files run ok with VASP4.6, error and crash with VASP5.3

[peterklaver]

Hello people,

I came across a case where files for the calculation of a Ti oxide run ok with VASP 4.6 but they cause things to fail with an error with VASP 5.3. First there are countless warnings like

WARNING: Sub-Space-Matrix is not hermitian in DAV 7
-2.357326009921828E+113

then

WARNING: DENTET: can't reach specified precision
Number of Electrons is NELECT = 327.875000000000

and then

the old and the new charge density differ
old charge density: 328.00000 new 327.87499
0.161E+02
Error EDDDAV: Call to ZHEGV failed. Returncode = 9 2 16

Since that particular calculation doesn't require any VASP5 specific features, I can run it fine with 4.6, no problem for me. But it does seem a bit bad that files that run ok in 4.6 crash out with v5.3. Maybe the VASP people would care to see if this is related to a problem in the code? If so, the files (stdout messages in files called 'vasp.out') can be downloaded from

http://x120141.tudelft.net/files.zip


[update: oops, our university did a network overhaul and now that link for downloading has gone dead. Will put in a working one once network upgrade problems have been fixed.]
<span class='smallblacktext'>[ Edited ]</span>

[peterklaver]

Ok, url to the input/output files download for this problem has now been restored to something working again......