Hi Kun: Here QoS seems to be a real-time computation constraint, right? Kepler currently has no provisions to deal with such RT constraints.
On the other hand, you might find some of the work on the underlying Ptolemy system of interest. Googling Ptolemy + realtime will find you many works by the Ptolemy folks (mainly on design of RT system I suppose): http://www.mail-archive.com/ptolemy-hackers at bennett.eecs.berkeley.edu/msg00355.html http://ptolemy.eecs.berkeley.edu/~johnr/papers/thesis.html Bertram -- Bertram Ludaescher, Assoc. Professor Dept of Computer Science & Genome Center University of California, Davis One Shields Avenue, Davis, CA 95616 Ph: (530) 754-8576 ludaesch at ucdavis.edu >>> On Tue, 03 Oct 2006 23:15:34 -0500 >>> Kun Xiao <kxiao3 at iit.edu> wrote: KX> KX> Hi, KX> How can Kepler deal with the QoS constraints(e.g. a function should be KX> finished in less than 1 second)? KX> KX> Thanks! KX> KX> KX> ---------------- KX> Best Regards, KX> Kun Xiao KX> KX> KX> -----Original Message----- KX> From: Dan Higgins [mailto:higgins at nceas.ucsb.edu] KX> Sent: 2006年10月2日 11:44 KX> To: Kun Xiao KX> Cc: 'Efrat Frank'; Kepler-users at ecoinformatics.org KX> Subject: Re: [kepler-users] Kepler-users at ecoinformatics.org KX> KX> Thanks for discovering this problem! KX> KX> Dan Higgins KX> KX> Kun Xiao wrote: >> I have found the reason of high usage of cpu. >> >> The user guide of kepler missed some details about the addition >> sample, so after I build a model as it said, the usage of cpu is KX> very high. >> Fortunately, I found it's a paramter about iteration in sdf director >> cause the problem, the default value of it is 0. After I change the >> value to 1, everything is fine. >> >> Thank you very much! >> >> ---------------- >> Best Regards, >> Kun Xiao >> >> >> -----Original Message----- >> From: Dan Higgins [mailto:higgins at nceas.ucsb.edu] >> Sent: 2006年9月28日 13:52 >> To: Efrat Frank >> Cc: Kun Xiao; Kepler-users at ecoinformatics.org >> Subject: Re: [kepler-users] Kepler-users at ecoinformatics.org >> >> I downloaded a new copy of Kepler-1.0.0beta1 (with jre), installed KX> it, >> and also ran it with no problem, so I see no evidence of a memory KX> leak. >> I have 2GB of memory, but the Windows Task manger reports that only KX> a >> bout 520 MB is being used when I am running Kepler. >> >> So I have no idea why you are having the reported problem! The only >> thing I can suggest is perhaps your download was corrupted, so you >> might try downloading the installer again and reinstalling? >> >> Dan Higgins >> >> Efrat Frank wrote: >> >>> I just ran the Kepler 1beta2 release on my windows machine (which KX> is >>> similar to yours) and didn't notice any problems. Perhaps there are >>> some heavy duty background processes running on your machine? >>> >>> EFrat >>> >>> Kun Xiao wrote: >>> >>> I found when I click the button "run the model", there is only 1% >>> memory left. My computer almost down! What's the problem? >>> Is there any memory leak? >>>> >>> BTW, which version do you use? >>>> >>> Thanks! >>>> >>> Kun Xiao >>> ----- Original Message ----- >>> From: Dan Higgins <higgins at nceas.ucsb.edu> >>> Date: Thursday, September 28, 2006 12:23 pm >>> Subject: Re: [kepler-users] Kepler-users at ecoinformatics.org >>>> >>> >>> >>>>> The javaw.exe process is just the java executable that Kepler is >>>>> running under. ('javaw' runs 'java' without displaying the >>>>> background command window). Kepler will try to use a lot of memory >>>>> (RAM)) so there might be a problem on a 1G machine if there are a >>>>> lot of other applications running simultaneously, but I haven't >>>>> seen the problem on a similarly configured Windows machine. >>>>> >>>>> I was not aware of a problem on some Windows machines for the non- KX> >>>>> JRE nstaller. Is this a version problem? >>>>> >>>>> Dan Higgins >>>>> >>>>> Efrat Frank wrote: >>>>> >>>>> >>>>>> For the Windows distribution it's best to download the installer >>>>>> >>>>>> >>>>> with >>>>> >>>>> >>>>>> the JRE (aparently on some Windows machines the non-JRE >>>>>> >>>>>> >>>>> installer >>>>> >>>>> >>>>>> seems to be causing problems). Is this what you installed? >>>>>> >>>>>> Efrat >>>>>> >>>>>> Kun Xiao wrote: >>>>>> >>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> I am a phd student doing some research on SWF and have >>>>>>> >>>>>>> >>>>> downloaded >>>>> >>>>> >>>>>>> kepler 1.0. But the software seems running very slowly even >>>>>>> >>>>>>> >>>>> when I >>>>> >>>>> >>>>>>> run some simple samples such as a workflow only include >>>>>>> >>>>>>> >>>>> add/subtract >>>>> >>>>> >>>>>>> function. And from Windows Task Manager, I found the process >>>>>>> >>>>>>> >>>>> named >>>>> >>>>> >>>>>>> javaw.exe consume too much cpu resource. Sometimes I have to >>>>>>> >>>>>>> >>>>> kill >>>>> >>>>> >>>>>>> that process to keep my computer running. >>>>>>> >>>>>>> >>>>>>> >>>>>>> My running environment is below: >>>>>>> >>>>>>> OS: Windows xp professional >>>>>>> >>>>>>> CPU: Intel duo 2.0 G >>>>>>> >>>>>>> Memory: 1G >>>>>>> >>>>>>> >>>>>>> >>>>>>> Could you tell me what the problem is? Thanks! >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ---------------- >>>>>>> >>>>>>> Best Regards, >>>>>>> >>>>>>> Kun Xiao >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ---------------------------------------------------------------- >>>>>>> >>>>>>> >>>>> -------- >>>>> >>>>> >>>>>>> _______________________________________________ >>>>>>> Kepler-users mailing list >>>>>>> Kepler-users at ecoinformatics.org >>>>>>> >>>>>>> >>>>>>> >>>>> KX> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/keple >>>>> r >>>>> - >>>>> >>>>> >>> users >>> >>> >>>>>>> >>>>>>> >>>>>>> >>>>>> ----------------------------------------------------------------- >>>>>> >>>>>> >>>>> ------- >>>>> >>>>> >>>>>> _______________________________________________ >>>>>> Kepler-users mailing list >>>>>> Kepler-users at ecoinformatics.org >>>>>> >>>>>> >>>>>> >>>>> KX> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/keple >>>>> r >>>>> - >>>>> >>>>> >>> users >>> >>> >>>>>> >>>>>> >>>>>> >>>>> -- >>>>> KX> ******************************************************************* >>>>> Dan Higgins KX> higgins at nceas.ucsb.edu >>>>> http://www.nceas.ucsb.edu/ Ph: 805-893-5127 >>>>> National Center for Ecological Analysis and Synthesis (NCEAS) >>>>> Marine Science Building - Room 3405 Santa Barbara, CA 93195 >>>>> KX> ******************************************************************* >>>>> >>>>> >>>>> >>>>> >>>>> >> >> >> -- >> ******************************************************************* >> Dan Higgins higgins at nceas.ucsb.edu >> http://www.nceas.ucsb.edu/ Ph: 805-893-5127 >> National Center for Ecological Analysis and Synthesis (NCEAS) Marine >> Science Building - Room 3405 Santa Barbara, CA 93195 >> ******************************************************************* >> >> >> >> >> >> KX> KX> KX> -- KX> ******************************************************************* KX> Dan Higgins higgins at nceas.ucsb.edu KX> http://www.nceas.ucsb.edu/ Ph: 805-893-5127 KX> National Center for Ecological Analysis and Synthesis (NCEAS) Marine KX> Science Building - Room 3405 Santa Barbara, CA 93195 KX> ******************************************************************* KX> KX> KX> _______________________________________________ KX> Kepler-users mailing list KX> Kepler-users at ecoinformatics.org KX> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-users
