[molpro-user] About very large memory usage in molpro2006
Chen Hui
chenh at yfaat.ch.huji.ac.il
Wed Dec 5 10:38:15 GMT 2007
Hi, molpro users,
I want to know if there is an upper limit on the amount of memory that
can be used by molpro.
I ran a molpro CASPT2 job with 16GB memory and the output file is like:
******
...
ITER. STATE ROOT NORM CORR.ENERGY TOTAL ENERGY ENERGY
CHANGE DEN1 VAR(S) VAR(P) TIME
1 1 1 1.06316402 -0.06316402 -2843.71484973
-0.06316402 -3.10737008 0.63D-01 0.73D+00 32794.57
insufficient memory available - require 1381725450 have
596719322
the request was for real words
ERROR EXIT
CURRENT STACK: CIPRO MAIN
******
So I increased the memory to about 22 GB, but this time another error
appears:
******
...
State Reference Energy
2 -2843.64123985
** ACML error: on entry to DSYEV parameter number 8 had an illegal value
******
I noticed that this time error appears even before the first iteration
of CASPT2. I change the library from ACML to the molpro's own lib, but
the error still exsit like:
******
...
State Reference Energy
2 -2843.64123984
** On entry to DSYEV parameter number 8 had an illegal value
******
I added "olddiag2" to the FTCFLAGS in CONFIG: FTCFLAGS="blas1 blas2
blas3 lapack olddiag2" After doing this, I deleted src/util.o and remade
the program. But the error still exsits.
Another user of molpro communicated with me and he also encountered the
same problem as mine when he used 28 GB memory for molpro. After he
reduced the amount of requested memory to 9GB, then the calculation went
on well.
His opinion is that Molpro can't handle the 28GB memory. So I want to
know is that true that molpro can not handle memory of more than 20GB or
is the error appeared in my calculation just from a bug of molpro?
Chen
--
Dr. Chen Hui
Department of Organic Chemistry
Hebrew University of Jerusalem
The Hebrew University, Givat Ram Campus
Jerusalem, 91904,Israel
Email: chenh at yfaat.ch.huji.ac.il
phone: 02 6585069
fax: 02 6584033
More information about the Molpro-user
mailing list