November 2010 - Comp-phys-alps-users

Core parallelization and GPU porting
by degliesposti 03 Dec '10

03 Dec '10

Dear ALPS users, I wonder if some of you is experienced and could give us recent insights about these two points: i) core parallelization with ALPS: which parallel environment is better (OpenMP, MPICH, ...) ii) are there any porting/tests of ALPS on GPU's ? Thanks for your time, Cristian Degli Esposti Boschi

3 2

Questions regarding DMRG
by Chong You Quan 03 Dec '10

03 Dec '10

Hi, I'm using ALPS 1.3.5 on Linux currently. I'm new to ALPS and DMRG and I'd just like to clarify a few questions:1. For full DMRG application with verbose = 1 using the Heisenberg (spin) model, what is the basis of the RHO's eigenvalues in the output? DMRG truncates the density matrix, so, the basis of the density matrix is that of the spin sites is it? For example, L = 2, then the basis would be |00>, |01>, |10>, |11>.. is it correct? What about the basis of the matrix after truncation? Also, if I would like to get the full wavefunction in the basis of the sites, what should I do? Also if I specify that N is conserved, what is the default value of N? Or does it try all possible N to give the lowest energy state?2. I can see that ALPS can be integrated into C++ but I tried googling for examples and documentation to no avail.. Is it in the Trac wiki? If so, must one register for it in order to view the documentation? The default documentation seems to specify what the input valu es represent but don't describe much of the output or how to integrate ALPS into C++.. Am I missing out on some way I can find the documentation for ALPS and how to integrate it in C++?3. I noticed there were some .dat files after running the DMRG application.. Are those temporary files stored during calculation or are those actual output files? I didn't seem to be able to extract any form of data from those files.. Is it that all the useful output for the DMRG is written to the screen?4. I'm not sure why but my parameter2xml does not seem to work on any of the example parameter files.. I'm currently just typing out the xml file by hand. Maybe I'm just missing out on something.. Does parameter2xml output to a specific directory or something?5. Almost all of the files on http://www.comp-phys.org/lugano04/Talks.html seem to be broken.. Is it that the links are outdated or something? I'm sorry if some of these questions sound too amateur but this seems to be the best place I can find an answer to these questions.Thanks in advance. Regards,YQ

2 3

Re: [ALPS-users] ALPS linux binary packages in repositories?
by Ryo IGARASHI 30 Nov '10

30 Nov '10

Dear Prof. Troyer, On Tue, Nov 30, 2010 at 2:59 PM, Matthias Troyer <troyer(a)phys.ethz.ch> wrote: > Regarding lp_solve, we are testing an ALPS replacement for it, so that user do not need to download it in the future. Is implementing own Linear Programming solver easier than updating to lp_solve 5.5? Since Debian/Ubuntu/RHEL all have official lp_solve package (5.5), updating should be enough from the packaging purpose. (See #68 on trac). Best regards, -- Ryo IGARASHI, Ph.D. rigarash(a)hosi.phys.s.u-tokyo.ac.jp OpenPGP fingerprint: BAD9 71E3 28F3 8952 5640 6A53 EC79 A280 6A19 2319

2 1

ALPS linux binary packages in repositories?
by Krunoslav Prsa 30 Nov '10

30 Nov '10

Dear ALPS developers, I have a simple question. Are ALPS binary packages available for download/automatic update through standard Ubuntu/Debian repositories? This appears to be the main argument of the clusters' system administrators against requests to install alps libraries and applications. They claim that it is not possible and that it involves extra work for them to update it manually (which they are not paid for). I also cannot seem to find it in packages.debian.org and packages.ubuntu.com. There are only some old binaries for Ubuntu 6.06 Dapper Drake. The manual installation of ALPS on a cluster without administration privileges is difficult. This needs to be said, although I understand that you are continuously doing steps to make it easier. I think that many end users would appreciate having a standardized way of installing and updating the packages through the repositories. Please note that all this is written bona fide (from a happy ALPS user) and assuming, of course, that this issue is not already resolved. Best wishes and thanks in advance, Kruno ------------------------------------------------- Dr. Krunoslav Prsa, Postdoc Dynamics of strongly correlated materials Lab. for Solid State Physics Building/Office: HPF/E15.1 Schafmattstrasse 16, ETH Zürich Hönggerberg, CH-8093 Switzerland. Office phone: +41 44 633 22 96 Lab phone: +41 44 633 23 01 Mobile phone: +41 76 386 17 99 -------------------------------------------------

3 3

some bugs on the ALPS web site
by Ruben Ghulghazaryan 25 Nov '10

25 Nov '10

Hi, Recently the ALPS web site was updated. It seems the web site http://alps.comp-phys.org/mediawiki/index.php/Tutorials:LatticesAndGraphs contains some language problems. I got some messages like the one below with Firefox 3.6.10 +++++++++ Invalid language. You need to specify a language like this: <source lang="html4strict">...</source> Supported languages for syntax highlighting: 4cs, 6502acme, 6502kickass, 6502tasm, 68000devpac, abap, actionscript, actionscript3, ada, algol68, apache, applescript, apt_sources, asm, asp, autoconf, autohotkey, autoit, avisynth, awk, bash, basic4gl, bf, bibtex, blitzbasic, bnf, boo, c, c_mac, caddcl, cadlisp, cfdg, cfm, chaiscript, cil, clojure, cmake, cobol, cpp, cpp-qt, csharp, css, cuesheet, d, dcs, delphi, diff, div, dos, dot, e, ecmascript, eiffel, email, erlang, f1, fo, fortran, freebasic, fsharp, gambas, gdb, genero, genie, gettext, glsl, gml, gnuplot, go, groovy, gwbasic, haskell, hicest, hq9plus, html4strict, icon, idl, ini, inno, intercal, io, j, java, java5, javascript, jquery, kixtart, klonec, klonecpp, latex, lb, lisp, locobasic, logtalk, lolcode, lotusformulas, lotusscript, lscript, lsl2, lua, m68k, magiksf, make, mapbasic, matlab, mirc, mmix, modula2, modula3, mpasm, mxml, mysql, newlisp, nsis, oberon2, objc, objeck, ocaml, ocaml-brief, oobas, oracle11, oracle8, oxygene, oz, pascal, pcre, per, perl, perl6, pf, php, php-brief, pic16, pike, pixelbender, plsql, postgresql, povray, powerbuilder, powershell, progress, prolog, properties, providex, purebasic, python, q, qbasic, rails, rebol, reg, robots, rpmspec, rsplus, ruby, sas, scala, scheme, scilab, sdlbasic, smalltalk, smarty, sql, systemverilog, tcl, teraterm, text, thinbasic, tsql, typoscript, unicon, vala, vb, vbnet, verilog, vhdl, vim, visualfoxpro, visualprolog, whitespace, whois, winbatch, xbasic, xml, xorg_conf, xpp, z80, zxbasic +++++++++ Regards, Ruben

2 1

ALPS 2.0.0rc2 binary packages for Debian (amd64)
by Ryo IGARASHI 25 Nov '10

25 Nov '10

Hi, I have created binary packages of ALPS 2.0.0rc2 for Debian GNU/Linux (primary for sid(unstable), but should work also on squeeze(testing)). The apt line is as follows: deb http://alps.comp-phys.org/static/software/debian/sid ./ deb-src http://alps.comp-phys.org/static/software/debian/sid ./ Currently, I have only compiled amd64 packages. I can run python-based tutorials with these binary packages[1][2], but if you find any bugs related to debian packaging, please open the trac ticket and assign to me(rigarash) or just report to the ML. [1] Since all binaries/scripts are installed system-wide, use 'python' instead of 'alpspython'. [2] This package does not contain dirloop-sse binary because of the incompatibility of lpsolve library. Best regards, -- Ryo IGARASHI, Ph.D. rigarash(a)hosi.phys.s.u-tokyo.ac.jp OpenPGP fingerprint: BAD9 71E3 28F3 8952 5640 6A53 EC79 A280 6A19 2319

1 0

ALPS 2.0 release candidate
by Matthias Troyer 25 Nov '10

25 Nov '10

Dear all, we finally have an ALPS 2 release candidate available on the Web. The MacOS X 10.5 build will take a bit longer but will also come soon. Please try to use is, install it, check your codes with it and report any bug to the mailing list. We are nearing completion of ALPS 2.0 Matthias Troyer

3 9

Re: [ALPS-users] Comp-phys-alps-users Digest, Vol 56, Issue 11
by Xiao Quan Yu 24 Nov '10

24 Nov '10

Dear ALPS community, Just informing you that we change the compiler to gnu4.4 with alps 1.3.5, it works well on the cluster. Thanks for your help. Best, Xiaoquan Send Comp-phys-alps-users mailing list submissions to > comp-phys-alps-users(a)lists.phys.ethz.ch > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.phys.ethz.ch/listinfo/comp-phys-alps-users > or, via email, send a message with subject or body 'help' to > comp-phys-alps-users-request(a)lists.phys.ethz.ch > > You can reach the person managing the list at > comp-phys-alps-users-owner(a)lists.phys.ethz.ch > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Comp-phys-alps-users digest..." > > > Today's Topics: > > 1. Re: asking for help (Fabien Alet) > 2. Re: extractxmgr problem (Ruben Ghulghazaryan) > 3. Re: extractxmgr problem (Matthias Troyer) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Mon, 15 Nov 2010 12:05:16 +0100 > From: Fabien Alet <alet(a)irsamc.ups-tlse.fr> > Subject: Re: [ALPS-users] asking for help > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Message-ID: <7F8AEC91-8B1B-4064-8E42-3F5C8F76B1C2(a)irsamc.ups-tlse.fr> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed; delsp=yes > > Hi, > > indeed I can reproduce the problem (located in alea/histogrameval.h) > with alps 1.3.5 and icc 11.0. > > It goes away with a more recent version of icc 11.1 (as Matthias > suggested). I used boost 1.42.0 but that's probably irrelevant. > > Alternatively, you can use alps 1.3.3 which goes fine with icc 11.0. > > Fabien > > Le 13 nov. 10 ? 19:00, Xiao Quan Yu a ?crit : > >> Dear ALPS community, >> >> We are trying to compile the alps 1.3.5 libraries on an intel >> cluster. We >> already installed it successfully on a PC and a laptop but now we have >> problems doing the installation on a cluster. The configuration >> seems to >> have gone OK. The line >> we have used is this: >> ./configure --with-boost=/scratch/warter/alps/boost_1_41_0 CC=icc >> CXX=icc >> --with-mpi-incdir=/opt/openmpi/1.3.4/intel-11.0/include/ >> >> >> And the summary is the following: >> Summary of configuration: >> ALPS version 1.3.5 >> prefix /u/he/warter/ALPS >> compiler type gnu [icc] >> optimization yes >> exceptions yes >> warnings no >> Boost version 1_41 >> MPI support yes >> OpenMP support no >> HDF5 support no >> LAPACK support yes >> SQLite support yes >> XML parser native >> Documentation no >> >> >> Then when I run make I am getting the following error: >> compilation aborted for ../../src/alps/alea/observableset.C (code 4) >> make[2]: *** [observableset.lo] Error 1 >> make[2]: Leaving directory `/scratch/warter/alps/alps-1.3.5/lib/alps' >> make[1]: *** [all-recursive] Error 1 >> make[1]: Leaving directory `/scratch/warter/alps/alps-1.3.5/lib' >> make: *** [all-recursive] Error 1 >> >> >> We really have no idea what this problem is. Could you please help >> us with >> this problem? >> >> Many thanks >> >> >> >> >> ---------------------------------------------------------------- >> SISSA Webmail https://webmail.sissa.it/ >> Powered by SquirrelMail http://www.squirrelmail.org/ >> >> >> >> ---------------------------------------------------------------- >> SISSA Webmail https://webmail.sissa.it/ >> Powered by SquirrelMail http://www.squirrelmail.org/ > > > > ------------------------------ > > Message: 2 > Date: Mon, 15 Nov 2010 05:44:00 -0800 (PST) > From: Ruben Ghulghazaryan <ghulr(a)yahoo.com> > Subject: Re: [ALPS-users] extractxmgr problem > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Message-ID: <358127.7561.qm(a)web30303.mail.mud.yahoo.com> > Content-Type: text/plain; charset=us-ascii > > Hi ALPS team, > > It seems I had missing xsltproc library. > > After installing this library the problem was fixed. > > Sorry for disturbing you. > > With best regards, > Ruben > > > > ----- Original Message ---- > From: Ruben Ghulghazaryan <ghulr(a)yahoo.com> > To: Comp-phys-alps-users(a)lists.phys.ethz.ch > Sent: Mon, November 15, 2010 2:28:01 PM > Subject: [ALPS-users] extractxmgr problem > > Hi, > > I have successfully installed ALSP 1.3.5 on Kubuntu 10.04. > I did not get any error nether during installation nor during "make test". > > When I try to follow test case for Monte Carlo simulation parm2a and > performed > simulation. Then I could not generate xmgr output file with > > > extractxmgr plot.xml parm2a.task*.out.xml > plot2a.xmgr > > It generates the following errors: > > exit: 19: Illegal number: -1 > exit: 19: Illegal number: -1 > exit: 19: Illegal number: -1 > > Would you please help me to understand what is wrong? > I have just followed the instruction according to web page: > > http://alps.comp-phys.org/mediawiki/index.php/Tutorial:ClassicalMCSimulatio… > > My plot.xml file content is as in the web site > > ++++++++++++++++ > <?xml version="1.0" encoding="UTF-8"?> > <plot name="Susceptibility versus temperature for one-dimensional > Heisenberg > models"> > > <legend show="true"/> > <xaxis label="Temperature" type="PARAMETER" name="T"/> > <yaxis label="Susceptibility" type="SCALAR_AVERAGE"/> > <set label="One-dimensional chain"/> > </plot> > +++++++++++++++ > > I got the same error with other extract* tools. > > With best regards, > Ruben > > > > > > ------------------------------ > > Message: 3 > Date: Tue, 16 Nov 2010 11:55:55 +0100 > From: Matthias Troyer <troyer(a)phys.ethz.ch> > Subject: Re: [ALPS-users] extractxmgr problem > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Message-ID: <CBBBA6A5-CA21-4C48-BB42-960FCBE4A043(a)phys.ethz.ch> > Content-Type: text/plain; charset=us-ascii > > I'm glad to hear that it works > > Matthias > > On Nov 15, 2010, at 2:44 PM, Ruben Ghulghazaryan wrote: > >> Hi ALPS team, >> >> It seems I had missing xsltproc library. >> >> After installing this library the problem was fixed. >> >> Sorry for disturbing you. >> >> With best regards, >> Ruben >> >> >> >> ----- Original Message ---- >> From: Ruben Ghulghazaryan <ghulr(a)yahoo.com> >> To: Comp-phys-alps-users(a)lists.phys.ethz.ch >> Sent: Mon, November 15, 2010 2:28:01 PM >> Subject: [ALPS-users] extractxmgr problem >> >> Hi, >> >> I have successfully installed ALSP 1.3.5 on Kubuntu 10.04. >> I did not get any error nether during installation nor during "make >> test". >> >> When I try to follow test case for Monte Carlo simulation parm2a and >> performed >> simulation. Then I could not generate xmgr output file with >> >> >> extractxmgr plot.xml parm2a.task*.out.xml > plot2a.xmgr >> >> It generates the following errors: >> >> exit: 19: Illegal number: -1 >> exit: 19: Illegal number: -1 >> exit: 19: Illegal number: -1 >> >> Would you please help me to understand what is wrong? >> I have just followed the instruction according to web page: >> >> http://alps.comp-phys.org/mediawiki/index.php/Tutorial:ClassicalMCSimulatio… >> >> My plot.xml file content is as in the web site >> >> ++++++++++++++++ >> <?xml version="1.0" encoding="UTF-8"?> >> <plot name="Susceptibility versus temperature for one-dimensional >> Heisenberg >> models"> >> >> <legend show="true"/> >> <xaxis label="Temperature" type="PARAMETER" name="T"/> >> <yaxis label="Susceptibility" type="SCALAR_AVERAGE"/> >> <set label="One-dimensional chain"/> >> </plot> >> +++++++++++++++ >> >> I got the same error with other extract* tools. >> >> With best regards, >> Ruben >> >> >> >> > > > > End of Comp-phys-alps-users Digest, Vol 56, Issue 11 > **************************************************** > ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by SquirrelMail http://www.squirrelmail.org/

1 0

"fulldiag": Can it handle magnetic field coupled to a non-conserved quantity?
by Itamar Kimchi 24 Nov '10

24 Nov '10

Hi all, I'm new to ALPS, but before I try to learn how to use it, I wanted to resolve an ambiguity in the documentation to figure out whether it's applicable to my problem. I want to find the magnetic susceptibility for a spin system, where the magnetic field h couples to Sz_tot as usual. However, I wish to use a Hamiltonian that does not commute with Sz_tot, hence does not leave it conserved. In the fulldiag "documentation" file http://alps.comp-phys.org/mediawiki/index.php/Documentation:FullDiagonaliza… there is a red warning claiming that I cannot use fulldiag if Sz_tot is not conserved. However, this documentation looks like it's two years old. The new tutorial, at http://alps.comp-phys.org/mediawiki/index.php/ALPS_2_Tutorials:ED-06_FullDi… has no such warning, but it only gives an example for calculating the susceptibility for a Heisenberg Hamiltonian, which conserves Sz_tot. Do you know if I can use fulldiag (or anything else) if my Sz_tot is not conserved? Thanks! Itamar

3 2

asking for help
by Xiao Quan Yu 23 Nov '10

23 Nov '10

Dear ALPS community, We are trying to use SSE to study 3D BoseGlass, our input file is the following: LATTICE_LIBRARY="lattices.xml"; LATTICE="inhomogeneous simple cubic lattice"; L=2; W=2; H=2; MODEL_LIBRARY="models.xml"; MODEL="hardcore boson"; t = 1.0; T = 0.2; SWEEPS=500000; THERMALIZATION=10000; {DISORDERSEED = 34275; mu=delta*2*(random()-0.5);} {delta=0.0} {delta=1.0} {delta=2.0} {delta=3.0} {delta=4.0} {delta=5.0} . The "parameter2xml input" step is fine, but after doing "dirloop_sse ", it says"too many positional options". We do know how to solve this problem. We also tried the example of 2D BoseGlass model in the ALPS wiki webpage, the code of worm algorithm is LATTICE_LIBRARY="../lattices.xml"; LATTICE="inhomogeneous square lattice"; L=4; MODEL_LIBRARY="../models.xml"; MODEL="boson Hubbard"; NONLOCAL=0; U = 1.0; Nmax = 2; t = 1.0; T = 0.1; delta = 5.0; SWEEPS=500000; THERMALIZATION=10000; { DISORDERSEED = 34275; mu=delta*2*(random()-0.5); } { DISORDERSEED = 49802; mu=delta*2*(random()-0.5); } { DISORDERSEED = 82529; mu=delta*2*(random()-0.5); } , we just use the same code but for SSE, and meet the same "too many positional options" problem. It seems one should modify the input file, but we don't know how. Thank you in advance. Best, Xiaoquan ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by SquirrelMail http://www.squirrelmail.org/

2 1