LAPACK Archives

[Lapack] bug in c,d,s,z)drvrfp.f?

Dear Julie,
It seems like there is an error in the following files using for the 
packed format (in the case when N=0):
  /TESTING/LIN/(c,d,s,z)drvrfp.f
For example, if we run schkrfp.f, the test fails in sdrvrfp.f (line 451) 
by FPE:

(dbx) where
=>[1] sdrvrfp(nout = 6, nn = 9, nval = ARRAY, nns = 3, nsval = ARRAY, 
nnt = 9, ntval = ARRAY, thresh = 30.0, a = ARRAY, asav = ARRAY, afac = 
ARRAY, ainv = ARRAY, b = ARRAY, bsav = ARRAY, xact = ARRAY, x = ARRAY, 
arf = ARRAY, arfinv = ARRAY, s_work_slatms = ARRAY, s_work_spot01 = 
ARRAY, s_temp_spot02 = ARRAY, s_temp_spot03 = ARRAY, s_work_slansy = 
ARRAY, s_work_spot02 = ARRAY, s_work_spot03 = ARRAY), line 451 in 
"sdrvrfp.f"
   [2] MAIN(), line 243 in "schkrfp.f"

Line 451 (sdrvrfp.f):
                         RCONDC = ( ONE / ANORM ) / AINVNM
ANORM = 0.0, AINVNM = 0.0    --> 0/0 --> FPE

Calculations provided for N=0.
See the lines (344-346) of sdrvrfp.f:

344*              If N.EQ.0, only consider the first type
345*
346               IF( N.EQ.0 .AND. IIT.*GT*.1 ) GO TO 120
In my opinion, there has to be:
346               IF( N.EQ.0 .AND. IIT.*GE*.1 ) GO TO 120

The same changes - for (c,d,z)drvrfp.f

Thank you,
Elena Ivanova.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
http://lists.eecs.utk.edu/mailman/private/lapack/attachments/20130619/aeb3863b/attachment.html
 

<Prev in Thread] Current Thread [Next in Thread>


For additional information you may use the LAPACK/ScaLAPACK Forum.
Or one of the mailing lists, or