DADiSP / MatrixXL
Matrix Acceleration Module
A matrix is a rectangular array of values that lies at the heart of a wide variety
of technical applications, including signal processing, medical,
geophysical, acoustic, statistics, image processing and many more.
Matrices are used to find solutions to systems of equations, perform
numeric optimizations and compute linear transformations. Matrix decompositions provide
simplified factorizations of matrices to implement efficient matrix computations for a
large class of matrix problems.
LAPACK (Linear Algebra PACKage) is a software library for numerical
linear algebra and provides industry standard routines for solving systems of linear
equations and linear least squares, eigenvalue problems, and
singular value decomposition. LAPACK also includes routines to
implement matrix factorizations such as LU, QR,
Cholesky and Schur decomposition.
Processor Tuned Performance
The MatrixXL Module is based on the LAPACK implementation from Intel's MKL
library to provide outstanding performance on Intel based
processors. MatrixXL automatically takes advantage of the latest instruction sets,
parallelism, and algorithms to yield a highly optimized matrix
functionality. Performance gains of 3x to 50x over the standard
built-in matrix functions are acheived.
MatrixXL is completely automatic, simply install the module and
matrix functions immediately run faster - no settings to change, no
code to rewrite. In addition to core matrix routines, any
custom or built-in function that relies on matrix processing experiences
the same performance gain.
MatrixXL is a straightforward way to accelerate any matrix based
data analysis and is the perfect complement to DADiSP/FFTXL
, the FFT Acceleration Module and
, the Vector Math
Acceleration Module. Or for the most cost effective solution to
speeding up nearly any numeric calculation, see the DADiSP/ProPac
module that combines all three
accelerators into a single, highly optimized numeric computation
engine to provide some of the most efficient technical data
analysis routines available today.
DADiSP/MatrixXL requires DADiSP 6.5 B04
for information about updating your current
version of DADiSP.