SCALAPACK 2.2.2
LAPACK: Linear Algebra PACKage
Loading...
Searching...
No Matches

◆ pderrscal()

subroutine pderrscal ( double precision  errbnd,
double precision  psclr,
double precision  x,
double precision  prec 
)

Definition at line 3245 of file pdblas1tst.f.

3246*
3247* -- PBLAS test routine (version 2.0) --
3248* University of Tennessee, Knoxville, Oak Ridge National Laboratory,
3249* and University of California, Berkeley.
3250* April 1, 1998
3251*
3252* .. Scalar Arguments ..
3253 DOUBLE PRECISION ERRBND, PREC, PSCLR, X
3254* ..
3255*
3256* Purpose
3257* =======
3258*
3259* PDERRSCAL serially computes the product PSCLR * X and returns a sca-
3260* led relative acceptable error bound on the result.
3261*
3262* Notes
3263* =====
3264*
3265* If s1 = PSCLR*X and s2 are two different computed results, and s1 is
3266* being assumed to be correct, we require
3267*
3268* abs( s1 - s2 ) <= ERRBND = ERRFACT * abs( s1 ),
3269*
3270* where ERRFACT is computed as two times the machine precision.
3271*
3272* Arguments
3273* =========
3274*
3275* ERRBND (global output) DOUBLE PRECISION
3276* On exit, ERRBND specifies the scaled relative acceptable er-
3277* ror bound.
3278*
3279* PSCLR (global input) DOUBLE PRECISION
3280* On entry, PSCLR specifies the scale factor.
3281*
3282* X (global input/global output) DOUBLE PRECISION
3283* On entry, X specifies the scalar to be scaled. On exit, X is
3284* the scaled entry.
3285*
3286* PREC (global input) DOUBLE PRECISION
3287* On entry, PREC specifies the machine precision.
3288*
3289* -- Written on April 1, 1998 by
3290* Antoine Petitet, University of Tennessee, Knoxville 37996, USA.
3291*
3292* =====================================================================
3293*
3294* .. Parameters ..
3295 DOUBLE PRECISION TWO
3296 parameter( two = 2.0d+0 )
3297* ..
3298* .. Intrinsic Functions ..
3299 INTRINSIC abs
3300* ..
3301* .. Executable Statements ..
3302*
3303 x = psclr * x
3304*
3305 errbnd = ( two * prec ) * abs( x )
3306*
3307 RETURN
3308*
3309* End of PDERRSCAL
3310*
Here is the caller graph for this function: