thanks for the report, we noticed the forst mismatched you mentionned a
few times ago and still need to push this in a new release. We'll have a
look at the second bug.
On Wed, 16 Jul 2008, John Mulcahy wrote:
The comments in the xHSEQR routines about ispec values for Ilaenv are wrong,
refer to 1, 2, 3... when they should be 12, 13,
The code is also wrong in CHSEQR, SHSEQR and ZHSEQR where it tests for the
matrix size to cross over from xLAHQR to xLAQR0,
*??????? ==== ZLAHQR/ZLAQR0 crossover point ====
???????? NMIN = ILAENV( 1, 'ZHSEQR', JOB( : 1 ) // COMPZ( : 1 ), N, ILO,
???? $????????? IHI, LWORK )
should be calling with an ispec of 12 there, DHSEQR uses the correct value.