Message: Re: Optical photon transport and cpu time Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

Feedback Re: Optical photon transport and cpu time 

Forum: Processes Involving Optical Photons
Re: Question Optical photon transport and cpu time (Frederic Yermia)
Date: 04 Feb, 2009
From: Gumplinger Peter <Gumplinger Peter>

Hello Frederic,

I didn't see anything wrong with your code; nothing that would cause an inflation of CPU time. I don't know how much energy you hope to detect in your scintillator but from:


I see that you request 7000 scintillation photons for every MeV dumped into your scintillator. That's a lot of tracks to simulate. You may find that you will get a statistical representative sample by biasing your yield (down) if you just simulated a fraction of them (you'll weight your result accordingly).

Now, there is one more, probably significant, CPU gain you can achieve by defining your interface between the scintillator and your ground reflector as:


This option allows you to define your thin air gap not as a distinct volume in the G4 geometry, but as something 'known' to the G4OpBoundaryProcess internally. Please, see the documentation for details (Example 5.8). It looks to me that your surface specifications are perfectly suited for this build in speed-up:

OpAirHousingSurface->SetFinish(ground); OpAirHousingSurface ->SetPolish(0.);

This is the same as Lambertian reflection, which is what the back reflector is simulated as in 'groundbackpainted'.

Regards, Peter

 Add Message Add Message
to: "Re: Optical photon transport and cpu time"

 Subscribe Subscribe

This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews