[Beowulf] thanks for the array issue
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
f.costen at cs.man.ac.uk fumie.costen at manchester.ac.ukMon Dec 17 01:47:11 PST 2007
- Previous message: [Beowulf] large array to run
- Next message: [Beowulf] Interesting Webinar: Multi-core in HPC
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Dear Donald(Shillady),James(Cownie), Alan(scheinine),Toon(Moene),Geoff(Jacobs) Thank you very much for your feedback. This array is meant to be used to build an FDTD radio environment at the master ( MPI_WORLD == 0 ). And part of it is planned to be sent to each slave. That's why this particular array is used at the master. Sadly, as you pointed out, the size of the memory per core restricted my calculation although the shell environment setting in .cshrc is done like: limit coredumpsize 0 unlimit stacksize These two lines did not turn the computers infinite ;-) > Mike(Frese) So the solution, which I came up with, is, as Brian(Dobbins) said on 13th, that #define ntmax 820 integer*4, ALLOCATABLE :: a( : , : , :) allocate(a(ntmax,ntmax,localsmalldimension)) to distribute this matrix a(,,) to each slave to fit to the size of the memory available at each slave (and master) Thank you very much Fumie
- Previous message: [Beowulf] large array to run
- Next message: [Beowulf] Interesting Webinar: Multi-core in HPC
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
