July 2006
NEC Corporation

New in MathKeisan 2.0.0

• BLAS library tuned for the latest SX-8 family of processors
  • Significant improvements in the serial and parallel subroutines CGEMV, ZGEMV, CSYMV, ZSYMV, CHEMV, ZHEMV, SGEMM, DGEMM, CGEMM, ZGEMM
• BLAS library fixed to handle NaN arguments correctly
• Performance improvement for LAPACK subroutines SGETRF, DGETRF, CGETRF, ZGETRF
• Added LAPACK subroutines STTQRE, DTTQRE, CTTQRE, ZTTQRE
• ScaLAPACK library updated to version 1.7.4
• Performance improvements in ScaLAPACK subroutines PSGETRS, PDGETRS, PCGETRS, PZGETRS
• Corrected instability in ScaLAPACK subroutine PDSYEVX
• Performance improvements in subroutines SFFTS, DFFTS, CFFTS, ZFFTS for the inverse transform when data is contiguous
• Correction to interface subroutines SFFTS, DFFTS, CFFTS, and ZFFTS when iopt=0
• Memory reduction and improved parallel scaling for FFT and PARFFT subroutines SCFFT3D, CSFFT3D, DZFFT3D, ZDFFTED
• Correction of accuracy error for FFT and PARFFT subroutines DZFFTM and ZDFFTM
• Performance improvement for subroutines RFFTMLT and DFFTMLT

Installation

The MathKeisan 2.0.0 distribution for SX contains these files:

• MK2_0_0.pkg SUPER-UX package manager file
• MK2_0_0.tar tar file
• install.sh installation script
• README information on how to install MathKeisan

User's Guide

For information about using MathKeisan, please see the MathKeisan User's Guide (also available at www.mathkeisan.com).

Contact

If you have any questions or feedback about MathKeisan, please contact us at www.mathkeisan.com.