mpblas.f
SUBROUTINE MPBLAS (I1)
C***BEGIN PROLOGUE MPBLAS
....
....
Warning: this routine is not intended to be user-callable.
....
....
C***SUBSIDIARY
C***PURPOSE Subsidiary to DQDOTA and DQDOTI
C***LIBRARY SLATEC
C***TYPE ALL (MPBLAS-A)
C***AUTHOR (UNKNOWN)
C***DESCRIPTION
C
C This subroutine is called to set up Brent's 'mp' package
C for use by the extended precision inner products from the BLAS.
C
C In the SLATEC library we require the Extended Precision MP number
C to have a mantissa twice as long as Double Precision numbers.
C The calculation of MPT (and MPMXR which is the actual array size)
C in this routine will give 2x (or slightly more) on the machine
C that we are running on. The INTEGER array size of 30 was chosen
C to be slightly longer than the longest INTEGER array needed on
C any machine that we are currently aware of.
C
C***SEE ALSO DQDOTA, DQDOTI
C***REFERENCES R. P. Brent, A Fortran multiple-precision arithmetic
C package, ACM Transactions on Mathematical Software 4,
C 1 (March 1978), pp. 57-70.
C R. P. Brent, MP, a Fortran multiple-precision arithmetic
C package, Algorithm 524, ACM Transactions on Mathema-
C tical Software 4, 1 (March 1978), pp. 71-81.
C***ROUTINES CALLED I1MACH, XERMSG
C***COMMON BLOCKS MPCOM
C***REVISION HISTORY (YYMMDD)
C 791001 DATE WRITTEN
C 890531 Changed all specific intrinsics to generic. (WRB)
C 890531 REVISION DATE from Version 3.2
C 891214 Prologue converted to Version 4.0 format. (BAB)
C 900402 Added TYPE section. (WRB)
C 920501 Reformatted the REFERENCES section. (WRB)
C 930124 Increased Array size in MPCON for SUN -r8, and calculate
C size for Quad Precision for 2x DP. (RWC)
C***END PROLOGUE MPBLAS