|Message: Re: Parallel the Dicom...how to make it more easily and more efficiently||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)|
I am not familiar with parallelizing Geant4, however regarding MPI and efficiency in general...
Depending on how you did your MPI implementation, my guess is that you're attempting to run multiple identical processes on one computer, but each requiring the GB's of ram that a voxelized geometry would need.
The benefit of MPI would be best seen in a cluster of nodes that each have their own cpu/ram resources, where you can run multiple instances of the same simulation, just starting from different RN seeds for your incident particles.
I am not familiar with topC. But if you are attempting to obtain speedup through multi-core usage on one standalone system, have you tried looking into openMP? The basics of that framework is well suited to leveraging multiple cores on the same system, with access to the same global memory.
On Fri, 29 Jun 2012 14:33:34 GMT, Geng wrote:
> Dear experts: > > I am studying for my master degree with the simulation of Geant4. My > application is perfectly like the example "/extend/medical/Dicom". I > have meeting a very serious problem with it------time consuming. I do > not know how do you deal with it. If I want to get a relative precise > result,it may consume few days or even weeks. That may be very terrible. > So have you think about the parallel of "DICOM"? I have tried to using > the MPI, but it even slower than one core. I really cannot understand > what is wrong with the application. In addition I also tried to using > the Top-c,But another problem occurred. How to port the "Dicom"to > available to top-c? I have read the example"parN02"and "parN04",the > similar of them is that they both used the trackerhit class,and also > marshaled it. But in "Dicom",multifunctionDetector was used to collect > the information we interest. So I do not know how to deal that. > > Could anyone give me some advise,that will be great appreciate. > > Thanks very much. Best regards. Geng >