[Gate-users] 32 bit vs 64 bit

Tianyu Ma maty at tsinghua.edu.cn
Fri Aug 25 11:05:29 CEST 2006


	For reference, I list a simple benchmark as follows:

	#System 1: xeon 2.8GHz CPU, 2GB mem, 32bit linux, Intel C++ complier 32bit
	#System 2: Itanium 3.0GHz CPU, 2GB mem, 64bit linux, Intel C++ complier 64bit

	A Focus120 MicroPET system with a slit hole tungsten collimator was described. The 
source was a cylindrical uniform 140keV single gamma source. 555M photon histories were 
simulated. On eath system, only one single GATE process was running although both system
have dual CPUs. In the 32bit system, the real simulation time was 74059s and in the 
64 bit system it was 69606s. 
	One can see that 74059s * 2.8GHz ~= 69606s* 3.0GHz. So it seems that the 64-bit
system didn't make difference.

From: komarovs
Date: 2006-08-24 22:22:39
To: gate-users at lists.healthgrid.org
CC To: 
Subject: [Gate-users] 32 bit vs 64 bit

Dear Gate users

I have a few questions:

1. Does GATE and/or Geant 4 runs faster in a 64-bit system? or there is no
significant difference
between a 32 bit vs a 64 bit system?

2. Was GATE and/or Geant 4 written to take advantage of a dual-core CPU?
Alternatively, what's the most efficient way to run GATE on a dual core
One simulation at a time? or 2 or more?

And one more question:

The initial random seed number (the first int number in .rndm  file) is the
same in the files: beginOfRun.rndm and endOfRun.rndm. Is it supposed to be
so or it's something wrong with my system?

Thank you.
Gate-users mailing list
Gate-users at lists.healthgrid.org

More information about the Gate-users mailing list