Hello,
First of all, thanks al lot for your reply. Our in-laboratory technician
looked at the 'Out of memory' issue and detected apparently a conflict in
perl versions installed on my computer which may induce this 'Out of Memory'
message. I uninstalled the appropriate version and now it works again (so no
automatic crashes anymore at the start up). I have 4 GB of RAM space on my
computer so it seemed surprising to me, being a layman in computer science,
that all of the initially estimated free space would be consumed.
However, I encounter another crash of the software when I load an artemis
project file (1 specific, other projects function perfectly). The log file
is included in attachment. Could someone please have a look and give any
suggestion about what goes wrong and what can be done to resolve this
problem?
Thanks in advance!
Matthias
_
Laboratory for Chemical Technology
Department of Chemical Engineering, Ghent University
Krijgslaan 281 S5, B-9000 Ghent, Belgium
Disclaimer: http://helpdesk.UGent.be/e-maildisclaimer.php
-Original Message-
From: ifeffit-boun...@millenia.cars.aps.anl.gov
[mailto:ifeffit-boun...@millenia.cars.aps.anl.gov] On Behalf Of Ravel, Bruce
Sent: woensdag 16 januari 2013 23:26
To: XAFS Analysis using Ifeffit
Subject: Re: [Ifeffit] (D)Artemis and (D)Athena crash - perl not responding
Matthias,
The Artemis log file that you sent opens and runs its fit to completion in a
reasonable way on the Windows 7 machine here at my beamline.
As you no doubt saw, the message in the log fle is Out of memory.
I suspect that this gives a hint about the problem. One of the shortcomings
of the current software is that it is based on Ifeffit, a library of
exafs-related routines which is written in Fortran and which is related to
code that Matt wrote way back (in some previous century) when he and I were
in grad school. One of the problems with the dialect of Fortran that
Ifeffit is written in is that it does not have dynamic memory allocation.
This means that Ifeffit is not able to request memory from the operating
system as needed. Instead, it requests a rather large block of memory at
start-up and holds it in reserve in case it is needed. Eventually, it will
run out of memory. If you search the mailing list, you will find many
example of problems related to people attempting to load many dozens of data
sets into Athena and eventually exhausting Ifeffit's memory store.
If your computer has rather little on-board memory, or if you are running
lots of other memory-intensive program, or if you are running several
instances of Ifeffit, you may exhaust your computer's capacity. Because of
the out of memory message in the log file, I suspect that is what is
happening to you.
The very long time that it took Artemis to give up the ghost and die,
suggests that large amounts of memory were being swapped to disk as Ifeffit
was trying to start and then trying to fail because memory had been
exhausted.
Do you think you are in one of those situations? Do you observe this
behavior on another computer? (I suspect not, because your project file
runs just fine on my computer here at my beamline.) Do you observe this
behavior if you manually halt all the other things running on your computer?
Or right after rebooting (when, presumably, lots of other stuff isn't
running)?
B
PS: Among the many advantages to Matt's Larch project, which will eventually
replace Ifeffit in my software and elsewhere, is that it is written in a way
that does dynamic memory allocation.
From: ifeffit-boun...@millenia.cars.aps.anl.gov
[ifeffit-boun...@millenia.cars.aps.anl.gov] on behalf of Matthias Filez
[matthias.fi...@ugent.be]
Sent: Wednesday, January 16, 2013 5:32 AM
To: 'XAFS Analysis using Ifeffit'
Subject: Re: [Ifeffit] (D)Artemis and (D)Athena crash - perl not responding
Hello,
I looked inside the dartemis.log file before, but nothing was inside. Sorry
for not mentioning that! Now, I waited for 20 minutes during the startup of
(D)Artemis and the software seemed to work. I opened a project file (see
attachment), but then Artemis gets stuck again. This time, a bug report
could be obtained (see attachment).
Matthias
_
Laboratory for Chemical Technology
Department of Chemical Engineering, Ghent University Krijgslaan 281 S5,
B-9000 Ghent, Belgium
Disclaimer: http://helpdesk.UGent.be/e-maildisclaimer.php
-Original Message-
From: ifeffit-boun...@millenia.cars.aps.anl.gov
[mailto:ifeffit-boun...@millenia.cars.aps.anl.gov] On Behalf Of Ravel, Bruce
Sent: dinsdag 15 januari 2013 18:03
To: XAFS Analysis using Ifeffit
Subject: Re: [Ifeffit] (D)Artemis and (D)Athena crash - perl not responding
As I have said many, many times before, there's not much I can do without a
proper bug report.
http://bruceravel.github.com/demeter/pods/bugs.pod.html#Capturing_error_mess
ages_from_Demete
B
