take a look at your memory consumption. you will find that you are using an in-memory model so all your data will be held in RAM and swap space. you can imagine that your machine at some point will constantly swap data, ultimately to its own demise.
here are two things to look into: persistent storage & runtime GC. 1. start to use a persistent jena model interface 2. switch to a 64bit java run-time to improve overall performance. On Mon, Apr 2, 2012 at 5:39 AM, [email protected] <[email protected]> wrote: > Hi Marco and hi all, > > sorry to bother you. I have restarted the ontology population trying not to > use a reasoner. > So, instead of using this instruction: > > model = ModelFactory.createOntologyModel(org.mindswap.pellet.jena. > PelletReasonerFactory.THE_SPEC, null); > > I'm using this one: > > model = ModelFactory.createOntologyModel(); > > but, after some hours of population, it seems to me that the performances are > slowing down again. > Is it correct to use the previous instruction instead of the one related to > Pellet? Do I have to take into > account other instructions? > > Thanks a lot! > > Paola > >>----Messaggio originale---- >>Da: [email protected] >>Data: 30/03/2012 15.43 >>A: "[email protected]"<[email protected]> >>Cc: <[email protected]> >>Ogg: Re: Re: Re: Re: Re: Jena slow ontology population >> >>you can take any tool of your choice but SPARQL updates and Jena API >>calls on the model come to mind here, of course. >> >>so try updates without reasoning and let us know if you experience >>performance advantages. >> >> >> >>On Fri, Mar 30, 2012 at 9:34 AM, [email protected] <[email protected]> > wrote: >>> thanks a lot! >>> >>> Sorry, but what do you mean with "plain update"? >>> >>> Paola >>> >>>>----Messaggio originale---- >>>>Da: [email protected] >>>>Data: 30/03/2012 15.21 >>>>A: "[email protected]"<[email protected]> >>>>Cc: <[email protected]> >>>>Ogg: Re: Re: Re: Re: Jena slow ontology population >>>> >>>>so why don't you try a plain update and see how that performs? >>>> >>>> >>>>On Fri, Mar 30, 2012 at 8:59 AM, [email protected] <[email protected]> >>> wrote: >>>>> yes, I insert individuals in every step. >>>>> >>>>> thanks! >>>>> >>>>>>----Messaggio originale---- >>>>>>Da: [email protected] >>>>>>Data: 30/03/2012 14.47 >>>>>>A: "[email protected]"<[email protected]> >>>>>>Cc: <[email protected]> >>>>>>Ogg: Re: Re: Re: Jena slow ontology population >>>>>> >>>>>>So all updates are A-box related? >>>>>> >>>>>>On Fri, Mar 30, 2012 at 8:22 AM, [email protected] <[email protected]> >>>>> wrote: >>>>>>> Thanks a lot for your reply! >>>>>>> As you have assumed, I don't use neither TDB or SDB...I have an in >>> memory- >>>>>>> model at runtime. >>>>>>> It seems to me that my model does not require complicate inferences... > at >>>>> every >>>>>>> step, >>>>>>> it extracts a tuple from the DB and generates the corresponding >>> individuals >>>>>>> with: >>>>>>> model.createIndividual; the properties are added using the addProperty >>>>>>> method. >>>>>>> >>>>>>> The only thing that can be complicated to be calculated is that, if an >>>>>>> individual has the same >>>>>>> id of the last one processed, I extract the latter from the ontology to >>> add >>>>>>> some properties... >>>>>>> >>>>>>> thanks a lot again! >>>>>>> >>>>>>> Paola >>>>>>> >>>>>>> >>>>>>> >>>>>>>>----Messaggio originale---- >>>>>>>>Da: [email protected] >>>>>>>>Data: 30/03/2012 13.56 >>>>>>>>A: "[email protected]"<[email protected]> >>>>>>>>Cc: <[email protected]> >>>>>>>>Ogg: Re: Re: Jena slow ontology population >>>>>>>> >>>>>>>>your hardware looks good, what type of inference does your model >>>>>>>>require for updates? >>>>>>>> >>>>>>>>Pellet is a very powerful reasoning system which can result in >>>>>>>>significant computational complexity for the subsystem to deal with >>>>>>>>depending on your level of inference that is of course. So you better >>>>>>>>know exactly what level of reasoning your application requires in >>>>>>>>advance to optimize the system at run-time. >>>>>>>> >>>>>>>>TBD, SDB are persistent storage options, from what you have just said >>>>>>>>I can assume you are using an in-memory model at run-time. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>On Fri, Mar 30, 2012 at 7:41 AM, [email protected] <villa.va@libero. > it> >>>>>>> wrote: >>>>>>>>> Hi Marco, thanks a lot for your help! :-) >>>>>>>>> >>>>>>>>> Version of Jena: 2.6.4 >>>>>>>>> OS: Linux 64 bit >>>>>>>>> reasoner: Pellet >>>>>>>>> TDB and SDB: I don't use them, I believed that they were related to > RDF >>>>>>> while >>>>>>>>> I'm using OWL, but maybe I can be wrong... >>>>>>>>> >>>>>>>>> current hw: >>>>>>>>> *-CPU >>>>>>>>> description: CPU >>>>>>>>> product: Intel(R) Xeon(R) CPU >>>>>>>>> version: Intel(R) Xeon(R) CPU E5620 >>>>>>>>> slot: CPU 1 >>>>>>>>> size: 2400MHz >>>>>>>>> capacity: 2400MHz >>>>>>>>> width: 64 bits >>>>>>>>> clock: 133MHz >>>>>>>>> *-memory >>>>>>>>> description: System Memory >>>>>>>>> physical id: 24 >>>>>>>>> slot: System board or motherboard >>>>>>>>> size: 8GiB >>>>>>>>> >>>>>>>>> If you want to know other information, please tell me! And thanks a > lot >>>>>>> again! >>>>>>>>> >>>>>>>>> Paola >>>>>>>>> >>>>>>>>>>----Messaggio originale---- >>>>>>>>>>Da: [email protected] >>>>>>>>>>Data: 30/03/2012 11.36 >>>>>>>>>>A: <[email protected]>, "[email protected]"<villa. >>>>> va@libero. >>>>>>>>> it> >>>>>>>>>>Ogg: Re: Jena slow ontology population >>>>>>>>>> >>>>>>>>>>can you please be a bit more specific about the configuration you > have >>>>>>>>>>currently installed? e.g. version of Jena, back end type (TDB, SDB, >>>>>>>>>>in-memory), current hardware, and OS (32bit, 64bit, linux, >>>>>>>>>>windows,mac), and type of reasoner you use. >>>>>>>>>> >>>>>>>>>>--- >>>>>>>>>>Marco Neumann >>>>>>>>>>KONA >>>>>>>>>> >>>>>>>>>>Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% >>>>>>>>>>with the lotico community discount code 'STMN' >>>>>>>>>>http://www.lotico.com/evt/SemTechSF2012/ >>>>>>>>>> >>>>>>>>>>On Fri, Mar 30, 2012 at 4:22 AM, [email protected] <villa.va@libero. >>> it> >>>>>>>>> wrote: >>>>>>>>>>> Dear users, >>>>>>>>>>> >>>>>>>>>>> I am new to Jena and have a problem with Jena OWL-API. I have > written >>>>>>> some >>>>>>>>>>> simple Java code that: >>>>>>>>>>> >>>>>>>>>>> 1) extracts some tuples (around 1 million tuples) of data from a >>>>> database >>>>>>>>>>> 2) puts each of them in an ontology file using Jena library. In >>>>>>> particular, >>>>>>>>> it >>>>>>>>>>> takes every tuple and inserts >>>>>>>>>>> every element in a class of the ontology file, by using methods > like >>>>>>>>>>> createIndividual, addproperty, model.write. >>>>>>>>>>> >>>>>>>>>>> The insertion phase is very quick when my program starts, but the >>>>>>>>> performances >>>>>>>>>>> became worse and worse as >>>>>>>>>>> time goes by. The DB man has told me that the database has no >>> problems. >>>>> I >>>>>>>>> have >>>>>>>>>>> run the program with 5 gigabytes >>>>>>>>>>> of RAM...it simply ends up to use them all. Is that normal? Do I > have >>>>> to >>>>>>>>> use a >>>>>>>>>>> more powerful PC to manage such >>>>>>>>>>> data? I can give you more details about my problem if you tell me >>> what >>>>> I >>>>>>>>> have >>>>>>>>>>> to look for exactly. >>>>>>>>>>> >>>>>>>>>>> Thanks a lot, >>>>>>>>>>> >>>>>>>>>>> Paola >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>-- >>>>>>>> >>>>>>>> >>>>>>>>--- >>>>>>>>Marco Neumann >>>>>>>>KONA >>>>>>>> >>>>>>>>Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% >>>>>>>>with the lotico community discount code 'STMN' >>>>>>>>http://www.lotico.com/evt/SemTechSF2012/ >>>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>>-- >>>>>> >>>>>> >>>>>>--- >>>>>>Marco Neumann >>>>>>KONA >>>>>> >>>>>>Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% >>>>>>with the lotico community discount code 'STMN' >>>>>>http://www.lotico.com/evt/SemTechSF2012/ >>>>>> >>>>> >>>>> >>>> >>>> >>>> >>>>-- >>>> >>>> >>>>--- >>>>Marco Neumann >>>>KONA >>>> >>>>Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% >>>>with the lotico community discount code 'STMN' >>>>http://www.lotico.com/evt/SemTechSF2012/ >>>> >>> >>> >> >> >> >>-- >> >> >>--- >>Marco Neumann >>KONA >> >>Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% >>with the lotico community discount code 'STMN' >>http://www.lotico.com/evt/SemTechSF2012/ >> > > -- --- Marco Neumann KONA Join us at SemTech Biz in San Francisco June 3-7 2012 and save 15% with the lotico community discount code 'STMN' http://www.lotico.com/evt/SemTechSF2012/
