|Message: Re: What's the best choice for number of processes and threads?||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)|
On Fri, 13 Jan 2017 23:54:38 GMT, Andrea Dotti wrote:
> if your machine has 2 cores, it make no sense to go above 2 threads.
Unless each core supports 2 hardware threads -- then it can make sense to configure your compute-bound program to use 4 threads, or even 5.
This is dependent on both the hardware and the details of how the program accesses memory: for programs that have a small working set, the hardware threads don't gain very much. But for a typical Geant4 program with a large working set most of the time, the CPU is often blocked waiting for memory and using both hardware threads nearly doubles throughput.
My program, G4beamline, uses Geant4 with MPI (implemented long before Geant4 supported multithreading); rank 0 does I/O while the rest perform tracking. On my 8-core/2-thread Mac Pro, the best throughput is with 17 ranks. I usually use 15 ranks so the user interface and system services get adequate response.
|Inline Depth:||Outline Depth:||Add message:|