devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

First part of a big refactoring of alignment control to enable the handling of arbitrarily aligned buffers. It includes:

Browse Source 
- AlignedBit flag is deprecated. Alignment is now specified by the evaluator through the 'Alignment' enum, e.g., evaluator<Xpr>::Alignment. Its value is in Bytes.
 - Add several enums to specify alignment: Aligned8, Aligned16, Aligned32, Aligned64, Aligned128. AlignedMax corresponds to EIGEN_MAX_ALIGN_BYTES. Such enums are used to define the above Alignment value, and as the 'Options' template parameter of Map<> and Ref<>.
 - The Aligned enum is now deprecated. It is now an alias for Aligned16.
 - Currently, traits<Matrix<>>, traits<Array<>>, traits<Ref<>>, traits<Map<>>, and traits<Block<>> also expose the Alignment enum.
This commit is contained in:
Gael Guennebaud
2015-08-06 15:31:07 +02:00
parent 65186ef18d
commit 1f5024332e
24 changed files with 231 additions and 200 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
Eigen/src/Geometry/arch/Geometry_SSE.h
View File

@@ -16,14 +16,14 @@ namespace Eigen {
namespace internal {
template<class Derived, class OtherDerived>
struct quat_product<Architecture::SSE, Derived, OtherDerived, float, Aligned>
struct quat_product<Architecture::SSE, Derived, OtherDerived, float, Aligned16>
{
static inline Quaternion<float> run(const QuaternionBase<Derived>& _a, const QuaternionBase<OtherDerived>& _b)
{
Quaternion<float> res;
const __m128 mask = _mm_setr_ps(0.f,0.f,0.f,-0.f);
__m128 a = _a.coeffs().template packet<Aligned>(0);
__m128 b = _b.coeffs().template packet<Aligned>(0);
__m128 a = _a.coeffs().template packet<Aligned16>(0);
__m128 b = _b.coeffs().template packet<Aligned16>(0);
__m128 s1 = _mm_mul_ps(vec4f_swizzle1(a,1,2,0,2),vec4f_swizzle1(b,2,0,1,2));
__m128 s2 = _mm_mul_ps(vec4f_swizzle1(a,3,3,3,1),vec4f_swizzle1(b,0,1,2,1));
pstore(&res.x(),
@@ -55,8 +55,8 @@ struct cross3_impl<Architecture::SSE,VectorLhs,VectorRhs,float,true>
static inline typename plain_matrix_type<VectorLhs>::type
run(const VectorLhs& lhs, const VectorRhs& rhs)
{
__m128 a = lhs.template packet<VectorLhs::Flags&AlignedBit ? Aligned : Unaligned>(0);
__m128 b = rhs.template packet<VectorRhs::Flags&AlignedBit ? Aligned : Unaligned>(0);
__m128 a = lhs.template packet<traits<VectorLhs>::Alignment>(0);
__m128 b = rhs.template packet<traits<VectorRhs>::Alignment>(0);
__m128 mul1=_mm_mul_ps(vec4f_swizzle1(a,1,2,0,3),vec4f_swizzle1(b,2,0,1,3));
__m128 mul2=_mm_mul_ps(vec4f_swizzle1(a,2,0,1,3),vec4f_swizzle1(b,1,2,0,3));
typename plain_matrix_type<VectorLhs>::type res;