November 2011 - Comp-phys-alps-users

ALPS 2.0.1 released
by Matthias Troyer 19 Sep '12

19 Sep '12

Dear ALPS users, ALPS 2.0.1 has been released. This is the version included with the ALPS paper: B. Bauer et al., J. Stat. Mech. (2011) P05001, which is available at : http://www.iop.org/EJ/abstract/1742-5468/2011/05/P05001 .Besides references to this paper, this release contains many minor bug fixes. The MacOS X 10.5 binary release will be uploaded shortly. The others are ready for download. Best regards Matthias Troyer

4 33

Problem with large graph
by Volodymyr Derzhko 26 Dec '11

26 Dec '11

Dear Mr. Troyer, I have another problem with a graph, which consists of 32 sites. fulldiag works perfectly with smaller graphs (8 and 18 sites) but won't with 32-sites one. I tried it under linux (command line, Vistrails) and Windows(Vistrails). There is no error, but it can't diagonalize it (time of diagonalizing is longer than several hours; I had to interrupt the process). Note, that I'm taking {N_total=1; Sz_total=1/2;}, and it's hard to believe, that it can take hours and hours to diagonalize it. The lattice, model, and tasks files are attached. Please, let me know what can be wrong. Probably, there is a limitation for big lattice size? Thank you in advance, Volodymyr Derzhko

2 8

Re: [ALPS-users] Comp-phys-alps-users Digest, Vol 68, Issue 11
by Suddhasattwa Ghosh 02 Dec '11

02 Dec '11

Dear Sir, Thank you very much for the mail. The possible algorithm for generating random structures is 1. Begin with known correlations of ideal crystal 2. Determine good structures to represent ideal crystal 3. Convert lattice into Ising-like model 4. Replace atoms {A,B} with spins, Sigma={-1,+1} 5. Consider a set of “figures” .Define local correlation function as the product of spins 6. Calculate lattice average by summing over symmetrically equivalent figures over all sites Goal: Generate small unit cell that reproduces the correlations found in random alloys using a MC code Thank you Suddhasattwa Ghosh On Sun, Nov 27, 2011 at 4:30 PM, < comp-phys-alps-users-request(a)lists.phys.ethz.ch> wrote: > 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: Suggestions on ALPS Simulations (Matthias Troyer) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sat, 26 Nov 2011 17:22:38 +0100 > From: Matthias Troyer <troyer(a)phys.ethz.ch> > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Subject: Re: [ALPS-users] Suggestions on ALPS Simulations > Message-ID: <D69E3090-F741-4BBA-A182-89E32703FA79(a)phys.ethz.ch> > Content-Type: text/plain; charset="iso-8859-1" > > Hi, > > You need to tell us which model you want to simulate. Certainly ALPS can > help you write a MC code for your problem, but I don't think that a MC code > for your specific problem exists. > > Matthias > > On 23 Nov 2011, at 12:09, Suddhasattwa Ghosh wrote: > > > Dear Sir, > > > > We would like to address the following problems using the ALPS package. > > > > > > > > May I kindly request if you could inform us if these problems could be > done using this package? > > > > > > > > Problem 1 > > > > I consider a bcc phase with 3.64 A as lattice parameter. We assume A, B > or C (could be Ni,Ti and Nb) can either occupy (0,0,0) position. > > > > I also assume X(A)=0.40 X(B)=0.40 X(C)=0.20 as composition variables > > > > Now, I consider a 40 atom cell which can mimic the most random alloy > with this composition. > > > > Can I use your code to get the most random alloy for A, B and C. That is > to say, getting the structure of this alloy with a 40 atom cell satisfying > the > > > > compositions using MC simulations at T=298.15 K > > > > Can we add any number of elements if we are to construct a disordered > alloy of say, 5 elements? Can it be done? > > > > > > Problem 2 > > > > > > I consider a rhombohedral phase with 15-16 A approx. as lattice > parameters. I have five different inequivalent positions which can either > be occupied by A or B (say Ni and Fe) > > > > I assume composition variables as X(A)=0.60 X(B)=0.40 > > > > Can in this case a 40 atom cell be used to generate the most random > structure using MC simulations at T=298.15 K > > > > Could you please suggest on these issues Sir and inform if ALPS could be > used for these cases. > > > > > > With regards > > > > Suddhasattwa Ghosh > > > >

2 1

matrixelement=?
by Mateusz Łącki 30 Nov '11

30 Nov '11

Dear All, I have found here: http://alps.comp-phys.org/mediawiki/index.php/Tutorials:MeasureHOWTO a way to measure an observable (while working with QMC code) which is related to the basis in which we measure by some simple formula: sqrt (N). But i still have some question regarding implementing a non-standard situation. I would not like to go into really ugly details, so I have created a self contained example: I understood that: symbols "N" that appearing in that example are labels of quantum numbers defining the basis: if basis was enumerated by a few quantum numbers then the config should look like this: <SITEBASIS name="boson"> <QUANTUMNUMBER name="S" min="0" max="b1"/> <QUANTUMNUMBER name="C" min="0" max="b2"/> <OPERATOR name="Milka" matrixelement="1"> <CHANGE quantumnumber="S" change="-1"/> <CHANGE quantumnumber="C" change="1"/> </OPERATOR> </SITEBASIS> In this case operator "Milka" creates one bar of chocolate using one unit of sugar? Another issue is what makes a correct expression? I suppose that parsing of these things is done by "expression" binary. Which functions are supporter by this parser? sqrt(N) is supported but BesselJ(N) may not be, I also saw different examples of random number generators - matrixelement="random_gaussian(1)" (or sth like that). The other question: if matrix elements are not constant, a simple function but "conditional" in nature: Milka_{(S1,C1), (S2,C2)} = 1 is \lambda(S1,C1,S2,C2) is true and -7 otherwise then I suppose: <SITEBASIS name="boson"> <QUANTUMNUMBER name="S" min="0" max="b1"/> <QUANTUMNUMBER name="C" min="0" max="b2"/> <OPERATOR name="Milka" matrixelement=" f(S1,C1,S2,C2)-7*(1-f(S1,C1,S2,C2))"> <CHANGE ...> </OPERATOR> </SITEBASIS> will do provided f evaluates to 1,0 iff \lambda is true,false for any C1,C2,S2,S1. Is there a way to introduce "if" construct inside the matrixelement? Regards, Mateusz Lacki

2 1

FindLapack.cmake - proposed change
by Mateusz Łącki 29 Nov '11

29 Nov '11

Dear All, I heve succesfull debugged cmake's failure to properly set up my mkl. I use the newest Intel suite there is, and I had lots of errors when trying to link with it. In config/FindLapack.cmake there was: SET(MKL_PATHS $ENV{MKL_HOME}/lib/em${QMC_BITS}t $ENV{MKL_HOME}/lib/${QMC_BITS} $ENV{MKL_HOME}/lib ${MKL_PATHS} ) after changing to: $ENV{MKL_HOME}/lib/intel64${QMC_BITS} $ENV{MKL_HOME}/lib/em${QMC_BITS}t $ENV{MKL_HOME}/lib/${QMC_BITS} $ENV{MKL_HOME}/lib ${MKL_PATHS} ) as well as: STRING(REGEX MATCHALL "[-][L]([^ ;])+" MKL_PATH_WITH_PREFIX "$ENV{MKL}") should be modified to make it possible to build successfully when MKL does not contain "-L" just before the path First change is needed because new mkl versions contain libs in $MKL/lib/intel64 (here $MKL also does not include -L) Regards, Mateusz Lacki

2 1

alps-2.0.2 compile error
by Tadeusz Wasiutyński 29 Nov '11

29 Nov '11

Dears, After successful installation of alps-2.0.2 on my desktop (Scientific Linux 6.1) I moved to the cluster with RedHat5(el5) with gcc 4.1.2 and python 2.4.3 . Make produced error: Linking CXX executable compactrun ../src/boost/libboost.so: undefined reference to `pthread_create' ../src/boost/libboost.so: undefined reference to `pthread_key_create' ../src/boost/libboost.so: undefined reference to `pthread_getspecific' ../src/boost/libboost.so: undefined reference to `pthread_once' ../src/boost/libboost.so: undefined reference to `pthread_join' ../src/boost/libboost.so: undefined reference to `pthread_setspecific' ../src/boost/libboost.so: undefined reference to `pthread_detach' collect2: ld returned 1 exit status make[2]: *** [tool/compactrun] Error 1 make[1]: *** [tool/CMakeFiles/compactrun.dir/all] Error 2 make: *** [all] Error 2 Is the above to the old compiler or there something else? Regards Tadeusz

3 2

compiling 2.0.2: libboost_mpi missing?
by Mateusz Łącki 29 Nov '11

29 Nov '11

Dear All, I am still trying to compile alps using icc. I have run to a new error: lacki@scud:~/alps/alps-2.0.2-r5790-src-with-boost/alps-2.0.2-r5790-src-with-boost/alps $ make xmlparser [100%] Built target alps Linking CXX executable xmlparser ../../src/alps/libalps.so.2.0.2: undefined reference to `void boost::mpi::broadcast<boost::mpi::packed_iarchive>(boost::mpi::communicator const&, boost::mpi::packed_iarchive&, int)' ../../src/alps/libalps.so.2.0.2: undefined reference to `boost::mpi::exception::exception(char const*, int)' ../../src/alps/libalps.so.2.0.2: undefined reference to `boost::mpi::communicator::barrier() const' (............) ../../src/alps/libalps.so.2.0.2: undefined reference to `boost::mpi::communicator::operator ompi_communicator_t*() const' ../../src/alps/libalps.so.2.0.2: undefined reference to `boost::mpi::environment::~environment()' ../../src/alps/libalps.so.2.0.2: undefined reference to `boost::mpi::communicator::size() const' ../../src/alps/libalps.so.2.0.2: undefined reference to `void boost::mpi::communicator::send<boost::mpi::packed_oarchive>(int, int, boost::mpi::packed_oarchive const&) const' make[3]: *** [test/parser/xmlparser] Error 1 make[2]: *** [test/parser/CMakeFiles/xmlparser.dir/all] Error 2 make[1]: *** [test/parser/CMakeFiles/xmlparser.dir/rule] Error 2 while trying to do "make". The exact make subcommand it corresponds to is "make xmlparser" I enclose my config files. here is a quotation from: alps/config/FindBoostForALPS.cmake # We do not require Boost.MPI, therefore check whether Boost.MPI exists # before actual find_package for Boost. # - Ubuntu 10.10 does not have Boost.MPI package. find_package(Boost 1.43.0 COMPONENTS mpi) if (Boost_FOUND) it seem contradictory with previous message. Also when i try to set boost_mpi directory to /usr/lib64/openmpi/lib where there is libboost_mpi.so.1.46.0 pressing "c" for doing "configure" ereases the path that i have just put in - why is that ? The same linking error appears if I set the path to boost_root_dir correctly - in that case manual specification of location of binary libboost_mpi is no longer an option Regards, Mateusz

3 22

Suggestions on ALPS Simulations
by Suddhasattwa Ghosh 26 Nov '11

26 Nov '11

Dear Sir, We would like to address the following problems using the ALPS package. * * *May I kindly request if you could inform us if these problems could be done using this package? * * * Problem 1 I consider a bcc phase with 3.64 A as lattice parameter. We assume A, B or C (could be Ni,Ti and Nb) *can either occupy* (0,0,0) position. I also assume X(A)=0.40 X(B)=0.40 X(C)=0.20 as composition variables Now, I consider a 40 atom cell which can mimic the most random alloy with this composition. Can I use your code to get the most random alloy for A, B and C. That is to say, getting the structure of this alloy with a 40 atom cell satisfying the compositions using MC simulations at T=298.15 K Can we add any number of elements if we are to construct a disordered alloy of say, 5 elements? Can it be done? Problem 2 I consider a rhombohedral phase with 15-16 A approx. as lattice parameters. I have *five* different inequivalent positions which can *either be occupied * by A or B (say Ni and Fe) I assume composition variables as X(A)=0.60 X(B)=0.40 Can in this case a 40 atom cell be used to generate the most random structure using MC simulations at T=298.15 K Could you please suggest on these issues Sir and inform if ALPS could be used for these cases. With regards Suddhasattwa Ghosh

2 1

Re: [ALPS-users] Comp-phys-alps-users Digest, Vol 68, Issue 9
by Tadeusz Wasiutyński 25 Nov '11

25 Nov '11

Dear Mathias, thank you very much for the hint. Indeed when 64 bit is off, make goes smoothly and make test passed all but 97,98 and 99. I was misled by the package lapack-3.2.1-4.el6(x86_64) which went to /usr/lib64/. By the way my system is Scientific Linux 6.1. Best regards Tadeusz 2011/11/25 <comp-phys-alps-users-request(a)lists.phys.ethz.ch> > 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. alps-2.0.2 compile error (Tadeusz Wasiuty?ski) > 2. Re: alps-2.0.2 compile error (Mateusz ??cki) > 3. Re: alps-2.0.2 compile error (Matthias Troyer) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 25 Nov 2011 10:07:52 +0100 > From: Tadeusz Wasiuty?ski <tadeusz.wasiutynski(a)ifj.edu.pl> > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Subject: [ALPS-users] alps-2.0.2 compile error > Message-ID: > <CAKNmYXinqnEPFuFv0X32Z=rrvwhU556W1aaOiJwEeAgiNgaoKQ(a)mail.gmail.com > > > Content-Type: text/plain; charset="iso-8859-1" > > Dears, > I had compiled successfully 2.0.1 version and made running with cmake as > follows: > > /home/wasiutyn/opt/bin/cmake -D BLAS_LIBRARY=/usr/lib64/libblas.so.3.2.1 > -D LAPACK_LIBRARY=/usr/lib64/liblapack.so.3.2.1 -D > LPSolve_LIBRARIES=/usr/lib64/liblpsolve55.so -D > LPSolve_INCLUDE_DIR=/usr/include/ -D CMAKE_INSTALL_PREFIX=$HOME/ALPS -D > LAPACK_64_BIT=ON -D HDF5_LIBRARIES=/home/wasiutyn/opt/lib/libhdf5.so.6 -D > HDF5_INCLUDE_DIR=/home/wasiutyn/opt/include/ > ~/alps-2.0.1-r5565-src-with-boost/alps/ > > However with about the same cmake I got: > > -- CMAKE_GENERATOR: Unix Makefiles > -- CMAKE_CL_64: > -- CMAKE_SIZEOF_VOID_P: > -- CMAKE_GENERATOR: Unix Makefiles > -- CMAKE_CL_64: > -- Build type: Release > -- Python interpreter /usr/bin/python > -- Python interpreter ok : version 2.6.6 > -- PYTHON_INCLUDE_DIRS = /usr/include/python2.6 > -- PYTHON_NUMPY_INCLUDE_DIR = > /usr/lib64/python2.6/site-packages/numpy/core/include > -- PYTHON_SITE_PKG = /usr/lib/python2.6/site-packages > -- PYTHON_LIBRARY = /usr/lib64/python2.6/config/libpython2.6.so > -- PYTHON_EXTRA_LIBS =-lpthread -ldl -lutil > -- PYTHON_LINK_FOR_SHARED = -Xlinker -export-dynamic > -- ALPS version: 2.0.2 > -- Looking for Boost Source > -- Found Boost Source: /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/boost > -- Boost Version: 1_47_0 > -- Could NOT find MPI (missing: MPI_LIBRARY MPI_INCLUDE_PATH) > -- MPI compiler was MPI_COMPILER-NOTFOUND > -- Could NOT find SZIP (missing: SZIP_LIBRARIES SZIP_INCLUDE_DIRS) > -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ > -- Python interpreter /usr/bin/python > -- Python interpreter ok : version 2.6.6 > -- PYTHON_INCLUDE_DIRS = /usr/include/python2.6 > -- PYTHON_NUMPY_INCLUDE_DIR = > /usr/lib64/python2.6/site-packages/numpy/core/include > -- PYTHON_SITE_PKG = /usr/lib/python2.6/site-packages > -- PYTHON_LIBRARY = /usr/lib64/python2.6/config/libpython2.6.so > -- PYTHON_EXTRA_LIBS =-lpthread -ldl -lutil > -- PYTHON_LINK_FOR_SHARED = -Xlinker -export-dynamic > -- Numpy include in /usr/lib64/python2.6/site-packages/numpy/core/include > -- ALPS XML dir is /home/wasiutyn/ALPS2/lib/xml > -- Adding Boost dir: /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/boost > -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ > -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ > -- tebd will not be built > -- Configuring done > -- Generating done > -- Build files have been written to: > /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/alps > > during make I receive: > > make[2]: *** > [example/ietl/CMakeFiles/jacobidavidson2.dir/jacobidavidson2.cpp.o] Error 1 > make[1]: *** [example/ietl/CMakeFiles/jacobidavidson2.dir/all] Error 2 > make: *** [all] Error 2 > > The only differece I see in cmake output is Boost Version: now 1_47_0 was > before 1_46_1 > > The same error was already reported earlier by Mateusz Lacki. What is the > solution? > Regards > Tadeusz >

1 0

alps-2.0.2 compile error
by Tadeusz Wasiutyński 25 Nov '11

25 Nov '11

Dears, I had compiled successfully 2.0.1 version and made running with cmake as follows: /home/wasiutyn/opt/bin/cmake -D BLAS_LIBRARY=/usr/lib64/libblas.so.3.2.1 -D LAPACK_LIBRARY=/usr/lib64/liblapack.so.3.2.1 -D LPSolve_LIBRARIES=/usr/lib64/liblpsolve55.so -D LPSolve_INCLUDE_DIR=/usr/include/ -D CMAKE_INSTALL_PREFIX=$HOME/ALPS -D LAPACK_64_BIT=ON -D HDF5_LIBRARIES=/home/wasiutyn/opt/lib/libhdf5.so.6 -D HDF5_INCLUDE_DIR=/home/wasiutyn/opt/include/ ~/alps-2.0.1-r5565-src-with-boost/alps/ However with about the same cmake I got: -- CMAKE_GENERATOR: Unix Makefiles -- CMAKE_CL_64: -- CMAKE_SIZEOF_VOID_P: -- CMAKE_GENERATOR: Unix Makefiles -- CMAKE_CL_64: -- Build type: Release -- Python interpreter /usr/bin/python -- Python interpreter ok : version 2.6.6 -- PYTHON_INCLUDE_DIRS = /usr/include/python2.6 -- PYTHON_NUMPY_INCLUDE_DIR = /usr/lib64/python2.6/site-packages/numpy/core/include -- PYTHON_SITE_PKG = /usr/lib/python2.6/site-packages -- PYTHON_LIBRARY = /usr/lib64/python2.6/config/libpython2.6.so -- PYTHON_EXTRA_LIBS =-lpthread -ldl -lutil -- PYTHON_LINK_FOR_SHARED = -Xlinker -export-dynamic -- ALPS version: 2.0.2 -- Looking for Boost Source -- Found Boost Source: /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/boost -- Boost Version: 1_47_0 -- Could NOT find MPI (missing: MPI_LIBRARY MPI_INCLUDE_PATH) -- MPI compiler was MPI_COMPILER-NOTFOUND -- Could NOT find SZIP (missing: SZIP_LIBRARIES SZIP_INCLUDE_DIRS) -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ -- Python interpreter /usr/bin/python -- Python interpreter ok : version 2.6.6 -- PYTHON_INCLUDE_DIRS = /usr/include/python2.6 -- PYTHON_NUMPY_INCLUDE_DIR = /usr/lib64/python2.6/site-packages/numpy/core/include -- PYTHON_SITE_PKG = /usr/lib/python2.6/site-packages -- PYTHON_LIBRARY = /usr/lib64/python2.6/config/libpython2.6.so -- PYTHON_EXTRA_LIBS =-lpthread -ldl -lutil -- PYTHON_LINK_FOR_SHARED = -Xlinker -export-dynamic -- Numpy include in /usr/lib64/python2.6/site-packages/numpy/core/include -- ALPS XML dir is /home/wasiutyn/ALPS2/lib/xml -- Adding Boost dir: /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/boost -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ -- HDF5 version found is 1.8.6 in /home/wasiutyn/opt/include/ -- tebd will not be built -- Configuring done -- Generating done -- Build files have been written to: /home/wasiutyn/alps-2.0.2-r5790-src-with-boost/alps during make I receive: make[2]: *** [example/ietl/CMakeFiles/jacobidavidson2.dir/jacobidavidson2.cpp.o] Error 1 make[1]: *** [example/ietl/CMakeFiles/jacobidavidson2.dir/all] Error 2 make: *** [all] Error 2 The only differece I see in cmake output is Boost Version: now 1_47_0 was before 1_46_1 The same error was already reported earlier by Mateusz Lacki. What is the solution? Regards Tadeusz

3 2