Hi John, Ok !
In documentation, the EGI option mention : *openMOLEMemory: the memory of attributed to the OpenMOLE runtime on the execution node, if you run external tasks you can reduce the memory for the OpenMOLE runtime to 256MB in order to have more memory for you program on the execution node, for instance openMOLEMemory = 256* Perhaps we need just to precise/warn users here that *Java/Scala program run into the same JVM than openMOLE *to reduce the ambiguity; Best R. SR 2016-11-14 10:22 GMT+01:00 Jonathan Passerat-Palmbach < [email protected]>: > Hi Seb, > > memoryByCore/Node would clash with the memory option available on the > other environments for any type of jobs (not the openmole memory). > > https://next.openmole.org/Documentation_Language_Environment > s_Clusters.html > > Problem is at the moment, as Romain was saying, we can't specify the > memory or number of cores requirements on all the EGI VOs through Dirac. > > When this is possible, there will normally be a memory and coreByNodes > options on the EGIEnvironment as well. > > Cheers > J. > > > On 13 November 2016 20:06:32 GMT+00:00, reyman <[email protected]> wrote: >> >> Thanks for this information, this is very important for Gama users to >> know this limitation for large simulation ;) >> >> Why not "memoryByCore" or "memoryByNode" ? >> >> >> >> 2016-11-13 19:01 GMT+01:00 Romain Reuillon <[email protected]>: >> >>> What would you propose for a more explicit name? >>> >>> EGI is 2G by core. If you set openmole to 2G of heap size it will take >>> arrond 2.5G of RAM for stacks and other part of the jvm on the node. I you >>> are desparte for RAM you can go up to 3G, but you might ear some complaints >>> from the grid admins :) >>> >>> Romain >>> >>> >>> Le 13/11/2016 à 18:36, sebastien rey coyrehourcq a écrit : >>> >>> Ok ! The name "openMOLEMemory" is a little ambiguous. >>> For information, the maximum amount of Ram on node on EGI is limited to >>> 2048Mo or more ? >>> >>> PS : thanks for git lfs tips, it works well now :) >>> >>> --- >>> Sébastien Rey Coyrehourcq >>> Research Engineer, BAP E, UMR IDEES >>> Sent from Nylas N1 <https://nylas.com/n1?ref=n1>, the extensible, open >>> source mail client. >>> >>> On nov. 13 2016, at 6:05 pm, Romain Reuillon <[email protected]> >>> <[email protected]> wrote: >>> >>>> Don't reduce the memory used by openmole. The memory used by openmole >>>> is the same as the one use by gama (that's the memory allocated to the >>>> JVM). By default it is set to 1024 megs, make it 2048. >>>> >>>> Romain >>>> >>>> Le 13/11/2016 à 17:02, reyman a écrit : >>>> >>>> Hi guys, >>>> >>>> It seems the gama model i try to run on grid take lot of memory ... >>>> because i have multiple *java.lang.OutOfMemoryError: Java heap space >>>> error with some parameters.* >>>> >>>> I already reduce the openmole memory with the command >>>> >>>> val env = EGIEnvironment("vo.complex-systems.eu",openMOLEMemory = 256) >>>> >>>> * Is there also a way to change the memory used by gama on jvm on node >>>> ?* >>>> >>>> Thanks, >>>> SR. >>>> >>>> >>>> _______________________________________________ >>>> OpenMOLE-users mailing >>>> [email protected]http://fedex.iscpif.fr/mailman/listinfo/openmole-users >>>> >>>> >>>> >>> >>> _______________________________________________ >>> OpenMOLE-users mailing >>> [email protected]http://fedex.iscpif.fr/mailman/listinfo/openmole-users >>> >>> >>> >>> _______________________________________________ >>> OpenMOLE-users mailing list >>> [email protected] >>> http://fedex.iscpif.fr/mailman/listinfo/openmole-users >>> >>> >> >> >> -- >> <http://stackoverflow.com/users/385881/reyman64> >> >> ------------------------------ >> >> OpenMOLE-users mailing list >> [email protected] >> http://fedex.iscpif.fr/mailman/listinfo/openmole-users >> >> > -- > Jonathan Passerat-Palmbach, PhD > Research Associate > Department of Computing > Imperial College London > > South Kensington Campus > Huxley Building - room 344 > 180 Queen's Gate > London SW7 2AZ > -- <http://stackoverflow.com/users/385881/reyman64>
_______________________________________________ OpenMOLE-users mailing list [email protected] http://fedex.iscpif.fr/mailman/listinfo/openmole-users
