|Message: RE: Performance of homogeneous voxelised phantom||Not Logged In (login)|
Click on the Forum title, e.g. on the "Forums by Category" page, to read a sequence of postings to the Forum and its threads all in one page. If you are only interested in one thread or the thread following a specific posting, click the thread or the posting, which takes you to a smaller page, which contains only the part you are interested in and may be easier to navigate.
Messages are "chained" if there are only replies at the first level, i.e. 1/1.html, 1/1/1.html etc. In case of "chained" messages the message number is replaced by the icon and there is no indentation.
Inline: Display the subject line only or also the text of the posting(s); for the choice "All" the "Outline" choices are switched off.
|1||0||1||no text / full text of posting|
|2||1||All||text for level 1 only / text for All postings|
Outline: Choose the depth of the posting thread, successive toggle controls provide increasing detail.
|1||2||1||2 levels / 1 level (original posting)|
|2||3||2||3 levels / 2 levels|
|3||3||All||3 levels / all levels (all postings)|
> After some reading, I realise that my implementation is probably > worst case as performance is concerned, ... I'm afraid so. But not due to the scorer but most-likely due to the way how you implemented your voxels. Could you please post the code where you implement the voxel geometry? In particular, how did you divide the phantom? Makoto > -----Original Message----- > From: Sven Lotze [mailto:firstname.lastname@example.org] > Sent: Thursday, November 08, 2007 3:23 AM > To: publichypernews > Subject: Performance of homogeneous voxelised phantom > > *** Discussion title: Geometry > Email replies to PublicHyperNews@slac.stanford.edu must include: > In-Reply-To: <"/geometry/814"@geant4-hn.slac.stanford.edu> > Subject: ...change this to be about your reply. > > Hi, > > in my simulation I have a homogeneous water phantom voxelised in > 3d by dividing the box geometry into many G4Boxes in the tracking > geometry and attaching a primitive scorer for deposited energy > to each. As I increase the resolution (smaller voxels), calculation > time skyrockets. > > After some reading, I realise that my implementation is probably > worst case as performance is concerned, because the tracking has > to stop at every voxel boundary. The alternatives I see now are > > A) using a parallel readout geometry and do the scoring there, > like e.g. in the hadrontherapy example > > B) doing it manually by using G4UserStepping action, like > in TestEM11/12. > > I would assume that B should be faster, as more low level > G4 functions are used - is that correct? > > On the other hand, as I have to limit the step length as > done in TestEM11/12 according to voxel size, doesn´t that > lead to the same problems as my original geometry with > small volumes limiting the steps? > > Looking at TestEM11/12, for the deposited energy in each > step a position along the step is randomly chosen before > filling it to a histogram. Why is that necessary if max. > step size is limited corresponding to voxel (=bin) size > anyway? Why not deposit it at the center of the step? > > Thanks for your help! > > Sven > >
|Inline Depth:||Outline Depth:||Add message:|