[Trilinos-Users] [EXTERNAL] Anasazi solver with ML preconditioner won't converge

[Thornquist, Heidi K]

Hi Andris,

The two examples I pointed you to should be looking for the smallest
eigenvalues (--sort="SM"), but the default for some reason is the largest
eigenvalues (--sort="LM").  I will be changing those defaults today, but
if you add the --sort="SM" to the command line the examples will converge.
 I don't think there is anything wrong with the installation.

Thanks,
Heidi

-- 

Heidi K. Thornquist
Electrical Models & Simulation
Sandia National Laboratories
Albuquerque, NM  87185-1323





On 2/7/17, 8:58 AM, "Andris Freimanis" <andris.freimanis_1 at edu.rtu.lv>
wrote:

>Heidi and Alicia,
>
>yes the M was singular in BCSST[K/M]01.
>
>I tried LOBPCG with both Ifpack and ML preconditioners for BCSST[K/M]08
>matrices, but solver didn't converge in both cases. I attached output
>from both solves and source files.
>
>Additionally copy+pasted, compiled and ran
>"LOBPCGEpetraExGenPrecIfpack.cpp" and
>"BlockDavidsonEpetraExGenPrecIfpack.cpp". They didn't converge for
>problems built in those files, could this indicate some kind of a
>problem with my trilinos installation?
>
>
>On 2017.02.07. 03:19, Thornquist, Heidi K wrote:
>> Good point Alicia, I didn't pay enough attention to bcsstm01, that can
>> easily cause problems!
>>
>> Andris, if M is indeed non-singular (it seems to be a diagonal in the
>>spy
>> plot), would you mind trying this again with a generalized eigenproblem
>> that has a positive definite mass matrix?  For example,
>>bcsst[k/m][6,8,9]?
>>   However, it may be easier to modify an example like
>> 
>>anasazi/epetra/example/BlockDavidson/BlockDavidsonEpetraExGenPrecIfpack.c
>>pp
>>   or anasazi/epetra/example/LOBPCG/LOBPCGEpetraExGenPrecIfpack.cpp.
>>These
>> examples use a simple incomplete Cholesky preconditioner and will
>>converge
>> to the solution with their current settings if the eigensolver sorts for
>> the smallest magnitude (--sort="SM").  Changing just the preconditioner,
>> from Ifpack to ML, will give you an easy problem to try it out on.
>>
>> Thanks,
>> Heidi
>>
>
>-- 
>Andris Freimanis
>PhD student
>
>Riga Technical university
>Institute of transporation
>engineering
>
>??psalas iela 6a, R?ga
>00371 263 823 52
>andris.freimanis_1 at edu.rtu.lv
>www.rtu.lv
>