mm_add_ps(vmul, shuf); // [x+y, y+y, z+w, w+w]
// [x*x+y*y, y*y+y*y, z*z+w*w, w*w+w*w]
__m128 sum = _mm_add_ps(vmul, shuf); shuf = _mm_movehl_ps(shuf, sum); // [z+w, w+w, w, w]
sum = _mm_add_ss(sum, shuf); // [x+y+z+w, y+y, z+w, w+w]
// [z*z+w*w, w*w+w*w, ?, ?]
shuf = _mm_movehl_ps(shuf, sum); return _mm_cvtss_f32(sum);
// [x*x+y*y+z*z+w*w, ?, ?, ?]
sum = _mm_add_ss(sums, shuf);
result = __mm_cvtss_f32(sum);
#elif defined (SIMD_ARCH) #elif defined (SIMD_ARCH)
float32x4_t va = vld1q_f32(a.data); float32x4_t va = vld1q_f32(a.data);
@@ -162,11 +157,10 @@ float vec4f_dot(Vec4f_t a, Vec4f_t b)
float32x2_t sum_pair = vadd_f32(vget_low_f32(vmul), vget_high_f32(vmul)); float32x2_t sum_pair = vadd_f32(vget_low_f32(vmul), vget_high_f32(vmul));
float32x2_t final_sum = vpadd_f32(sum_pair, sum_pair); float32x2_t final_sum = vpadd_f32(sum_pair, sum_pair);
result = vget_lane_f32(final_sum, 0); return vget_lane_f32(final_sum, 0);
#else #else
result = a.x * b.x + a.y * b.y + a.z * b.z + a.w * b.w; return a.x * b.x + a.y * b.y + a.z * b.z + a.w * b.w;
#endif #endif
return result;
} }
// float vec4_dot(Vec4_t a, Vec4_t b) // float vec4_dot(Vec4_t a, Vec4_t b)