[Gate-users] Variance Reduction Technique on GATE v6...

[Park MinJae]

Dear Gaters,

I tested the performance of VRTs, one by one as following:
[ I add a VRT command in the middle of benchSPECT, ]

/gate/physics/addProcess PhotoElectric
/gate/physics/addProcess LowEnergyCompton
/gate/physics/addProcess LowEnergyRayleighScattering

/gate/physics/addProcess ElectronIonisation
/gate/physics/addProcess Bremsstrahlung
/gate/physics/addProcess MultipleScattering e-

#/gate/physics/processes/Bremsstrahlung/activateSplitting e- 100
#/gate/physics/processes/Bremsstrahlung/activateRussianRoulette e- 50

#/gate/physics/processes/PhotoElectric/activateSplitting gamma 100
/gate/physics/processes/PhotoElectric/activateRussianRoulette gamma 50

/gate/physics/processList Enabled
/gate/physics/processList Initialized


However, I got the weird results.

[turbo978 at dmip0 VRT_none]
It took 1584 seconds

[turbo978 at dmip0 VRT_split]
It took 3147 seconds

[turbo978 at dmip0 VRT_RR]
It took 1460 seconds

Splitting is 2times slower than no VRT.
RR is a little bit faster.


Moreover, the energy spectra of VRTs is not normal as I attached.
[Single count of spliting is quite low.]


Am I doing right?


Cordially yours,
MJ

//=================================================
//  Ph.D. Park MinJae
//
//  Johns Hopkins University
//  School of Medicine, Dept of Radiology
//  Division of Medical Imaging Physics
//
//  DialTo:+1-410-955-2684 (Office)
//  MailTo:MJ at ToTi.us, MJ at JHMI.edu
//
//  Make the Dream True...
//=================================================



On Wed, Jun 23, 2010 at 11:09 AM, Park MinJae <MJ at toti.us> wrote:
> Dear Thibault,
>
> Yes, It's obviously GATE v6 p01 :)
>
>
> Anyway, let me clarify my question.
>
> What I want to know is...
>
> 1) Is there any other VRTs like forced detection on GATE v6?
>
> 2) Is it possible to use VRTs simultaneously?
>
> When I use following two commands...
> /gate/physics/processes/PhotoElectric/activateSplitting gamma 100
> /gate/physics/processes/PhotoElectric/activateRussianRoulette gamma 50
>
> I got the messange like this.
> src/GateVProcessMessenger.cc (l.126): Splitting or Russian Roulette
> already activate for the process PhotoElectric.
>
> C++ source of GATE is below:
> if(( command==pSetSplit || command==pSetRussianR) &&
> pProcess->GetIsWrapperActive())
>   GateError("Splitting or Russian Roulette already activate for the
> process "<<pProcess->GetG4ProcessName()  <<".");
>
> There is "or" between Split and RussianR.
>
>
> Cordially yours,
> MJ
>
> //=================================================
> //  Ph.D. Park MinJae
> //
> //  Johns Hopkins University
> //  School of Medicine, Dept of Radiology
> //  Division of Medical Imaging Physics
> //
> //  DialTo:+1-410-955-2684 (Office)
> //  MailTo:MJ at ToTi.us, MJ at JHMI.edu
> //
> //  Make the Dream True...
> //=================================================
>
>
>
> On Wed, Jun 23, 2010 at 5:19 AM, Thibault Frisson
> <frisson at creatis.insa-lyon.fr> wrote:
>> Hi,
>>
>>
>>> 1) is there those two implemented VRTs on ROOT v6.0 p01?
>>
>> yes (I think you want to say "GATE v6.0 p01")
>>
>>
>>> I saw there is the other VTR on CT projector, which is not NM.
>>>
>>> I saw Beenhouwer's paper which described two VRTs and enhanced about 5
>>> times.
>>> On the Staelens's paper, they are working on the forced detection.
>>>
>>> I guess...Using filter could be an option to enhance the speed.
>>> but it's not primary VRTs, right?
>>>
>>
>> Not sure to understand your question. Filters are used to improve the
>> efficiency of splitting or russian roulette.
>>
>>
>>> 2) Is it working on the other process except Bremsstrahlung?
>>
>> yes, it should work for all processes.
>>
>>
>> Best regards,
>>
>> Thibault
>>
>> --
>> Thibault Frisson, Phd
>> Post-doctorant
>>
>> Centre de lutte contre le cancer Léon Bérard
>> Service de radiothérapie
>> 28 rue Laënnec
>> 69373 Lyon cedex 08
>> Tel : 04 78 78 51 51
>>      06 70 37 05 21
>>
>> Laboratoire CREATIS (UMR CNRS 5220)
>> INSA - Bâtiment Blaise Pascal
>> 7, avenue Jean Capelle
>> 69621 Villeurbanne cedex
>>
>> http://www.creatis.insa-lyon.fr/rio
>> _______________________________________________
>> Gate-users mailing list
>> Gate-users at lists.opengatecollaboration.org
>> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>>
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: VRT_none.png
Type: image/png
Size: 6613 bytes
Desc: not available
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20100624/2d03e087/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: VRT_Split.png
Type: image/png
Size: 6185 bytes
Desc: not available
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20100624/2d03e087/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: VRT_RR.png
Type: image/png
Size: 5815 bytes
Desc: not available
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20100624/2d03e087/attachment-0005.png>