[no subject]

[Cristina Domènech <cdomenech@xxxxxxxxxxx>]

Dear David Parkhurst,
Taking into account the advice you told me in the last mail, I have been checking the memories of several computers of the office to see when and where my "complicated" input files worked properly.
I have tested the inputs in 7 computers for the moment. In four of them, whose RAM memories were from 120 MB to 384 MB, the phast code begun the calculations after the
              

"Successfully 
processed database file

Successfully processed chemistry data file"

Then, they went more or less fast depending on their capabilities,
but the program run.
In the other three computers, whose RAM memories range between 256MB and
1GB (this is the RAM memory of the computer for calculations), the
program failed after the previous message and said

"ERROR: NULL pointer returned from malloc or realloc

ERROR: Program terminating

Stopping

Forrtl: severe (157) Program exception - access violation"

The methodology used to test the program was the same for each computer. We have some common directories were the code to install and the input files were placed.
I installed the phast code in the computer, I copied the input files from the common directory to the computer and finally I run the program.
The only difference we have observed between the computers is that the four computers where the program run had WINDOWS 98 while the three others had WINDOWS XP.
It seems clear that given the variability of RAM tested it can not be a problem of RAM memory. It seems a problem related to WINDOWS XP.
Do you think that this is possible?
Any way, in this moment this situation is quite dangerous for us, because we are changing the computers for new ones and these are coming with WINDOWS XP, but at the same time we are interested in using PHAST. And it will be very interesting to know if there is a solution or not.
As I told you in the last mail, we have no problems of this kind if we simplify the model. But the question is why a more complicated model works in one computer and not in another.
If you have a new idea to test or you need more information, please told me.
Thank you very much
Cristina Domènech

--------------------------------- ATENCIÓ, nova adreça de correu i lloc d'Internet
ATENCIÓN, nueva dirección de correo y sitio de Internet
ATTENTION, new e-mail address and Internet site
---------------------------------
Dr. Cristina Domenech
ENVIROS Spain S.L.
Passeig de Rubí, 29-31
08197-Valldoreix
Spain
Tel +34 93 583 05 00
Fax +34 93 589 00 91
mailto: cdomenech@xxxxxxxxxxx
www.enviros.biz
---------------------------------