Benoit Jacob 03d19f3bae quick temporary fix for a perf issue we just identified with ...

vectorization....
now the sum benchmark runs 3x faster with vectorization than without.

2008-06-23 11:23:05 +00:00

..

arch

* implement slice vectorization. Because it uses unaligned

2008-06-22 15:02:05 +00:00

util

split sum away from redux and vectorize it.

2008-06-23 10:32:48 +00:00

Assign.h

split sum away from redux and vectorize it.

2008-06-23 10:32:48 +00:00

Block.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

CacheFriendlyProduct.h

added an *optional* Eigen2 dynamic library.

2008-05-31 23:21:49 +00:00

CMakeLists.txt

Various fixes in:

2008-04-25 23:10:37 +00:00

Coeffs.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

CommaInitializer.h

* add CommaInitializer::finished to allow the use of (Matrix3() << v0, v1, v2).finished()

2008-06-03 15:50:09 +00:00

CoreInstantiations.cpp

Add QR and Cholesky module instantiations in the lib.

2008-06-14 13:02:41 +00:00

CwiseBinaryOp.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

CwiseNullaryOp.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

CwiseUnaryOp.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

DiagonalCoeffs.h

* more cleaning in Product

2008-06-19 23:00:51 +00:00

DiagonalMatrix.h

meagre outcome for so much time spent!

2008-06-01 03:36:49 +00:00

DiagonalProduct.h

* refactoring of Product:

2008-06-19 17:33:57 +00:00

Dot.h

* move some compile time "if" to their respective unroller (assign and dot)

2008-06-07 01:07:48 +00:00

DummyPacketMath.h

move arch-specific code to arch/SSE and arch/AltiVec subdirs.

2008-05-12 08:30:42 +00:00

Extract.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

Flagged.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

Functors.h

split sum away from redux and vectorize it.

2008-06-23 10:32:48 +00:00

Fuzzy.h

* rewrite of the QR decomposition:

2008-06-07 22:47:11 +00:00

InverseProduct.h

many small fixes and documentation improvements,

2008-05-29 03:12:30 +00:00

IO.h

* introducte recursive Flags system for the expressions

2008-03-30 18:43:22 +00:00

Map.h

* make the _Flags template parameter of Matrix default to the corrected

2008-06-13 07:53:45 +00:00

MathFunctions.h

-fix certain #includes

2008-05-12 21:15:17 +00:00

Matrix.h

Override MatrixBase::eval() since matrices don't need

2008-06-20 15:26:39 +00:00

MatrixBase.h

* added a pseudo expression Array giving access to:

2008-06-20 12:38:03 +00:00

MatrixStorage.h

fix two bugs dicovered by the previous commit.

2008-06-16 16:39:58 +00:00

Minor.h

Introduce generic Flagged xpr, remove already Lazy.h and Temporary.h

2008-05-14 08:20:15 +00:00

NestByValue.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

NumTraits.h

* add typedefs for matrices/vectors with LargeBit

2008-04-10 10:33:50 +00:00

Part.h

* move some compile time "if" to their respective unroller (assign and dot)

2008-06-07 01:07:48 +00:00

Product.h

* implement slice vectorization. Because it uses unaligned

2008-06-22 15:02:05 +00:00

Redux.h

split sum away from redux and vectorize it.

2008-06-23 10:32:48 +00:00

Sum.h

quick temporary fix for a perf issue we just identified with

2008-06-23 11:23:05 +00:00

Swap.h

* start of the Geometry module with a cross product and quaternion expressions

2008-06-02 22:58:36 +00:00

Transpose.h

* Block: row and column expressions in the inner direction

2008-06-16 14:54:31 +00:00

Visitor.h

* remove Cross product expression: MatrixBase::cross() now returns a temporary

2008-06-07 13:18:29 +00:00