[Gate-users] inhomogeneous activity distribution in several outputfiles

Simon Stute gate.stute at gmail.com
Mon Jan 3 22:42:58 CET 2011


OK, at least it is good that your problem is solved !
And thanks for pointing out the Ranlux problem.
Which version of GATE are you using, and linking with which version of clhep
But anyway, the MersenneTwister is probably the best random number generator
(in efficiency and in accuracy), even if it is not obvious to notice that in
daily's work ! So if your simulations run safe using the JamesRandom one,
just keep it :-)


On Mon, Jan 3, 2011 at 4:56 PM, Hans-Martin Schwab <
hans-martin.schwab at web.de> wrote:

> Hello,
> thanks a lot, I tested that but it didn´t have any effect on my results.
> Moreover I did not use a back-to-back source.. Well I made some attempts
> using the JamesRandom-generator and now everything seems to work properly. I
> don´t know why, but this appears to have been a problem with the Ranlux64..
> Best regards,
> Hans
>  ------------------------------
> *Von:* "Simon Stute" <gate.stute at gmail.com>
> *Gesendet:* 28.12.2010 19:40:50
> *An:* "Hans-Martin Schwab" <hans-martin.schwab at web.de>
> *Betreff:* Re: [Gate-users] inhomogeneous activity distribution in several
> outputfiles
> Hey,
> Didi you use backtoback sources ?
> Which version of Gate do you use ?
> It could be a problem with a non-initialized value in the code, concerning
> the non-collinearity flag when using backtoback sources.
> I encountered such a bug before, and it seemed to be fixed once I forced
> acolinearity to be true or false (whatever you want, it is just that this
> flag has no default value in the source code, and thus if the flag "decided"
> to be true, a non-initialized acolinearity angle is also used ... thus
> giving funny thing ...).
> Hope this helps,
> Simon
> On Thu, Dec 23, 2010 at 3:24 PM, Hans-Martin Schwab <
> hans-martin.schwab at web.de> wrote:
>>   Hello everybody,
>>   for my recent simulations I wrote a script in Root that creates a
>>   threedimensional .raw-file storing the activity distribution as derived
>>   from the sourcePos information in the coincidence dataset. Observing
>>   that raw-file one can see irregularities in the source-distribution
>>   that can even be recognized in the reconstructed image of the
>>   simulation data. There are several artifacts, that become visible. For
>>   some phantoms a area of a higher number of counts is noticable in the
>>   x-z-plane, although a homogeneous distribution is expected. Other
>>   images show hot planes orientated diagonally across each subvolume. Of
>>   course the data stored in the sourcePos dataset does not yet contain
>>   corrections of sensitivity and scatter, but the shape of these areas
>>   with more counts can not be explained by those issues. What
>>   is surprising, is that when I run the same script for several times
>>   (where only the output name distiguishes) these irregularaties
>>   sometimes occur and sometimes not. Does anybody have an idea about
>>   that? We suppose the random generator to play a role in this (in the
>>   macros the Ranlux64 is used), but still we canA't really find a
>>   consistent idea of what might cause those artifacts.
>>   Best regards,
>>   Hans
>>   Neu: WEB.DE De-Mail - Einfach wie E-Mail, sicher wie ein Brief!
>>   Jetzt De-Mail-Adresse reservieren:
>>   [1]https://produkte.web.de/go/demail02
>> References
>>   1. https://produkte.web.de/go/demail02
>> _______________________________________________
>> Gate-users mailing list
>> Gate-users at lists.opengatecollaboration.org
>> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
> NEU: FreePhone - kostenlos mobil telefonieren und surfen!
> Jetzt informieren: *http://produkte.web.de/go/webdefreephone*<http://produkte.web.de/go/webdefreephone>

More information about the Gate-users mailing list