Page 1 of 1

Diagonalization

Posted: Fri Mar 23, 2012 4:43 pm
by mailoliver
Dear all:

I try to diagonalize the BS matrix to get the properties of excitons of SiC. It is ok for convergence tests. But when I try to run yambo with large nkpt(20*20*20), it reports:
Diagonalization solverforrtl: severe (179): Cannot allocate array - overflow on array size calculation.

The calculation process as following:
1. yambo -b -o b -y h
2. yambo -o b -y d -V resp

Thanks in advance for your help!

Re: Diagonalization

Posted: Fri Mar 23, 2012 4:48 pm
by andrea marini
How big is the matrix? Maybe it is only too large to be diagonalized. If you need the wavefunctions of low energy excitons maybe few bands are enough. To get absorption remeber to use Haydock iterative approach when matrices are too big.

Re: Diagonalization

Posted: Sat Mar 24, 2012 8:46 am
by mailoliver
Thank you for your reply. It can work normally.