diff options
| author | Brad Payne (Vir Linden) <vir@lindenlab.com> | 2016-09-28 15:43:07 -0400 |
|---|---|---|
| committer | Brad Payne (Vir Linden) <vir@lindenlab.com> | 2016-09-28 15:43:07 -0400 |
| commit | 8d6bb9ee4e9def306454e9ecda1980793dcf7ad5 (patch) | |
| tree | f7c4bbbb98f7deac7ea54d6dca49fced89f64d23 /indra/llmath | |
| parent | e6297ab3d6262014d3b501703154437bea523be1 (diff) | |
SL-451 - SSE-optimized matrix multiply, used in initSkinningMatrixPalette()
Diffstat (limited to 'indra/llmath')
| -rw-r--r-- | indra/llmath/llmatrix4a.h | 23 |
1 files changed, 23 insertions, 0 deletions
diff --git a/indra/llmath/llmatrix4a.h b/indra/llmath/llmatrix4a.h index e11fa1bf72..216334752a 100644 --- a/indra/llmath/llmatrix4a.h +++ b/indra/llmath/llmatrix4a.h @@ -153,4 +153,27 @@ public: } }; +inline LLVector4a rowMul(const LLVector4a &row, const LLMatrix4a &mat) +{ + LLVector4a result; + result = _mm_mul_ps(_mm_shuffle_ps(row, row, _MM_SHUFFLE(0, 0, 0, 0)), mat.mMatrix[0]); + result = _mm_add_ps(result, _mm_mul_ps(_mm_shuffle_ps(row, row, _MM_SHUFFLE(1, 1, 1, 1)), mat.mMatrix[1])); + result = _mm_add_ps(result, _mm_mul_ps(_mm_shuffle_ps(row, row, _MM_SHUFFLE(2, 2, 2, 2)), mat.mMatrix[2])); + result = _mm_add_ps(result, _mm_mul_ps(_mm_shuffle_ps(row, row, _MM_SHUFFLE(3, 3, 3, 3)), mat.mMatrix[3])); + return result; +} + +inline void matMul(const LLMatrix4a &a, const LLMatrix4a &b, LLMatrix4a &res) +{ + LLVector4a row0 = rowMul(a.mMatrix[0], b); + LLVector4a row1 = rowMul(a.mMatrix[1], b); + LLVector4a row2 = rowMul(a.mMatrix[2], b); + LLVector4a row3 = rowMul(a.mMatrix[3], b); + + res.mMatrix[0] = row0; + res.mMatrix[1] = row1; + res.mMatrix[2] = row2; + res.mMatrix[3] = row3; +} + #endif |