LAPACK Archives

[Lapack] syev complexity


Actually maybe 35/3 n^3 for SYEV with an 'V' flag ... 

On Aug 9, 2011, at 11:45 AM, Langou, Julien wrote:

Hello, 

O(n^3) for both. The computation of the constant is somewhat more 
problematic. (Since there is a convergence story behind that, so you need a 
tolerance, hope for a cubic convergence, etc.)

Golub and Van Loan (3rd Edition, p.421) textbook declares that this is 4/3 
n^3 for STEV and 9n^3 for SYEV. 

(I am not sure how to read GVL on this one actually, maybe they mean 31/3 n^3 
for SYEV. I send an email to Charles, we'll see if there is an answer.) 

Anyone on the mailing list wants to help answering? 

Julien.


On Aug 6, 2011, at 4:55 PM, Andrea Arteaga wrote:

Dear all,
I did not find any specification of the complexity of the syev
routine. I'm using it to completely diagonalize a matrix (i.e. with
JOBZ = "V"). I'm also interested on the stev complexity, for the same
task.

Thanks for the help.
Best regards.
Andrea Arteaga
_______________________________________________
Lapack mailing list
Lapack@Domain.Removed
http://lists.eecs.utk.edu/mailman/listinfo/lapack



<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