- Comp-phys-alps-users - lists.phys.ethz.ch

Re: [ALPS-users] FW: error at installation for pyalea_c.so
by Munehisa Matsumoto 23 Jan '11

23 Jan '11

Dear Benoît, Thank you for sending the file "cmake_install.cmake" which was taken from lib/pyalps under your build directory as I understood. It looks like your build directory is the same as your installation directory. Is that the case? In line 41 of the file your install destination reads ${CMAKE_INSTALL_PREFIX}/lib/pyalps with ${CMAKE_INSTALL_PREFIX} being "/home/gremaud/alps/alps-2.0.0-r5363" as defined in line 5, while from the build directory "pyalea_c.so" is referenced as "/home/gremaud/alps/alps-2.0.0-r5363/lib/pyalps/pyalea_c.so" I guess your installation moves the file "pyalea.so" from your build directory to your installation directory and thus the file "pyalea.so" is deleted after this stage of installation is supposed to be passed. If the build directory and the installation directory is the same, it might happen that a target file is copied onto itself and then deleted. To avoid such complication having a separate installation directory might help. At least now I am reproducing the problem by setting the build directory and the installation directory to be the same. best regards, Munehisa

2 1

Re: [ALPS-users] FW: error at installation for pyalea_c.so
by Munehisa Matsumoto 22 Jan '11

22 Jan '11

Dear Kruno and Benoît, I finally reproduced the problem by setting the following in my initial cmake command -DCMAKE_INSTALL_PREFIX=~/.... instead of -DCMAKE_INSTALL_PREFIX=${HOME}/... which I usually do. It might help to do "${HOME}/" instead of "~/" in your cmake. best regards, Munehisa Matsumoto Munehisa MATSUMOTO Department of Physics University of California, One Shields Avenue, Davis, CA 95616 E-mail: matsumoto(a)physics.ucdavis.edu matumott(a)mac.com matumott(a)gmail.com On 2011/01/19, at 3:02, Krunoslav Prsa wrote: Hi, I get the same problem when installing on the brutus clustur... Best wishes, Kruno ------------------------------------------------- Dr. Krunoslav Prsa, Postdoc Dynamics of strongly correlated materials Laboratory 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 ------------------------------------------------- -----Original Message----- From: Prsa Krunoslav [mailto:Krunoslav.Prsa@psi.ch] Sent: Wednesday, January 19, 2011 8:50 AM To: kprsa(a)phys.ethz.ch Subject: FW: [ALPS-users] error at installation for pyalea_c.so ------------------------------------------- From: comp-phys-alps-users-bounces(a)lists.phys.ethz.ch on behalf of benoît grémaud[SMTP:GREMAUD.BENOIT@GMAIL.COM <SMTP%3AGREMAUD.BENOIT(a)GMAIL.COM>] Sent: Wednesday, January 19, 2011 8:50:21 AM To: comp-phys-alps-users(a)phys.ethz.ch Subject: [ALPS-users] error at installation for pyalea_c.so Auto forwarded by a Rule Hello, I am compiling alps2.0.0-- -r5363 on a linux machine (fedora) with intel compilers and mkl I use boost provided by alps. python version is 2.6.4 the cmake command is as follows: cmake -D Boost_ROOT_DIR:PATH=~/alps/alps-2.0.0-r5363-src-with-boost/boost -DCMAKE_INSTALL_PREFIX=~/alps/alps-2.0.0-r5363 -D LAPACK_64_BIT=ON -DLPSOLVE_ROOT=~/alps/lp_solve_4.0 ~/alps/alps-2.0.0-r5363-src-with-boost/alps cmake, make, make test, all went fine. however, I got the following error during the install: CMake Error at lib/pyalps/cmake_install.cmake:41 (FILE): file INSTALL cannot find "/home/gremaud/alps/alps-2.0.0-r5363/lib/pyalps/pyalea_c.so". Call Stack (most recent call first): cmake_install.cmake:94 (INCLUDE) I checked that before the installation the file pyalea_c.so existed. Strangely enough, after the preceding error, it has disappeared. any clue of what happened? regards, Benoît Grémaud -------------------- Researcher at CNRS, Laboratoire Kastler Brossel (Paris, France) Visiting Associate Professor at the Centre for Quantum Technologies (NUS, Singapore)

3 5

Dzyaloshinskii-Moriya in spinmc
by Francesco Casola 21 Jan '11

21 Jan '11

Dear ALPS community, as you know classical Monte Carlo in ALPS does not make use of the XML model descriptions, contrary to the quantum case. On the other hand, I would be intereseted in simulating a Heisenberg model including a Dzyaloshinskii-Moriya interaction. I could not find any comment about this possibility on the page describing spinmc : http://alps.comp-phys.org/mediawiki/index.php/Documentation:ClassicalMCSimu…. - Would it be possible to implement such a Hamiltonian? -Secondly, from that same tutorial page it is not clear to me whether the field can be specified as a 3D vector or it is intended as a scalar in the default z-direction. Many thanks for your kind help. Francesco -- Francesco Casola, Ph. D. Student Laboratory for Solid State Physics ETH Zürich Hönggerberg HPF E15.1 Schafmattstrasse 16 CH-8093 Switzerland Tel. Office: +41 (0)44 633 22 96 / 03 Tel. Lab: +41 (0)44 633 23 01 E-mail: fcasola(a)phys.ethz.ch Group website: https://protos.ethz.ch/dscm

1 0

Re: [ALPS-users] How to run Spin Glass systems
by Hakob Abajyan 21 Jan '11

21 Jan '11

Doesn't anyone have answer ? ________________________________ From: Hakob Abajyan <hakob.abajyan(a)yahoo.com> To: comp-phys-alps-devel(a)lists.phys.ethz.ch; comp-phys-alps-users(a)lists.phys.ethz.ch Sent: Wed, January 19, 2011 2:03:54 AM Subject: Re: How to run Spin Glass systems Any help would be appreciated. ________________________________ From: Hakob Abajyan <hakob.abajyan(a)yahoo.com> To: comp-phys-alps-devel(a)lists.phys.ethz.ch; comp-phys-alps-users(a)lists.phys.ethz.ch Sent: Sat, January 15, 2011 5:23:02 AM Subject: How to run Spin Glass systems ________________________________ Hi ALPS Team, I have investigated tutorials on ALPS web page but could not find any instructions how to run "classical Heisenberg spin-glass model" on ALPS. Could you please, answer to the following questions ? 1. Does ALPS support "classical Heisenberg spin-glass model" (Interaction constants are generated randomly) ? 2. If the answer of first question is positive, how to do that ? 3. Does ALPS support ensemble simulation or just simulation of one spin chain?Thanks in advance, Hakob

2 2

problem with alps - python
by Carlos Pineda 20 Jan '11

20 Jan '11

Dear friends, I am trying to install ALPS and my interest is to use it in combination with python, which I am still learning. The installation has so far been more or less easy. However I am stuck with a problem with python. Possibly its only python I downloaded alps-2.0.0-r5363-src-with-boost.tar.gz and followed the instructions in http://alps.comp-phys.org/mediawiki/index.php/Download_and_install_ALPS_2#B… for a generic distribution, until ******************************************** make test ******************************************** where some errors are reported: ******************************************** $ make test ... The following tests FAILED: 1 - ising_single (Failed) 2 - loop_single (Failed) 3 - exchange (Failed) 90 - python_hlist_test (Failed) 91 - python_mcdata (Failed) 92 - python_pyhdf5 (Failed) Errors while running CTest make: *** [test] Error 8 ******************************************** Some of them are trivial, I think. Namely, for the Ising model instead of having an error of exactly 0, the error is of the order of 10^-8. However for the python part it is mode problematic. I managed to trace back the errors to the command ******************************************** carlosp@papa ~/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps $ /usr/bin/ctest --force-new-ctest-process -V UpdateCTestConfiguration from :/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps/DartConfiguration.tcl UpdateCTestConfiguration from :/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps/DartConfiguration.tcl Test project /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps Constructing a list of tests Done constructing a list of tests Checking test dependency graph... test 1 Start 1: python_hlist_test 1: Test command: /usr/bin/cmake -Dcmd=hlist_test.py -Dinput=hlist_test -Doutput=hlist_test -Dsourcedir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dbinarydir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dcmddir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps -P /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake 1: Test timeout computed to be: 9.99988e+06 1: CMake Error at /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake:33 (message): 1: error runing test 'python_hlist_test.py': Traceback (most recent call 1: last): 1: 1: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps/hlist_test.py", line 28, in <module> 1: from pyalps.hlist import HList 1: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/__init__.py", line 38, in <module> 1: from tools import * 1: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/tools.py", line 31, in <module> 1: import tempfile 1: File "/usr/lib/python2.6/tempfile.py", line 34, in <module> 1: from random import Random as _Random 1: File "/usr/lib/python2.6/random.py", line 45, in <module> 1: from math import log as _log, exp as _exp, pi as _pi, e as _e, ceil as _ceil 1: 1: ImportError: cannot import name pi 1: 1: ; shell output: 1! 1: 1: 1/3 Test #1: python_hlist_test ................***Failed 0.09 sec test 2 Start 2: python_mcdata 2: Test command: /usr/bin/cmake -Dcmd=mcdata.py -Dinput=mcdata -Doutput=mcdata -Dsourcedir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dbinarydir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dcmddir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps -P /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake 2: Test timeout computed to be: 9.99988e+06 2: CMake Error at /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake:33 (message): 2: error runing test 'python_mcdata.py': Traceback (most recent call last): 2: 2: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps/mcdata.py", line 30, in <module> 2: from pyalps.alea import * 2: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/__init__.py", line 38, in <module> 2: from tools import * 2: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/tools.py", line 31, in <module> 2: import tempfile 2: File "/usr/lib/python2.6/tempfile.py", line 34, in <module> 2: from random import Random as _Random 2: File "/usr/lib/python2.6/random.py", line 45, in <module> 2: from math import log as _log, exp as _exp, pi as _pi, e as _e, ceil as _ceil 2: 2: ImportError: cannot import name pi 2: 2: ; shell output: 1! 2: 2: 2/3 Test #2: python_mcdata ....................***Failed 0.09 sec test 3 Start 3: python_pyhdf5 3: Test command: /usr/bin/cmake -Dcmd=pyhdf5.py -Dinput=pyhdf5 -Doutput=pyhdf5 -Dsourcedir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dbinarydir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps -Dcmddir=/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps -P /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake 3: Test timeout computed to be: 9.99988e+06 3: CMake Error at /home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/config/run_python_test.cmake:33 (message): 3: error runing test 'python_pyhdf5.py': Traceback (most recent call last): 3: 3: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps/pyhdf5.py", line 30, in <module> 3: import pyalps.hdf5 as h5 3: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/__init__.py", line 38, in <module> 3: from tools import * 3: File "/home/carlosp/Desktop/alps-2.0.0-r5363-src-with-boost/alps/lib/pyalps/tools.py", line 31, in <module> 3: import tempfile 3: File "/usr/lib/python2.6/tempfile.py", line 34, in <module> 3: from random import Random as _Random 3: File "/usr/lib/python2.6/random.py", line 45, in <module> 3: from math import log as _log, exp as _exp, pi as _pi, e as _e, ceil as _ceil 3: 3: ImportError: cannot import name pi 3: 3: ; shell output: 1! 3: 3: 3/3 Test #3: python_pyhdf5 ....................***Failed 0.09 sec 0% tests passed, 3 tests failed out of 3 Total Test time (real) = 0.28 sec The following tests FAILED: 1 - python_hlist_test (Failed) 2 - python_mcdata (Failed) 3 - python_pyhdf5 (Failed) Errors while running CTest ******************************************************************** however the "bad" line seems to be working in normal python: ******************************************************************** carlosp@papa ~/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps $ python Python 2.6.6 (r266:84292, Jan 19 2011, 11:02:33) [GCC 4.4.4] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> from math import log as _log, exp as _exp, pi as _pi, e as _e, ceil as _ceil >>> exit() ******************************************************************** my python version: ******************************************************************** carlosp@papa ~/Desktop/alps-2.0.0-r5363-src-with-boost/alps/test/pyalps $ python -V Python 2.6.6 ******************************************************************** Thanks a lot for your help! carlos

2 2

FW: error at installation for pyalea_c.so
by Krunoslav Prsa 19 Jan '11

19 Jan '11

Hi, I get the same problem when installing on the brutus clustur... Best wishes, Kruno ------------------------------------------------- Dr. Krunoslav Prsa, Postdoc Dynamics of strongly correlated materials Laboratory 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 ------------------------------------------------- -----Original Message----- From: Prsa Krunoslav [mailto:Krunoslav.Prsa@psi.ch] Sent: Wednesday, January 19, 2011 8:50 AM To: kprsa(a)phys.ethz.ch Subject: FW: [ALPS-users] error at installation for pyalea_c.so ------------------------------------------- From: comp-phys-alps-users-bounces(a)lists.phys.ethz.ch on behalf of benoît grémaud[SMTP:GREMAUD.BENOIT@GMAIL.COM] Sent: Wednesday, January 19, 2011 8:50:21 AM To: comp-phys-alps-users(a)phys.ethz.ch Subject: [ALPS-users] error at installation for pyalea_c.so Auto forwarded by a Rule Hello, I am compiling alps2.0.0-- -r5363 on a linux machine (fedora) with intel compilers and mkl I use boost provided by alps. python version is 2.6.4 the cmake command is as follows: cmake -D Boost_ROOT_DIR:PATH=~/alps/alps-2.0.0-r5363-src-with-boost/boost -DCMAKE_INSTALL_PREFIX=~/alps/alps-2.0.0-r5363 -D LAPACK_64_BIT=ON -DLPSOLVE_ROOT=~/alps/lp_solve_4.0 ~/alps/alps-2.0.0-r5363-src-with-boost/alps cmake, make, make test, all went fine. however, I got the following error during the install: CMake Error at lib/pyalps/cmake_install.cmake:41 (FILE): file INSTALL cannot find "/home/gremaud/alps/alps-2.0.0-r5363/lib/pyalps/pyalea_c.so". Call Stack (most recent call first): cmake_install.cmake:94 (INCLUDE) I checked that before the installation the file pyalea_c.so existed. Strangely enough, after the preceding error, it has disappeared. any clue of what happened? regards, Benoît Grémaud -------------------- Researcher at CNRS, Laboratoire Kastler Brossel (Paris, France) Visiting Associate Professor at the Centre for Quantum Technologies (NUS, Singapore)

1 0

Re: [ALPS-users] How to run Spin Glass systems
by Hakob Abajyan 19 Jan '11

19 Jan '11

Any help would be appreciated. ________________________________ From: Hakob Abajyan <hakob.abajyan(a)yahoo.com> To: comp-phys-alps-devel(a)lists.phys.ethz.ch; comp-phys-alps-users(a)lists.phys.ethz.ch Sent: Sat, January 15, 2011 5:23:02 AM Subject: How to run Spin Glass systems ________________________________ Hi ALPS Team, I have investigated tutorials on ALPS web page but could not find any instructions how to run "classical Heisenberg spin-glass model" on ALPS. Could you please, answer to the following questions ? 1. Does ALPS support "classical Heisenberg spin-glass model" (Interaction constants are generated randomly) ? 2. If the answer of first question is positive, how to do that ? 3. Does ALPS support ensemble simulation or just simulation of one spin chain?Thanks in advance, Hakob

1 0

error at installation for pyalea_c.so
by benoît grémaud 19 Jan '11

19 Jan '11

Hello, I am compiling alps2.0.0-- -r5363 on a linux machine (fedora) with intel compilers and mkl I use boost provided by alps. python version is 2.6.4 the cmake command is as follows: cmake -D Boost_ROOT_DIR:PATH=~/alps/alps-2.0.0-r5363-src-with-boost/boost -DCMAKE_INSTALL_PREFIX=~/alps/alps-2.0.0-r5363 -D LAPACK_64_BIT=ON -DLPSOLVE_ROOT=~/alps/lp_solve_4.0 ~/alps/alps-2.0.0-r5363-src-with-boost/alps cmake, make, make test, all went fine. however, I got the following error during the install: CMake Error at lib/pyalps/cmake_install.cmake:41 (FILE): file INSTALL cannot find "/home/gremaud/alps/alps-2.0.0-r5363/lib/pyalps/pyalea_c.so". Call Stack (most recent call first): cmake_install.cmake:94 (INCLUDE) I checked that before the installation the file pyalea_c.so existed. Strangely enough, after the preceding error, it has disappeared. any clue of what happened? regards, Benoît Grémaud -------------------- Researcher at CNRS, Laboratoire Kastler Brossel (Paris, France) Visiting Associate Professor at the Centre for Quantum Technologies (NUS, Singapore)

1 0

fulldiag with complex/imaginary numbers
by Itamar Kimchi 19 Jan '11

19 Jan '11

Hi all, I'm trying to run fulldiag in python on a spin model that involves terms like Sx(i)Sy(j). I can define Sy using the imaginary unit uppercase "I" with no immediate error message. I can also use terms such as Sy(i)Sy(j). (My Hamiltonian definition includes the line <PARAMETER name="COMPLEX" default="True"/> .) However, if I try to use a term with a complex coefficient such as (1+I), or if I try to have an imaginary term in the Hamiltonian such as Sx(i) Sy(j), I get an "Execution error in evaluateFulldiagVersusT". Alternatively, if I don't use evaluateFulldiagVersusT, but instead set writexml=True in the options for running fulldiag, then I get no error message but the fulldiag program appears not to run (and no xml file is created) if I have terms with complex/imaginary coefficients. What can I do to include terms such as Sx+Sy in my Hamiltonian? Thanks, Itamar

1 0

Building ALPS 2.0.0 with the Intel MKL libraries
by Baker D.J. 18 Jan '11

18 Jan '11

Hello, I'm fairly new to the ALPS package. I'm installing the most recent formal 2.0.0 release on our computation cluster (running RHEL 5.3) for one of our users at the university, and ideally I would like to build the package with the Intel MKL libraries. I took a brief look through the mailing list archive yesterday to see if I could find some clues. A recent posting suggested that the following recipe should do the trick: export MKL_HOME=/local/software/intel/mkl/10.2.1.017 ccmake -DQMC_BITS=64 -DLAPACK_LINKER_FLAGS="-lmkl_sequential -lmkl_core -lmkl_intel_lp64" /local/software/rh53/alps/2.0.0/source/alps-2.0.0-r5363-src-with-boost/alps Unfortunately this doesn't appear to work. In fact I think that the MKL references are being ignored, and the package is being built using the internal blas/lapack libraries - I could be wrong. I'm building with the GNU compilers by the way. Does anyone please have any useful advice on how to build ALPS with the MKL libraries. It would be very much appreciated. Best - David.

2 2