<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.2180" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=Arial size=2><FONT face="Times New Roman" size=3>Hello
everybody<BR><BR>I am calculating Alk-H series.<BR><BR>The problem we have is
connected to calculation of elastic constants.<BR>Everything went fine with Li,
Na and K cases, but We cannot obtain<BR>meaningful results for CsH and
RbH.<BR><BR>For CsH eos and tetra results look OK, but rhomb energy variations
are<BR>very small, tipically on 4th decimal place, it looks like system is
not<BR>sensible enough on this type of distorsion. Good results for eos
and<BR>tetra cases emerged only after increasing Gmax factor to
20.<BR><BR>Situation is far worse for RbH with large fluctuations for tetra
and<BR>again weak sensibility for rhomb case.<BR><BR>We already did elastic
constants (only C11 and C12) for same systems few<BR>years ago by hand with
WIEN97 and everything went fine (sensible results <BR>comparable with previous
findings).<BR>Perhaps it has something to do with APW+lo basis
set???<BR><BR>Details of calculations:<BR><BR>Complete apw+lo basis
set.<BR><BR>RbH: RMT 2.2 a.u for Rb and 1.6 a.u for H (large interatomic
distance<BR>throughout whole series)<BR>CsH: RMT 2.3 a.u. for Cs and 1.6 a.u.
for H<BR><BR>RmtKmax 8<BR>GGA96, with -7 ryd separation energy<BR>4000 k-points
in whole unit cell for all configurations<BR>Gmax 20<BR><BR>The other question
is regarding compiled binaries on our system. LAPW1<BR>crashes with SIGSEGV
whenever we try to calculate system with more then<BR>few atom per unit cell
(includeing test case of course). This does not<BR>happen if we use supplied
precompiled binaries.<BR><BR>I read article by Gerhard H. Fecher about
compiling<BR>Wien2k on Intel systems where he described similar problem
with<BR>calculation of more complicated systems, due to combination
of<BR>statically linked binaries and some GNU libraries. We have tried
both<BR>statically and dynamically liking with same result. Does anyone
knows<BR>the origin of this problem?<BR><BR>BTW, with latest update of wien2K,
calculations crash in first cycle in <BR>lapw2 with infamous 'error in LAPW2'
message description (same compiler <BR>options as for previous version of
wien).<BR><BR>Operating system is Slackware 10.1 kernel
2.6.18.2<BR><BR>Hardware: dual core 2 E6400 on 2.17 GHz with 1GB
RAM.<BR><BR>Compiler ifc 9.1 + mkl 8.1.<BR>current compiler options:<BR>FOPT
= -FR -mp1 -w -prec_div -pad -ip -DINTEL_VML -O3 -xT -axT -openmp<BR>FPOPT
= -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML<BR>LDFLAGS =
-L/opt/intel/mkl/8.1/lib/32 -i-static -lguide -lguide_stats <BR>-lsvml
-lpthread<BR>R_LIBS = -L/opt/intel/mkl/8.1/lib/32 -lmkl_lapack
-lmkl_ia32<BR>C_LIBS = $(R_LIBS)<BR><BR>Best regards<BR><BR>Nenad
Ivanovic</FONT><BR><BR><BR><BR><BR><BR><BR><BR></FONT></DIV></BODY></HTML>