[Gate-users] Re: benchmark

Fri Jun 1 17:13:44 CEST 2007

Hi all,

I am experiencing the same problems as Suhaili Zakaria, Antigoni Divoli
and Daniel Woodsworth (see below) with the SPECT benchmark. I have run
the benchmark on Gate version 3.0.0 and 3.1.1 and get similar result on
both versions. The biggest discrepancies are these:

Percentage of unscattered photons
My average value: 35.72%  ,   Benchmark value: 32.90%  ,   Out by 9%

Percentage of photons whose last scattered event occured in the
collimator:
My average value: 0.34%  ,   Benchmark value: 2.10%  ,   Out by 84%

Percentage of photons whose last scattered event occured in the
crystal:
My average value: 6.55%  ,   Benchmark value: 8.50%  ,   Out by 23%

But, as Antigoni pointed out, some values agree well. My average number
of detected counts between 20 keV and 190 keV is 36066 and is only out
by 0.4%. 

I'm running Gate on Linux with Suse. 

I am a bit reluctant to trust my output values from my other
simulations before I'm confident that my version of Gate gives me
benchmark values which agrees with those recommended. 

Any suggestions for what we can do to sort this problem out?

Any comments much appreciated.

Kind regards,
Maria

**************************************************************

Message: 3
Date: Wed, 23 May 2007 17:40:34 +1200
From: "Suhaili Zakaria" <suhailiz at gmail.com>
Subject: Re: [Gate-users] benchmark
To: "Daniel Woodsworth" <danieljw at phas.ubc.ca>
Cc: gate-users at lists.healthgrid.org 
Message-ID:
	<f3df92200705222240h5f35f347s53f49b026d6f8d80 at mail.gmail.com>
Content-Type: text/plain; charset="windows-1252"

Hi,
Thanks for your help.

I can run benchmark_spectra and benchmark_projections now.
But the results obtained was different from the results showed in the
manual.
Is it allright, or something wrong somewhere?

Number of counts for the total energy spectrum = 36041
Number of counts for the Unscattered spectrum =12809
Number of counts for the total scatter spectrum =23232
Number of counts for the crystal scatter spectrum =2392
Number of counts for the shie scatter spectrum = 1
Number of counts for the coll scatter spectrum = 107
Number of counts for the comp scatter spectrum = 456
Number of counts for the tabl scatter spectrum = 1081
Number of counts for the movs scatter spectrum = 5123
Number of counts for the phan scatter spectrum = 0
Number of counts for the order1 scatter spectrum = 11176
Number of counts for the order2 scatter spectrum = 6120
Number of counts for the order3 scatter spectrum = 3004
Number of counts for the order4 scatter spectrum = 1500
Number of counts for scatter orders greater than 4 spectrum = 1432

Number of detected counts from 0.020 to 0.190 MeV : 36041
Unscattered events (percentage of total counts)= 35.54
Scatter in the crystal (percentage of total counts) = 6.64
Scatter in the shie volume (percentage of total scatter) = 0.00
Scatter in the coll volume (percentage of total scatter) = 0.30
Scatter in the comp volume (percentage of total scatter) = 1.27
Scatter in the tabl volume (percentage of total scatter) = 3.00
Scatter in the movs volume (percentage of total scatter) = 14.21
Scatter in the phan volume (percentage of total scatter) = 0.00
Scatter order1 (percentage of total scatter) = 48.11
Scatter order2 (percentage of total scatter) = 26.34
Scatter order3 (percentage of total scatter) = 12.93
Scatter order4 (percentage of total scatter) = 6.46
Scatter orders greater than 4 (percentage of total scatter) = 6.16

Appreciate your reply and help.
Many thanks,
S

**************************************

Message: 1
Date: Wed, 23 May 2007 09:25:28 +0100
From: "Antigoni Divoli" <Antigoni.Divoli at icr.ac.uk>
Subject: Re: [Gate-users] benchmark
To: <suhailiz at gmail.com>, <danieljw at phas.ubc.ca>
Cc: gate-users at lists.healthgrid.org 
Message-ID: <s6540892.000 at groupwise.icr.ac.uk>
Content-Type: text/plain; charset=Windows-1253

Dear all,
I would also like to take this opportunity to say that we are
experiencing some disagreement between the optimal results suggested by
Gate developers and the once obtained by myself (on a Mac Intel Duo
platform) and by a colleague on a LINUX platform (don't remember the
exact details). Our results agree with each other, but disagree with the
suggested values. For example:
* Percentage of photons whose last scattered event occured in
collimator: 0.35% (benchmark value is 2.10%). 
* Percentage of photons whose last scattered event occured in crystal:
6.43% (benchmark value is 8.50%). I'm out by -24.38%!
 etc. 
Some values agree well though, for example the number of detected
counts. We tried running the same benchmark on different versions of
gate with similar results.

Should we worry? Is it something that occurs often, but is being
overlooked?

I'll appreciate any comments

Antigoni

***************************************

Message: 4
Date: Wed, 23 May 2007 09:26:39 -0700 (PDT)
From: "Daniel Woodsworth" <danieljw at phas.ubc.ca>
Subject: Re: [Gate-users] benchmark
To: "Antigoni Divoli" <Antigoni.Divoli at icr.ac.uk>
Cc: gate-users at lists.healthgrid.org 
Message-ID: <1179.137.82.91.46.1179937599.squirrel at www.phas.ubc.ca>
Content-Type: text/plain;charset=iso-8859-1

Hi,

I have similar results to Antigoni. I have run the SPECT benchmark 3
times
on my computer and obtained:

Percentage of photons whose last scattered event occured in
collimator:
0.35%,0.36%,0.35%    Average=0.353%

Percentage of photons whose last scattered event occured in crystal:
6.43%,6.62%,6.62%    Average=6.56%

Similarly, all of the rest of my results agree reasonably well with
the
benchmark values. One thing I have noticed: from the file
results_benchmark.pdf (in $GATEHOME/benchmarks/benchmark_SPECT), which
seems to be an example data from a single run of the benchmark, the
data
'detected percentage of counts' is well out of the values given on the
Gate website or in the install guide (ie those expressed as VALUE +/-
SD).
Maybe this shows that the errors given are a bit optimistic? I'm not
sure.
At any rate, I am also curious if this is a problem.

Cheers,

Daniel

******************************************

The Institute of Cancer Research: Royal Cancer Hospital, a charitable Company Limited by Guarantee, Registered in England under Company No. 534147 with its Registered Office at 123 Old Brompton Road, London SW7 3RP.

This e-mail message is confidential and for use by the addressee only.  If the message is received by anyone other than the addressee, please return the message to the sender by replying to it and then delete the message from your computer and network.