Message: Diminishing MT returns for photons in voxelized geometries? Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

Question Diminishing MT returns for photons in voxelized geometries? 

Forum: Multithreading
Date: 08 Sep, 2016
From: Youming Yang <Youming Yang>

Hello Geant4 Community,

I am not sure whether the issue I am seeing is due to MT, physics, or geometry, so I will try here first.

My simulation setup is a simple low-energy (~1MeV) photon beam incident on a voxelized geometry (~200x200x200 voxels, 1-2mm dx/dy/dz, implemented as simple 3 layer nesting of PVReplica of G4_Water).

Though my simulation is multi-threaded to 64 cores, I often only see 15-20 core net utilization (all cores are running, but not at 100% load).

If I change the simulation to geantinos (rather than gamma), I see ~100% load on all cores, suggesting it isn't due the geometry or source implementation, but more likely the physics. I have tried StandardEM option3, as well as the livermore EM model.

Does anyone have experience in maximizing the CPU utilization in such scenarios?

Thank you in advance, Ming

Inline Depth:
 1 1
 All All
Outline Depth:
 1 1
 2 2
 All All
Add message: (add)

1 Feedback: Re: Diminishing MT returns for photons in voxelized geometries?   (Andrea Dotti - 13 Dec, 2016)
 Add Message Add Message
to: "Diminishing MT returns for photons in voxelized geometries?"

 Subscribe Subscribe

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