From jnkim at ncsa.uiuc.edu Wed Jul 21 15:19:59 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Wed Jul 21 16:20:14 2004 Subject: [Ohmms-dev] download Message-ID: <5.1.0.14.2.20040721161748.02bbc950@pop.ncsa.uiuc.edu> Download is updated: http://www.mcc.uiuc.edu/ohmms/sw/archive.html I will use mailing list for such information from now on. Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From jnkim at ncsa.uiuc.edu Fri Jul 23 09:46:38 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Fri Jul 23 10:46:40 2004 Subject: [Ohmms-dev] a patch is uploaded Message-ID: <5.1.0.14.2.20040723104325.02a673a8@pop.ncsa.uiuc.edu> Check out for a patch http://www.mcc.uiuc.edu/ohmms/sw/archive.html Ohmms-v0.2.4-patch-20040723.tar a patch for cmake and ParticleIO. Apply it in the top directory. Contains: CMakeLists.txt CMake/FindLapack.cmake src/FortranLib/CMakeLists.txt src/ParticleIO/RecordParticleAttrib.h src/ParticleIO/RecordParticle.cpp Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From jnkim at ncsa.uiuc.edu Fri Jul 23 10:09:20 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Fri Jul 23 11:09:28 2004 Subject: [Ohmms-dev] Re: lattice.set In-Reply-To: References: <5.1.0.14.2.20040723103017.02a508b8@pop.ncsa.uiuc.edu> <5.1.0.14.2.20040723100814.02a77b30@pop.ncsa.uiuc.edu> <5.1.0.14.2.20040723100814.02a77b30@pop.ncsa.uiuc.edu> <5.1.0.14.2.20040723103017.02a508b8@pop.ncsa.uiuc.edu> Message-ID: <5.1.0.14.2.20040723110348.02b33500@pop.ncsa.uiuc.edu> I was aware of the potential problem but let it go until you point it out. I added to the patch and updated cvs. Use myPtcl->Lattice.update(h0); instead of set. At 11:42 AM 7/23/2004 -0400, you wrote: >Hi Jeongnim, > >I added my name to the ohmms e-mail list and will let the others here know >about it. > >Forget about my comment on the restart file. I used a wrong stride value >and ohmms did not write out the last MD step results. Now I use stride=1 >and it writes out the correct lattice parameters. But the potential >routine still does not get the correct NN list. > >Thanks for your help. > >Cheers, > Richard > > >On Fri, 23 Jul 2004, Jeongnim Kim wrote: > > > I know why your xyz file was wrong. Forgot to use reference, since I > > totally forgot that you may change the lattice! I fixed them. > > > > For your question: I do not understand what you mean by the last one. > > Not having the original lattice is okay if you have changed it: > > If you have set it with a new value, the restart will use the value set > by it. > > But, what is the last one? > > > > BTW, tell people to subscribe to the mailing list. I will notice any cvs > > update that may affect to the calculations, like this case. > > > > mcc0.mcc.uiuc.edu/mailman/listinfo/ohmms-dev > > > > I will probably ignore any e-mail about the code and documentation sent to > > me directly. Or, at least, I will answer to the mailing list not to the > > questioner. > > > > At 11:29 AM 7/23/2004 -0400, you wrote: > > >Hi Jeongnim, > > > > > >The restart file contains the lattice parameter after the first MD step, > > >not the last one and also not the original lattice parameter either! > > >That is strange. Do you understand this? > > > > > >Richard > > > > > > > > > > > > > > >On Fri, 23 Jul 2004, Jeongnim Kim wrote: > > > > > > > Can you check the restart.xml file? > > > > > > > > At 10:47 AM 7/23/2004 -0400, you wrote: > > > > >Hi Jeongnim, > > > > > > > > > >I am debugging the EPN propagator using your latest ohmms version 2.4. > > > > >Using the lattice.set function, I still do not get the correct nearest > > > > >neighbor list in my potential evaluation after the unit cell has > > > > >changed shape. Also, when I am looking at the output xyz file, the > > > > >lattice vectors never change. > > > > > > > > > >I added a few debug print statements to EPNVerlet.cpp and > > > > >Lattice/CrystalLattice.cpp to see how the variables change and > that the > > > > >lattice gets updated. From what I can tell the lattice gets set in > > > > >Lattice.set correctly but when it is written out into the xyz file it > > > > >still writes the original value. This old value also seems to be > used in > > > > >the NNEngine. > > > > > > > > > >Do you have any idea what I am doing wrong here. Thanks for your help. > > > > > > > > > >Cheers, > > > > > Richard > > > > > > > > Best wishes, > > > > Jeongnim Kim > > > > > > > > --------------------------------------------------------------- > > > > Jeongnim Kim e-mail: jnkim@uiuc.edu > > > > NCSA Tel: 217-244-6319 (NCSA) > > > > Beckman Institute 217-244-4971 (MCC) > > > > 405 N Mathews Av. Fax: 217-244-2909 > > > > Urbana, IL 61801 > > > > --------------------------------------------------------------- > > > > > > > > Best wishes, > > Jeongnim Kim > > > > --------------------------------------------------------------- > > Jeongnim Kim e-mail: jnkim@uiuc.edu > > NCSA Tel: 217-244-6319 (NCSA) > > Beckman Institute 217-244-4971 (MCC) > > 405 N Mathews Av. Fax: 217-244-2909 > > Urbana, IL 61801 > > --------------------------------------------------------------- > > Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From jnkim at ncsa.uiuc.edu Fri Jul 23 12:11:16 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Fri Jul 23 13:11:25 2004 Subject: [Ohmms-dev] Re: sprng on local linuces In-Reply-To: Message-ID: <5.1.0.14.2.20040723130548.029fadf8@pop.ncsa.uiuc.edu> It is correct. Something is not working with sprng. Even though autoconfig finds it, it does not link it correctly. Using sprng is the hardest one and I have not fully autoconfig-ed or cmake-ed it. A practical fix: add -lsprng to LIBS in src/OhmmsApp/Makefile or you can have LIBS="-lsprng" with the configuration. It is supposed to append. Will find a better fix soon. At 01:42 PM 7/23/2004 -0400, you wrote: >Jeongnim, > >I have succesfully compiled Ohmms-v2.4.0. However, to compile properly, I >must use --without-sprng. If I try to compile with sprng, I get errors >like "double_rng not found", etc. The installed sprng library on the >local linuces here is very old -- could this be the problem, or am I >doing something wrong. > >Here's the configure options I use to try to enable sprng: > >../configure --with-cxx=gcc --with-sprng=/usr/local >--with-sprng-include=/usr/local/include --with-sprng-libdir=/usr/local/lib > >To repeat, is the problem my configure options or is it a problem with the >installed libraries? > >Cheers, >~Kaden > >Office phone: (614) 292-2887 Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From jnkim at ncsa.uiuc.edu Fri Jul 23 13:28:26 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Fri Jul 23 14:28:31 2004 Subject: [Ohmms-dev] Re: getting out the forces from a run In-Reply-To: Message-ID: <5.1.0.14.2.20040723142720.02b75968@pop.ncsa.uiuc.edu> Check out src/ParticleTags.cpp or code documentation. force does not exist but "f" does. At 03:21 PM 7/23/2004 -0400, you wrote: > So I'm trying to modify the bulksi example to have it spit out the >forces as well as the atom positions; so I modified part of the xml to >look like this... > > > > > > > > > > > >...by adding the line. I got a file, but it's >all zeroes... which makes me think that I'm not using the right name for >the attribute. > > Is there a place in the documentation where the attributes of a particle >are listed, so that I could find what name I'm supposed to use to get >particle forces? I was probably looking at the wrong documentation on the >ohmms website. Thanks; --d Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From hazzard at pacific.mps.ohio-state.edu Fri Jul 23 13:35:35 2004 From: hazzard at pacific.mps.ohio-state.edu (Kaden Hazzard) Date: Fri Jul 23 14:41:05 2004 Subject: [Ohmms-dev] Re: sprng on local linuces In-Reply-To: <5.1.0.14.2.20040723130548.029fadf8@pop.ncsa.uiuc.edu> References: <5.1.0.14.2.20040723130548.029fadf8@pop.ncsa.uiuc.edu> Message-ID: Jeongnim, I have tried several variations on this theme: in the makefile, at the end of the configure command, etc.. No luck, even though -lsprng is there in many ways. Can you have a look at what I'm doing? /oahu/disk2/hazzard/inter/Ohmms-v2.4.0.DuTest2/linuxkadenbin Cheers, ~Kaden Office phone: (614) 292-2887 On Fri, 23 Jul 2004, Jeongnim Kim wrote: > It is correct. Something is not working with sprng. Even though autoconfig > finds it, it does not link it correctly. Using sprng is the hardest one and > I have not fully autoconfig-ed or cmake-ed it. > > A practical fix: > add -lsprng to LIBS in src/OhmmsApp/Makefile > or > > you can have LIBS="-lsprng" with the configuration. It is supposed to append. > > Will find a better fix soon. From jnkim at ncsa.uiuc.edu Fri Jul 23 14:37:18 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Fri Jul 23 15:37:23 2004 Subject: [Ohmms-dev] tips for configure Message-ID: <5.1.0.14.2.20040723152829.00ac14a0@pop.ncsa.uiuc.edu> I found out a very simple trick after many trials and errors with numerous libraries. This applies to any package that uses autoconf/automake. You can always set the values like LIBS CXXFLAGS to by-pass potential problems during configure. A typical problem is with sprng/hdf5 installed in the standard directories: /usr/local/include and /usr/local/lib (don't know why but I'm also very new with autoconf/automake). For example, this will get you somewhere: configure --with-cxx=gcc LIBS="-lsprng -lgmp" Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 --------------------------------------------------------------- From jnkim at ncsa.uiuc.edu Mon Jul 26 17:17:24 2004 From: jnkim at ncsa.uiuc.edu (Jeongnim Kim) Date: Mon Jul 26 18:17:29 2004 Subject: [Ohmms-dev] core package Message-ID: <5.1.0.14.2.20040726181406.02b824d0@pop.ncsa.uiuc.edu> Hi All, I have made a package called ohmmscore which contains only the base classes. It is used for abstraction penalty timing analysis and I plan to add more test codes. If you are interested in, check out http://www.mcc.uiuc.edu/ohmms/others/ohmmscore.html Best wishes, Jeongnim Kim --------------------------------------------------------------- Jeongnim Kim e-mail: jnkim@uiuc.edu NCSA Tel: 217-244-6319 (NCSA) Beckman Institute 217-244-4971 (MCC) 405 N Mathews Av. Fax: 217-244-2909 Urbana, IL 61801 ---------------------------------------------------------------