[FLASH-USERS] A solution to the "MAXBLOCKS" problem

[Vincenzo Antonuccio]

As some other FLASH 2.x users, I have also recently experienced the error in 
amr_redist arising from the unability to move blocks among different tasks. 
The suggested solution is often that of increasing the number of processors, 
or the MAXBLOCKS variable at setup. However, the jobs I am running are using 
2D runs, with 24 processors and the average number of blocks per processor 
(task) never exceed 1000 blocks (in fact, its is much lesser). Yet, I was 
systematically running into this deadlock.

The way out I have found has to do with the compiler options, not with FLASH 
(iI am using v. 2.5 for this run). I am running on a cluster of Opteron's, 
using MPICH2/mpichf90, which call ifort v. 9.1. When I degraded from an 
aggressive optimization (-O3) to -O2, the problem apparently disappeared. I 
guess that the reason is that, in order to increase performance, the compiler 
makes some tradeoff with memory occupation, possibly forcing lesser freedom 
to the code in allocating memory. In any case, now FLASH does not run anymore 
into error because of that.


******************************************
     Vincenzo ANTONUCCIO-DELOGU

Astrophysics, University of Oxford
Denys Wilkinson Building
Keble Road, Oxford OX1 3RH
United Kingdom

Room 555A, Beecroft Inst. for Particle Astrophysics
Tlf.: +44-(0)1865 283019
Fax:  +44-(0)1865-273390

e-mail: van at astro.ox.ac.uk

'Malheur a l'homme d'etude qui n'est d'aucune 
coterie, on lui reprochera jusqu'a de petits 
succes fort incertains, et la haute vertu 
triomphera en le volant.'

Guai all'intellettuale che non appartiene a 
nessuna consorteria: gli sara' rimproverato
ogni successo, anche il piu' incerto, e la 
virtu' trionfera' derubandolo.

(Stendhal, Le Rouge et le Noir)