[Wien] Forward of moderated message

[Peter Blaha]

Could it be that the IBM (or the queuing system !) has some memory 
limitations like 2 or 4 GB ?
Your case seems pretty large (matrix size ??), no inversion,.... and
and error "at the beginning" of lapw1 could be due to this limitation.

Have you ever tried the mpi-fine grain version. It would probably pay 
off for such a case.

Response to the IBM expert: What means "code problem", when the same 
case runs on another machine and "traceback,.." does not give anything.

> Betreff:
> segmentation fault on IBM p590
> Von:
> "Ludwig, Christian" <ludwigc at uni-mainz.de>
> Datum:
> Fri, 12 Sep 2008 09:54:33 +0200
> An:
> "wien at zeus.theochem.tuwien.ac.at" <wien at zeus.theochem.tuwien.ac.at>
> 
> An:
> "wien at zeus.theochem.tuwien.ac.at" <wien at zeus.theochem.tuwien.ac.at>
> 
> 
> Hello,
> 
> I am running Wien2k Version 08.2 (Release 21.4.2008) on an IBM pSeries 590 machine with SLES 10 with 16 GB RAM.
> Libraries: essl 4.3, lapack
> compiler: xlf90, gcc -mpowerpc64 -m64
> 
> options in the OPTIONS file:
> 
> current:FOPT:-O4 -qarch=pwr5 -qessl -qstrict -qthreaded -q64 -qsmp=omp:opt
> current:FPOPT:-O
> current:LDFLAGS:$(FOPT) -q64 -lmass_64 -lmassvp4_64 -L/usr/lib64
> current:DPARALLEL:'-DParallel'
> current:R_LIBS:-lxlopt -L/usr/lib64 -lesslsmp6464 -L/usr/local/lib64 -llapack
> current:RP_LIBS:
> 
> case file attached
> 
> Parameters: 1 k-point, changed TETRA to GAUSS 0.06 in case.in2, also changed MAX L IN WF to 3.
> 
> With Rk_max 4.5 I can run a calculation, with 4.7, after the execution of lapw0, I get a segmentation fault at the beginning of lapw1. The same case on another computer does not produce the error. The compiler options -traceback and -checkbounds did not yield any output during runtime. The IT expert at IBM thinks that this is not a compilation problem, but a code problem.
> 
> Best regards,
> Christian

-- 
-----------------------------------------
Peter Blaha
Inst. Materials Chemistry, TU Vienna
Getreidemarkt 9, A-1060 Vienna, Austria
Tel: +43-1-5880115671
Fax: +43-1-5880115698
email: pblaha at theochem.tuwien.ac.at
-----------------------------------------