lgermann
Hey there,
I have a weird non-error message / exception caught appearing while I'm trying to refine multiple phases for some in-situ datasets: If I include more than 2 structural phases, I get an exception caught. If I remove one of the phases (without changing anything else) the refinement works. Funnily, the exact same input file works for some data set (3/6) measured under the exact same conditions (same beamline, same integration, format etc).
I restarted my computer & TOPAS (V5) several times. A colleague of mine tried my input file with his V6 version and it worked, however if he switches to V5 he gets the same error message.
Has anybody experienced similar issues & has an idea how to solve it?
Best,
Luzia
rowlesmr
Hi Luzia
Is it crashing with two phases, and any one phase that you remove will let it work?
Have you removed every non-important thing from the input file? (ie no reporting macros, no output macros...)
Can you post your input file?
Matthew
lgermann
Hey Matthew,
Yes, the input files work if I remove one of the phases. Surprisingly the exactly same input file (with having all 5 phases present) works if I use TOPAS 6 instead of 5. That doesn't make sense to me ?
I can send you the input file as PM.
Luzia
rowlesmr
Thanks.
It works for me with v6. Ill try v5 and v4 tomorrow.
Matthew
rowlesmr
Hi Luzia
It crashes 5, runs in 4 and 6.
There is a warning in 5 and 6 (this is for all atom types):
Invalid f1 and f11 for C in file C:\ta7b\ssf\c.nff
for the wavelength 0.20736.
Setting value(s) to zero
There is no warning in 4. Maybe the warning wasn't looked for in 4, not properly handled in 5, and fixed for 6?
.
This is weird. I can't see anything obviously wrong.
Can you try changing the order of the strs in your input file? Maybe there is some strange interaction going on.
Matthew
alancoelho
Hi Luzia
I just got a chance to look over your INP file and the error you are getting is due to the program running out of memory. Version 5 is 32 bit and before it crashed I saw 1.6Gbytes of memory used (2 Gbytes is the maximum).
The step calculation size seems way too small and you have very broad peaks; ie. you have:
convolution_step 3
x_calculation_step 0.0001
the step size in the data is 0.003407; I would use something like:
convolution_step 1
x_calculation_step 0.003407
Increase convolution_step to 2 if anything but I doubt its required.
cheers
alan