GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-06-04 18:54:00 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix alignement of statically allocated temporaries in gemv.

Browse Source
This commit is contained in:
Gael Guennebaud 2016-09-17 12:52:27 +02:00
parent ce3557ca69
commit 4cc2c73e6a
2 changed files with 11 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
Eigen/src/Core/GeneralProduct.h
View File

@ -159,20 +159,20 @@ struct gemv_static_vector_if<Scalar,Size,Dynamic,true>
template<typename Scalar,int Size,int MaxSize> template<typename Scalar,int Size,int MaxSize>
struct gemv_static_vector_if<Scalar,Size,MaxSize,true> struct gemv_static_vector_if<Scalar,Size,MaxSize,true>
{ {
#if EIGEN_MAX_STATIC_ALIGN_BYTES!=0
internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize),0> m_data;
EIGEN_STRONG_INLINE Scalar* data() { return m_data.array; }
#else
// Some architectures cannot align on the stack,
// => let's manually enforce alignment by allocating more data and return the address of the first aligned element.
enum { enum {
ForceAlignment = internal::packet_traits<Scalar>::Vectorizable, ForceAlignment = internal::packet_traits<Scalar>::Vectorizable,
PacketSize = internal::packet_traits<Scalar>::size PacketSize = internal::packet_traits<Scalar>::size
}; };
internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize)+(ForceAlignment?PacketSize:0),0> m_data; #if EIGEN_MAX_STATIC_ALIGN_BYTES!=0
internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize),0,EIGEN_PLAIN_ENUM_MIN(AlignedMax,PacketSize)> m_data;
EIGEN_STRONG_INLINE Scalar* data() { return m_data.array; }
#else
// Some architectures cannot align on the stack,
// => let's manually enforce alignment by allocating more data and return the address of the first aligned element.
internal::plain_array<Scalar,EIGEN_SIZE_MIN_PREFER_FIXED(Size,MaxSize)+(ForceAlignment?EIGEN_MAX_ALIGN_BYTES:0),0> m_data;
EIGEN_STRONG_INLINE Scalar* data() { EIGEN_STRONG_INLINE Scalar* data() {
return ForceAlignment return ForceAlignment
? reinterpret_cast<Scalar*>((internal::UIntPtr(m_data.array) & ~(size_t(EIGEN_MAX_ALIGN_BYTES-1))) + EIGEN_MAX_ALIGN_BYTES) ? reinterpret_cast<Scalar*>((internal::UIntPtr(m_data.array) & ~(std::size_t(EIGEN_MAX_ALIGN_BYTES-1))) + EIGEN_MAX_ALIGN_BYTES)
: m_data.array; : m_data.array;
} }
#endif #endif
@ -207,7 +207,7 @@ template<> struct gemv_dense_selector<OnTheRight,ColMajor,true>
typedef internal::blas_traits<Rhs> RhsBlasTraits; typedef internal::blas_traits<Rhs> RhsBlasTraits;
typedef typename RhsBlasTraits::DirectLinearAccessType ActualRhsType; typedef typename RhsBlasTraits::DirectLinearAccessType ActualRhsType;
typedef Map<Matrix<ResScalar,Dynamic,1>, Aligned> MappedDest; typedef Map<Matrix<ResScalar,Dynamic,1>, EIGEN_PLAIN_ENUM_MIN(AlignedMax,internal::packet_traits<ResScalar>::size)> MappedDest;
ActualLhsType actualLhs = LhsBlasTraits::extract(lhs); ActualLhsType actualLhs = LhsBlasTraits::extract(lhs);
ActualRhsType actualRhs = RhsBlasTraits::extract(rhs); ActualRhsType actualRhs = RhsBlasTraits::extract(rhs);

3
test/product_small.cpp
View File

@ -213,7 +213,8 @@ void test_product_small()
{ {
for(int i = 0; i < g_repeat; i++) { for(int i = 0; i < g_repeat; i++) {
CALL_SUBTEST_1( product(Matrix<float, 3, 2>()) ); CALL_SUBTEST_1( product(Matrix<float, 3, 2>()) );
CALL_SUBTEST_2( product(Matrix<int, 3, 5>()) ); CALL_SUBTEST_2( product(Matrix<int, 3, 17>()) );
CALL_SUBTEST_8( product(Matrix<double, 3, 17>()) );
CALL_SUBTEST_3( product(Matrix3d()) ); CALL_SUBTEST_3( product(Matrix3d()) );
CALL_SUBTEST_4( product(Matrix4d()) ); CALL_SUBTEST_4( product(Matrix4d()) );
CALL_SUBTEST_5( product(Matrix4f()) ); CALL_SUBTEST_5( product(Matrix4f()) );