| Mathematics, MATLAB® Version 7.6 (R2008a) Release Notes | |
New features and changes introduced in this version are:
MATLAB® software now uses new versions of the Basic Linear Algebra Subroutine (BLAS) libraries. For Windows®, Intel® Mac, and Linux® platforms, MATLAB software supports the Intel Math Kernel Library (MKL) version 9.1. For the Solaris™ platform, MATLAB software uses the Sun Performance Library from Sun Studio 12.
MATLAB software now uses Version 3.1.1 of the Linear Algebra Package (LAPACK) library.
MATLAB software uses Linear Algebra Package (LAPACK) for linear algebra computations. At start-up, MATLAB software selects the LAPACK library to use.
If you want to take advantage of the potential performance enhancements provided by a custom LAPACK library on your computer, set the value of the environment variable LAPACK_VERSION to the name of the custom library. MATLAB uses the LAPACK library specified by this environment variable, if it exists.
The mechanism for changing the LAPACK library is similar to changing the BLAS library, which is described in Solution Solution 1-18QUC.
For example, on a LINUX machine, to set the LAPACK_VERSION environment variable, enter the following command at the LINUX prompt:
% setenv LAPACK_VERSION mllapack.so
Then start MATLAB as usual.
MATLAB software uses the Fastest Fourier Transform in the West (FFTW) libraries to speed up discrete Fourier transform. At start-up, MATLAB selects the FFTW libraries to use.
If you want to take advantage of the potential performance enhancements provided by custom FFTW libraries on your computer, set the value of the environment variable FFTW_VERSION to the name of the custom libraries. MATLAB uses the FFTW specified by this environment variable, if it exists.
The mechanism for changing the FFTW libraries is similar to changing the BLAS library, which is described in Solution 1-18QUC.
For example, on a LINUX machine, to set the FFTW_VERSION environment variable, enter the following command at the LINUX prompt:
% setenv FFTW_VERSION libfftw3i.so,libfftw3f.so
Then start MATLAB as usual.
Note When specifying FFTW libraries, you need to specify libraries for both single-precision and double-precision FFT. In the example above, libfftw3i.so contains double-precision FFT routines, and libfftw3f.so contains single-precision FFT routines. |
Multithreaded support has been added to elementwise math functions that may generate warnings: rdivide, ldivide, log, log2, and rem.
The ldl, logm, and funm functions include new algorithms based on recent numerical methods research.
| Function or Property Name | What Happens When You Use Function or Property? | Use This Instead | Compatibility Considerations |
|---|---|---|---|
| betacore | Errors | betainc | Replace all existing instances of betacore with betainc. |
| colmmd | Errors | colamd | Replace all existing instances of colmmd with colamd. |
| flops | Errors | None | Remove all existing instances of flops. With the incorporation of LAPACK in MATLAB version 6, counting floating-point operations is no longer practical. |
| symmmd | Errors | symamd | Replace all existing instances of symmd with symamd. |
| quad8 | Errors | quadl | Replace all existing instances of quad8 with quadl. |
| table1 | Errors | interp1 or interp1q | Replace all existing instances of table1 with interp1 or interp1q. |
| table2 | Errors | interp2 | Replace all existing instances of table2 with interp2. |
| bessela | Errors | besselj | Replace all existing instances of bessela with besselj. |
| beta using three input arguments | Errors | betainc using three input arguments | Replace all existing instances of beta using three input arguments with betainc using three input arguments. |
| Desktop Tools and Development Environment, MATLAB® Version 7.6 (R2008a) | Data Analysis, MATLAB® Version 7.6 (R2008a) | |
| © 1984-2008- The MathWorks, Inc. - Site Help - Patents - Trademarks - Privacy Policy - Preventing Piracy - RSS |