From nek5000-users at lists.mcs.anl.gov Thu Dec 3 03:55:35 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 10:55:35 +0100 Subject: [Nek5000-users] endian problem Message-ID: <4B178B17.7000902@mech.kth.se> Hi, I am trying to start up a job on intrepid now with the newest version of the code, with a resulting endian problem. It seems the code can handle the geometry file correctly but breaks down after reading the restart files with the wrong endian. Is there any new file format that I should be aware of? The output file is attached. Thanks for any tip! Best regards, Johan -- Johan Ohlsson Department of Mechanics, KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7906876 E-mail: johan at mech.kth.se -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: duct.log1.8192 URL: From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:08:10 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 11:08:10 +0100 Subject: [Nek5000-users] endian problems Message-ID: <4B178E0A.4080304@mech.kth.se> Hi again, It seems to work when I use a slightly older version of the code. There was apparently a change in ic.f where the restart data was read in ('Reading restart data..' --> 'Reading checkpoint data..'). Johan -- Johan Ohlsson Department of Mechanics, KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7906876 E-mail: johan at mech.kth.se From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:14:40 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 04:14:40 -0600 (CST) Subject: [Nek5000-users] endian problem In-Reply-To: <4B178B17.7000902@mech.kth.se> References: <4B178B17.7000902@mech.kth.se> Message-ID: Hi Johan, are you reading two sets of .f0000 files, one w/ geometry and one w/ data? Do they have different endian characteristics? (I think that shouldn't matter - since endian is determined on a file-by-file basis, and the flag set accordingly...) Paul On Thu, 3 Dec 2009, nek5000-users at lists.mcs.anl.gov wrote: > Hi, > > I am trying to start up a job on intrepid now with the newest version of > the code, with a resulting endian problem. It seems the code can handle the > geometry file correctly but breaks down after reading the restart files > with the wrong endian. Is there any new file format that I should be > aware of? > > The output file is attached. Thanks for any tip! > > Best regards, > > Johan > > -- > Johan Ohlsson > Department of Mechanics, KTH > SE-100 44, Stockholm, Sweden > Phone: +46 8 7906876 > E-mail: johan at mech.kth.se > > From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:18:06 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 11:18:06 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <4B178E0A.4080304@mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> Message-ID: Is the restart data a .fld?? or a .f???? file? #Stefan On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: > Hi again, > > It seems to work when I use a slightly older version of the code. There > was apparently a change in ic.f where the restart data was read in > ('Reading restart data..' --> 'Reading checkpoint data..'). > > Johan > > -- > Johan Ohlsson > Department of Mechanics, KTH > SE-100 44, Stockholm, Sweden > Phone: +46 8 7906876 > E-mail: johan at mech.kth.se > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:30:03 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 11:30:03 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: References: <4B178E0A.4080304@mech.kth.se> Message-ID: <4B17932B.6000504@mech.kth.se> It's a blahh??.f0001 file. (16 files) Johan nek5000-users at lists.mcs.anl.gov wrote: > Is the restart data a .fld?? or a .f???? file? > > #Stefan > > On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: > > >> Hi again, >> >> It seems to work when I use a slightly older version of the code. There >> was apparently a change in ic.f where the restart data was read in >> ('Reading restart data..' --> 'Reading checkpoint data..'). >> >> Johan >> >> -- >> Johan Ohlsson >> Department of Mechanics, KTH >> SE-100 44, Stockholm, Sweden >> Phone: +46 8 7906876 >> E-mail: johan at mech.kth.se >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > -- Johan Ohlsson Department of Mechanics, KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7906876 E-mail: johan at mech.kth.se From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:29:03 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 11:29:03 +0100 Subject: [Nek5000-users] endian problem In-Reply-To: References: <4B178B17.7000902@mech.kth.se> Message-ID: <4B1792EF.9090801@mech.kth.se> Hi Paul, Both data and geometry are contained in the restart files. What I meant was that the binary re2 file was read correctly, but not the restart files. But yes, I agree, it should not matter -- that's the whole idea with this if_byte_swap - thing. But anyway, it seems to run with the slightly older version. Thanks, Johan P.S. Did you see the response from Peake? Looks promising, once I choose my words carefully in the revised version... nek5000-users at lists.mcs.anl.gov wrote: > > Hi Johan, > > are you reading two sets of .f0000 files, one w/ geometry > and one w/ data? > > Do they have different endian characteristics? (I think > that shouldn't matter - since endian is determined on a > file-by-file basis, and the flag set accordingly...) > > Paul > > > On Thu, 3 Dec 2009, nek5000-users at lists.mcs.anl.gov wrote: > >> Hi, >> >> I am trying to start up a job on intrepid now with the newest version of >> the code, with a resulting endian problem. It seems the code can >> handle the >> geometry file correctly but breaks down after reading the restart files >> with the wrong endian. Is there any new file format that I should be >> aware of? >> >> The output file is attached. Thanks for any tip! >> >> Best regards, >> >> Johan >> >> -- >> Johan Ohlsson >> Department of Mechanics, KTH >> SE-100 44, Stockholm, Sweden >> Phone: +46 8 7906876 >> E-mail: johan at mech.kth.se >> >> > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users -- Johan Ohlsson Department of Mechanics, KTH SE-100 44, Stockholm, Sweden Phone: +46 8 7906876 E-mail: johan at mech.kth.se From nek5000-users at lists.mcs.anl.gov Thu Dec 3 04:40:57 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 11:40:57 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <4B17932B.6000504@mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> Message-ID: Ok and do you need to swap the bytes on Intrepid meaning is your restart data of small endian type? On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: > It's a blahh??.f0001 file. (16 files) > > Johan > > nek5000-users at lists.mcs.anl.gov wrote: >> Is the restart data a .fld?? or a .f???? file? >> >> #Stefan >> >> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >> >> >>> Hi again, >>> >>> It seems to work when I use a slightly older version of the code. There >>> was apparently a change in ic.f where the restart data was read in >>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>> >>> Johan >>> >>> -- >>> Johan Ohlsson >>> Department of Mechanics, KTH >>> SE-100 44, Stockholm, Sweden >>> Phone: +46 8 7906876 >>> E-mail: johan at mech.kth.se >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> > > > -- > Johan Ohlsson > Department of Mechanics, KTH > SE-100 44, Stockholm, Sweden > Phone: +46 8 7906876 > E-mail: johan at mech.kth.se > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Thu Dec 3 06:54:13 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 13:54:13 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> Message-ID: <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> Yes, the files were written out on a AMD Opteron Linux system. J Quoting nek5000-users at lists.mcs.anl.gov: > Ok and do you need to swap the bytes on Intrepid meaning is your > restart data of small endian type? > > On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: > >> It's a blahh??.f0001 file. (16 files) >> >> Johan >> >> nek5000-users at lists.mcs.anl.gov wrote: >>> Is the restart data a .fld?? or a .f???? file? >>> >>> #Stefan >>> >>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>> >>> >>>> Hi again, >>>> >>>> It seems to work when I use a slightly older version of the code. There >>>> was apparently a change in ic.f where the restart data was read in >>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>> >>>> Johan >>>> >>>> -- >>>> Johan Ohlsson >>>> Department of Mechanics, KTH >>>> SE-100 44, Stockholm, Sweden >>>> Phone: +46 8 7906876 >>>> E-mail: johan at mech.kth.se >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >> >> >> -- >> Johan Ohlsson >> Department of Mechanics, KTH >> SE-100 44, Stockholm, Sweden >> Phone: +46 8 7906876 >> E-mail: johan at mech.kth.se >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. From nek5000-users at lists.mcs.anl.gov Thu Dec 3 07:13:51 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 14:13:51 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> Message-ID: <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> It turns out that the old version of the code hangs after a random number of time steps, so I need the new version (r374) to work. Quoting nek5000-users at lists.mcs.anl.gov: > Yes, the files were written out on a AMD Opteron Linux system. > > J > > Quoting nek5000-users at lists.mcs.anl.gov: > >> Ok and do you need to swap the bytes on Intrepid meaning is your >> restart data of small endian type? >> >> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >> >>> It's a blahh??.f0001 file. (16 files) >>> >>> Johan >>> >>> nek5000-users at lists.mcs.anl.gov wrote: >>>> Is the restart data a .fld?? or a .f???? file? >>>> >>>> #Stefan >>>> >>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>> >>>> >>>>> Hi again, >>>>> >>>>> It seems to work when I use a slightly older version of the code. There >>>>> was apparently a change in ic.f where the restart data was read in >>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>> >>>>> Johan >>>>> >>>>> -- >>>>> Johan Ohlsson >>>>> Department of Mechanics, KTH >>>>> SE-100 44, Stockholm, Sweden >>>>> Phone: +46 8 7906876 >>>>> E-mail: johan at mech.kth.se >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>> >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>> >>> >>> -- >>> Johan Ohlsson >>> Department of Mechanics, KTH >>> SE-100 44, Stockholm, Sweden >>> Phone: +46 8 7906876 >>> E-mail: johan at mech.kth.se >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. From nek5000-users at lists.mcs.anl.gov Thu Dec 3 07:19:58 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 14:19:58 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> Message-ID: <3B7FD281-B2BF-4346-AD37-AA18EB2C42A6@lav.mavt.ethz.ch> - Why do you think r374 will not have the same problem? - At which timestep did it hang? #Stefan On Dec 3, 2009, at 2:13 PM, nek5000-users at lists.mcs.anl.gov wrote: > It turns out that the old version of the code hangs after a random number of time steps, so I need the new version (r374) to work. > > > > Quoting nek5000-users at lists.mcs.anl.gov: > >> Yes, the files were written out on a AMD Opteron Linux system. >> >> J >> >> Quoting nek5000-users at lists.mcs.anl.gov: >> >>> Ok and do you need to swap the bytes on Intrepid meaning is your restart data of small endian type? >>> >>> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >>> >>>> It's a blahh??.f0001 file. (16 files) >>>> >>>> Johan >>>> >>>> nek5000-users at lists.mcs.anl.gov wrote: >>>>> Is the restart data a .fld?? or a .f???? file? >>>>> >>>>> #Stefan >>>>> >>>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>> >>>>> >>>>>> Hi again, >>>>>> >>>>>> It seems to work when I use a slightly older version of the code. There >>>>>> was apparently a change in ic.f where the restart data was read in >>>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>>> >>>>>> Johan >>>>>> >>>>>> -- >>>>>> Johan Ohlsson >>>>>> Department of Mechanics, KTH >>>>>> SE-100 44, Stockholm, Sweden >>>>>> Phone: +46 8 7906876 >>>>>> E-mail: johan at mech.kth.se >>>>>> >>>>>> _______________________________________________ >>>>>> Nek5000-users mailing list >>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>> >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>> >>>> >>>> >>>> -- >>>> Johan Ohlsson >>>> Department of Mechanics, KTH >>>> SE-100 44, Stockholm, Sweden >>>> Phone: +46 8 7906876 >>>> E-mail: johan at mech.kth.se >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Thu Dec 3 09:07:46 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 16:07:46 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> Message-ID: <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> I was able to identify where the problems is. Yes you're right there is bug in the latest repo version if you'll try to restart and you need to swap the bytes. Fixed it in my dev-version. #Stefan On Dec 3, 2009, at 2:13 PM, nek5000-users at lists.mcs.anl.gov wrote: > It turns out that the old version of the code hangs after a random number of time steps, so I need the new version (r374) to work. > > > > Quoting nek5000-users at lists.mcs.anl.gov: > >> Yes, the files were written out on a AMD Opteron Linux system. >> >> J >> >> Quoting nek5000-users at lists.mcs.anl.gov: >> >>> Ok and do you need to swap the bytes on Intrepid meaning is your restart data of small endian type? >>> >>> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >>> >>>> It's a blahh??.f0001 file. (16 files) >>>> >>>> Johan >>>> >>>> nek5000-users at lists.mcs.anl.gov wrote: >>>>> Is the restart data a .fld?? or a .f???? file? >>>>> >>>>> #Stefan >>>>> >>>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>> >>>>> >>>>>> Hi again, >>>>>> >>>>>> It seems to work when I use a slightly older version of the code. There >>>>>> was apparently a change in ic.f where the restart data was read in >>>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>>> >>>>>> Johan >>>>>> >>>>>> -- >>>>>> Johan Ohlsson >>>>>> Department of Mechanics, KTH >>>>>> SE-100 44, Stockholm, Sweden >>>>>> Phone: +46 8 7906876 >>>>>> E-mail: johan at mech.kth.se >>>>>> >>>>>> _______________________________________________ >>>>>> Nek5000-users mailing list >>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>> >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>> >>>> >>>> >>>> -- >>>> Johan Ohlsson >>>> Department of Mechanics, KTH >>>> SE-100 44, Stockholm, Sweden >>>> Phone: +46 8 7906876 >>>> E-mail: johan at mech.kth.se >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Thu Dec 3 09:20:21 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 03 Dec 2009 16:20:21 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> Message-ID: <20091203162021.qr25ysow00sc44k0@www.mech.kth.se> That's great Stefan! Is this version available in the svn-system? Quoting nek5000-users at lists.mcs.anl.gov: > I was able to identify where the problems is. Yes you're right there > is bug in the latest repo version if you'll try to restart and you > need to swap the bytes. > > Fixed it in my dev-version. > > #Stefan > > > On Dec 3, 2009, at 2:13 PM, nek5000-users at lists.mcs.anl.gov wrote: > >> It turns out that the old version of the code hangs after a random >> number of time steps, so I need the new version (r374) to work. >> >> >> >> Quoting nek5000-users at lists.mcs.anl.gov: >> >>> Yes, the files were written out on a AMD Opteron Linux system. >>> >>> J >>> >>> Quoting nek5000-users at lists.mcs.anl.gov: >>> >>>> Ok and do you need to swap the bytes on Intrepid meaning is your >>>> restart data of small endian type? >>>> >>>> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>> >>>>> It's a blahh??.f0001 file. (16 files) >>>>> >>>>> Johan >>>>> >>>>> nek5000-users at lists.mcs.anl.gov wrote: >>>>>> Is the restart data a .fld?? or a .f???? file? >>>>>> >>>>>> #Stefan >>>>>> >>>>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>>> >>>>>> >>>>>>> Hi again, >>>>>>> >>>>>>> It seems to work when I use a slightly older version of the code. There >>>>>>> was apparently a change in ic.f where the restart data was read in >>>>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>>>> >>>>>>> Johan >>>>>>> >>>>>>> -- >>>>>>> Johan Ohlsson >>>>>>> Department of Mechanics, KTH >>>>>>> SE-100 44, Stockholm, Sweden >>>>>>> Phone: +46 8 7906876 >>>>>>> E-mail: johan at mech.kth.se >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Nek5000-users mailing list >>>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Nek5000-users mailing list >>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>> >>>>> >>>>> >>>>> -- >>>>> Johan Ohlsson >>>>> Department of Mechanics, KTH >>>>> SE-100 44, Stockholm, Sweden >>>>> Phone: +46 8 7906876 >>>>> E-mail: johan at mech.kth.se >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>> >>> >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. From nek5000-users at lists.mcs.anl.gov Thu Dec 3 09:21:22 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 09:21:22 -0600 (CST) Subject: [Nek5000-users] endian problems In-Reply-To: <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> Message-ID: Hi Stefan, Just curious - do you plan to update the repo, or are you saving multiple upgrades for a single commit? I don't care either way, just trying to understand good (Stefan) programming practices, so that I can emulate them! :) Paul On Thu, 3 Dec 2009, nek5000-users at lists.mcs.anl.gov wrote: > I was able to identify where the problems is. Yes you're right there is bug in the latest repo version if you'll try to restart and you need to swap the bytes. > > Fixed it in my dev-version. > > #Stefan > > > On Dec 3, 2009, at 2:13 PM, nek5000-users at lists.mcs.anl.gov wrote: > >> It turns out that the old version of the code hangs after a random number of time steps, so I need the new version (r374) to work. >> >> >> >> Quoting nek5000-users at lists.mcs.anl.gov: >> >>> Yes, the files were written out on a AMD Opteron Linux system. >>> >>> J >>> >>> Quoting nek5000-users at lists.mcs.anl.gov: >>> >>>> Ok and do you need to swap the bytes on Intrepid meaning is your restart data of small endian type? >>>> >>>> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>> >>>>> It's a blahh??.f0001 file. (16 files) >>>>> >>>>> Johan >>>>> >>>>> nek5000-users at lists.mcs.anl.gov wrote: >>>>>> Is the restart data a .fld?? or a .f???? file? >>>>>> >>>>>> #Stefan >>>>>> >>>>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>>> >>>>>> >>>>>>> Hi again, >>>>>>> >>>>>>> It seems to work when I use a slightly older version of the code. There >>>>>>> was apparently a change in ic.f where the restart data was read in >>>>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>>>> >>>>>>> Johan >>>>>>> >>>>>>> -- >>>>>>> Johan Ohlsson >>>>>>> Department of Mechanics, KTH >>>>>>> SE-100 44, Stockholm, Sweden >>>>>>> Phone: +46 8 7906876 >>>>>>> E-mail: johan at mech.kth.se >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Nek5000-users mailing list >>>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Nek5000-users mailing list >>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>> >>>>> >>>>> >>>>> -- >>>>> Johan Ohlsson >>>>> Department of Mechanics, KTH >>>>> SE-100 44, Stockholm, Sweden >>>>> Phone: +46 8 7906876 >>>>> E-mail: johan at mech.kth.se >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>> >>> >>> >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> >> >> >> ---------------------------------------------------------------- >> This message was sent using IMP, the Internet Messaging Program. >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users > From nek5000-users at lists.mcs.anl.gov Thu Dec 3 11:46:31 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Thu, 3 Dec 2009 18:46:31 +0100 Subject: [Nek5000-users] endian problems In-Reply-To: <20091203162021.qr25ysow00sc44k0@www.mech.kth.se> References: <4B178E0A.4080304@mech.kth.se> <4B17932B.6000504@mech.kth.se> <20091203135413.m56xmer9wss8sgwg@www.mech.kth.se> <20091203141351.m57wsxi5cgs8ss0s@www.mech.kth.se> <34A832F5-9016-4734-9D11-3B5C1AAF62EF@lav.mavt.ethz.ch> <20091203162021.qr25ysow00sc44k0@www.mech.kth.se> Message-ID: I will submit it after doing some tests. Unfortunately I am quite busy these days so I don't think it will be available before next week. #Stefan On Dec 3, 2009, at 4:20 PM, nek5000-users at lists.mcs.anl.gov wrote: > That's great Stefan! Is this version available in the svn-system? > > > > Quoting nek5000-users at lists.mcs.anl.gov: > >> I was able to identify where the problems is. Yes you're right there is bug in the latest repo version if you'll try to restart and you need to swap the bytes. >> >> Fixed it in my dev-version. >> >> #Stefan >> >> >> On Dec 3, 2009, at 2:13 PM, nek5000-users at lists.mcs.anl.gov wrote: >> >>> It turns out that the old version of the code hangs after a random number of time steps, so I need the new version (r374) to work. >>> >>> >>> >>> Quoting nek5000-users at lists.mcs.anl.gov: >>> >>>> Yes, the files were written out on a AMD Opteron Linux system. >>>> >>>> J >>>> >>>> Quoting nek5000-users at lists.mcs.anl.gov: >>>> >>>>> Ok and do you need to swap the bytes on Intrepid meaning is your restart data of small endian type? >>>>> >>>>> On Dec 3, 2009, at 11:30 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>> >>>>>> It's a blahh??.f0001 file. (16 files) >>>>>> >>>>>> Johan >>>>>> >>>>>> nek5000-users at lists.mcs.anl.gov wrote: >>>>>>> Is the restart data a .fld?? or a .f???? file? >>>>>>> >>>>>>> #Stefan >>>>>>> >>>>>>> On Dec 3, 2009, at 11:08 AM, nek5000-users at lists.mcs.anl.gov wrote: >>>>>>> >>>>>>> >>>>>>>> Hi again, >>>>>>>> >>>>>>>> It seems to work when I use a slightly older version of the code. There >>>>>>>> was apparently a change in ic.f where the restart data was read in >>>>>>>> ('Reading restart data..' --> 'Reading checkpoint data..'). >>>>>>>> >>>>>>>> Johan >>>>>>>> >>>>>>>> -- >>>>>>>> Johan Ohlsson >>>>>>>> Department of Mechanics, KTH >>>>>>>> SE-100 44, Stockholm, Sweden >>>>>>>> Phone: +46 8 7906876 >>>>>>>> E-mail: johan at mech.kth.se >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Nek5000-users mailing list >>>>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Nek5000-users mailing list >>>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Johan Ohlsson >>>>>> Department of Mechanics, KTH >>>>>> SE-100 44, Stockholm, Sweden >>>>>> Phone: +46 8 7906876 >>>>>> E-mail: johan at mech.kth.se >>>>>> >>>>>> _______________________________________________ >>>>>> Nek5000-users mailing list >>>>>> Nek5000-users at lists.mcs.anl.gov >>>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>> >>>>> _______________________________________________ >>>>> Nek5000-users mailing list >>>>> Nek5000-users at lists.mcs.anl.gov >>>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>>>> >>>> >>>> >>>> >>>> ---------------------------------------------------------------- >>>> This message was sent using IMP, the Internet Messaging Program. >>>> >>>> _______________________________________________ >>>> Nek5000-users mailing list >>>> Nek5000-users at lists.mcs.anl.gov >>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >>> >>> >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> _______________________________________________ >>> Nek5000-users mailing list >>> Nek5000-users at lists.mcs.anl.gov >>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> >> _______________________________________________ >> Nek5000-users mailing list >> Nek5000-users at lists.mcs.anl.gov >> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >> > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Wed Dec 9 10:31:41 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Wed, 9 Dec 2009 17:31:41 +0100 Subject: [Nek5000-users] makenek for systemX In-Reply-To: <3F004F71-B82D-4A2E-9D6F-1BF79F61C764@rsmas.miami.edu> References: <3F004F71-B82D-4A2E-9D6F-1BF79F61C764@rsmas.miami.edu> Message-ID: <3A92A75F-A61F-4A52-9DE4-DD64E60E4380@lav.mavt.ethz.ch> Hi, just follow this example: https://nek5000.mcs.anl.gov/index.php/Guided_Tutorial and set in makenek the following variables F77 = mpif77 CC = mpixlc and you're all set. Stefan On Nov 26, 2009, at 3:46 PM, nek5000-users at lists.mcs.anl.gov wrote: > > Dear developers, > > I started using the latest version of nek5000 yesterday, shifting from > one that I got from Paul about 2 years ago (I have been running nek5000 > since 2002 or so). Paul compiled it on my 8-processor > workstation using PG. After a short warm up, I want to run on systemX which is based on > IBM compilers. Attached is the makefile that I have been using on systemX > for the older nek5000 version. Can you suggest changes in the makefile > that came with the svn version to compile on sysX? > > Thanks!! > > ps: the new nek5000 is looking very good! > > Tamay Ozgokmen > Associate Professor > Division of Meteorology and Physical Oceanography > Rosenstiel School of Marine and Atmospheric Science > 4600 Rickenbacker Causeway > Miami, FL 33149-1098, USA > Phone : 305 421 4053 > Fax : 305 421 4696 > Email : tozgokmen at rsmas.miami.edu > Web : http://www.rsmas.miami.edu/personal/tamay > > > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Fri Dec 11 07:12:00 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 11 Dec 2009 14:12:00 +0100 Subject: [Nek5000-users] Fixed mass flux Message-ID: <20091211141200.a21lzgjbwgkcsc4k@webmail.uni-karlsruhe.de> Hello, currently I'm trying to run some simple cases to getting used to the code. But unfortunately I didn't understand how it's possible to predefine a fixed mass flux. Do I just have to enter a fixed value for Parameter 54 resp. 55 in the rea-File? Maybe this question is a little stupid, but whatever I tried, the velocity in the result file .fld* was more or less equal zero.. Best regards, Fred From nek5000-users at lists.mcs.anl.gov Fri Dec 11 07:28:00 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Fri, 11 Dec 2009 14:28:00 +0100 Subject: [Nek5000-users] Fixed mass flux In-Reply-To: <20091211141200.a21lzgjbwgkcsc4k@webmail.uni-karlsruhe.de> References: <20091211141200.a21lzgjbwgkcsc4k@webmail.uni-karlsruhe.de> Message-ID: Hi Fred, in general there are two different ways to do that. Anyway just have a look how we do it in the turbChannel example. The general idea is the following: - call set_forcing() in userchk() (e.g. call set_forcing(ffx_,xv,1) - add the acceleration term ffx_ to the momentum eqn. (e.g. set ffx = ffx_ in userf()) Please note that the variable ffx_ has to be visible for userf() and userchk() (e.g. use a common block). The initial acceleration can be specified by param(71) . #Stefan On Dec 11, 2009, at 2:12 PM, nek5000-users at lists.mcs.anl.gov wrote: > Hello, > > currently I'm trying to run some simple cases to getting used to the code. But unfortunately I didn't understand how it's possible to predefine a fixed mass flux. Do I just have to enter a fixed value for Parameter 54 resp. 55 in the rea-File? Maybe this question is a little stupid, but whatever I tried, the velocity in the result file .fld* was more or less equal zero.. > > Best regards, > Fred > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users From nek5000-users at lists.mcs.anl.gov Tue Dec 29 17:56:26 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 29 Dec 2009 17:56:26 -0600 Subject: [Nek5000-users] genbox compiling problems Message-ID: <88fcdb37de7fa0ee268e82039979be6d.squirrel@webmail.uic.edu> Hi, I am new to Nek5000. I can compile the code with OpenMPI compilers 'mpi77' & 'mpicc' and run the code in parallel up to 8 processors. Though, I do get some warnings during compilation. Then, I wanted to setup a 2D case for flow over a backward facing step and run it. This is were I ran into hurdles. I could not compile genbox using maketools as it gave me compiling errors. The compilers I have specified in the maketools executable are: *************** # specify your compilers here F77="gfortran" CC="gcc" *************** Following are the compiling errors: **************** [harish at cmtl tools]$ ./maketools genbox ---------------------- Make genbox... ---------------------- make[1]: Entering directory `/home/harish/nek5_svn/trunk/tools/genbox' gfortran -mcmodel=medium -c genbox.f In file genbox.f:648 buf(8) = 'C' 1 Error: Can't convert CHARACTER(1) to REAL(4) at (1) In file genbox.f:1551 r = x(2) 1 Warning: Array reference at (1) is out of bounds make[1]: *** [genbox.o] Error 1 make[1]: Leaving directory `/home/harish/nek5_svn/trunk/tools/genbox' make: *** [all] Error 1 ***************** Then, I compiled prenek and used 'prex' for meshing. But, I had trouble understanding how to use it. Can you direct me to some resources which explain the use of 'prex' for making meshes. For the backward facing step, genbox would be most useful to make the mesh as it has only 3 boxes and not a complicated geometry. But knowing how to use 'prex' would be useful as well. Thanks, Harish Kanchi. From nek5000-users at lists.mcs.anl.gov Tue Dec 29 21:23:39 2009 From: nek5000-users at lists.mcs.anl.gov (nek5000-users at lists.mcs.anl.gov) Date: Tue, 29 Dec 2009 21:23:39 -0600 (CST) Subject: [Nek5000-users] genbox compiling problems In-Reply-To: <88fcdb37de7fa0ee268e82039979be6d.squirrel@webmail.uic.edu> References: <88fcdb37de7fa0ee268e82039979be6d.squirrel@webmail.uic.edu> Message-ID: Hi Harish, go to nek5_svn/trunk then svn update tools then try to edit / run maketools again with gfortran For a prenek manual, you might look at the (very dated) manual www.mcs.anl.gov/~fischer/nekmanual.pdf and also: www.mcs.anl.gov/~fischer/primer.pdf Paul On Tue, 29 Dec 2009, nek5000-users at lists.mcs.anl.gov wrote: > Hi, > > I am new to Nek5000. I can compile the code with OpenMPI compilers 'mpi77' > & 'mpicc' and run the code in parallel up to 8 processors. Though, I do > get some warnings during compilation. > > Then, I wanted to setup a 2D case for flow over a backward facing step and > run it. This is were I ran into hurdles. > > I could not compile genbox using maketools as it gave me compiling errors. > The compilers I have specified in the maketools executable are: > > *************** > # specify your compilers here > F77="gfortran" > CC="gcc" > *************** > > Following are the compiling errors: > > **************** > [harish at cmtl tools]$ ./maketools genbox > ---------------------- > Make genbox... > ---------------------- > make[1]: Entering directory `/home/harish/nek5_svn/trunk/tools/genbox' > gfortran -mcmodel=medium -c genbox.f > In file genbox.f:648 > > buf(8) = 'C' > 1 > Error: Can't convert CHARACTER(1) to REAL(4) at (1) > In file genbox.f:1551 > > r = x(2) > 1 > Warning: Array reference at (1) is out of bounds > make[1]: *** [genbox.o] Error 1 > make[1]: Leaving directory `/home/harish/nek5_svn/trunk/tools/genbox' > make: *** [all] Error 1 > ***************** > > Then, I compiled prenek and used 'prex' for meshing. But, I had trouble > understanding how to use it. Can you direct me to some resources which > explain the use of 'prex' for making meshes. > > For the backward facing step, genbox would be most useful to make the mesh > as it has only 3 boxes and not a complicated geometry. But knowing how to > use 'prex' would be useful as well. > > Thanks, > > Harish Kanchi. > > > _______________________________________________ > Nek5000-users mailing list > Nek5000-users at lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users >