Difference between revisions of "Setup and Installation (Applications)"

From ALPS
Jump to: navigation, search
(Prerequisites)
(Prerequisites)
 
(13 intermediate revisions by 3 users not shown)
Line 10: Line 10:
 
The ALPS Applications require the ALPS library, so you need to [[Download | download]] and [[Setup_and_Installation |  install]] a '''matching version''' of the library first. Additional libraries required by some of the applications are (some of them may be already installed on your system):  
 
The ALPS Applications require the ALPS library, so you need to [[Download | download]] and [[Setup_and_Installation |  install]] a '''matching version''' of the library first. Additional libraries required by some of the applications are (some of them may be already installed on your system):  
  
*  For the [http://alps.comp-phys.org/software/applications/qmc/sse/doc/index.htmlSSE Simulation] package:  
+
*  For the [http://alps.comp-phys.org/software/applications/qmc/sse/doc/index.html SSE Simulation] package:  
  
the linear programming package [http://groups.yahoo.com/group/lp_solve lp_solve] (version [http://alps.comp-phys.org/software/lp_solve_3.2.tar.gz 3.2] or [http://alps.comp-phys.org/software/lp_solve_4.0.tar.gz 4.0] can be installed from ALPS server). The latest versions of lp_solve and more information can be found [http://groups.yahoo.com/group/lp_solve here]. Just do
+
the linear programming package [http://groups.yahoo.com/group/lp_solve lp_solve] (version [http://alps.comp-phys.org/static/software/lp_solve_3.2.tar.gz 3.2] or [http://alps.comp-phys.org/static/software/lp_solve_4.0.tar.gz 4.0] can be installed from ALPS server). The latest versions of lp_solve and more information can be found [http://groups.yahoo.com/group/lp_solve here]. Just do
  
  $ wget [http://alps.comp-phys.org/software/lp_solve_4.0.tar.gz http://alps.comp-phys.org/software/lp_solve_4.0.tar.gz]
+
  $ wget [http://alps.comp-phys.org/static/software/lp_solve_4.0.tar.gz http://alps.comp-phys.org/static/software/lp_solve_4.0.tar.gz]
 
  $ tar zxf lp_solve_4.0.tar.gz
 
  $ tar zxf lp_solve_4.0.tar.gz
 
  $ cd lp_solve_4.0
 
  $ cd lp_solve_4.0
 
  $ make
 
  $ make
  
to setup lp_solve. Otherwise the SSE-binary will not be generated when using make.
+
to setup lp_solve. Otherwise the SSE-binary will not be generated when using make. Depending on your version of bison there might be a compilation problem with lp_solve. To solve this problem, comment out line 17 "extern int yyleng" of the file lpglob.h.
  
*  For the exact diagonalization example program:
 
  
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lpack/ LAPACK] libraries
+
 
 +
*  For the sparse diagonalization program:
 +
 
 +
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lapack/ LAPACK] libraries
  
 
*  For the full diagonalization program:  
 
*  For the full diagonalization program:  
  
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lpack/ LAPACK] libraries
+
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lapack/ LAPACK] libraries
 +
 
 +
*  For the classical Monte Carlo program:
 +
 
 +
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lapack/ LAPACK] libraries
  
 
*  For the loop QMC program:
 
*  For the loop QMC program:
  
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lpack/ LAPACK] libraries
+
The [http://www.netlib.org/blas/ BLAS] (or [http://www.netlib.org/atlas/ ATLAS]) and [http://www.netlib.org/lapack/ LAPACK] libraries
  
 
==  Installation  ==
 
==  Installation  ==
Line 38: Line 44:
 
For easy setup, we have provided a configure script. Just type  
 
For easy setup, we have provided a configure script. Just type  
  
$ ./configure --prefix=ALPS_DIR
+
$ ./configure --prefix=ALPS_DIR
  
 
in the top directory of the ALPS Applications. ALPS_DIR is the directory where the ALPS library is installed (default is $HOME/ALPS). For all directories, please give the '''full path''', starting from root. If certain libraries cannot be found, then the relevant applications will not be created when doing make (next step). The output of the configure scipt, indicates which applications will be generated when doing make (next step). Additional options can be specified to provide the paths to additonal optional libraries if the configure script cannot find them automatically:
 
in the top directory of the ALPS Applications. ALPS_DIR is the directory where the ALPS library is installed (default is $HOME/ALPS). For all directories, please give the '''full path''', starting from root. If certain libraries cannot be found, then the relevant applications will not be created when doing make (next step). The output of the configure scipt, indicates which applications will be generated when doing make (next step). Additional options can be specified to provide the paths to additonal optional libraries if the configure script cannot find them automatically:
  
 
{| border="1"
 
{| border="1"
 
 
|-
 
|-
 
|| --with-lp_solve=DIR
 
|| --with-lp_solve=DIR
Line 49: Line 54:
 
|-
 
|-
 
|| --with-blas=-lLIB
 
|| --with-blas=-lLIB
 +
|}
  
--with-atlas=-lLIB
+
The BLAS and LAPACK libraries need to be specified when configuring the ALPS libraries.
--with-lapack=-lLIB
 
 
 
|| specifies the library files for the BLAS, ATLAS, and LAPACK libraries. If more than one library needs to be linked, they have to be specifed in quotes, such as in --with-blas='-lblas -lf2c'.The BLAS (or ATLAS) and the LAPACK libraries are needed for the full and sparse diagonalization example application.
 
|-
 
|| --with-blas-dir=DIR
 
 
 
--with-atlas-dir=DIR
 
--with-lapack-dir=DIR
 
 
 
|| specifies the place where the BLAS, ATLAS, and LAPACK libraries are installed. The BLAS (or ATLAS) and the LAPACK libraries are needed for the full and sparse diagonalization example application.
 
|-
 
|| --with-dxml
 
|| specifies the use of Compaq Digital Extended Math Library for the BLAS and LAPACK functions.
 
|-
 
|| --with-mkl=PROC
 
|| specifies the use of Intel Math Kernel Library for the BLAS and LAPACK functions. Optional parameter PROC specifies the type of processor (p3 for Pentium 3, p4 for Pentium 4, or itp for Itanium).
 
|-
 
|| --with-scsl
 
|| specifies the use of SGL SCSL Scientific Library for the BLAS and LAPACK functions.
 
|-
 
|| --with-veclib
 
|| specifies the use of vecLib framework on Mac OS X for the BLAS and LAPACK functions.
 
|}
 
  
 
<br/>
 
<br/>
 
After configuration, just type  
 
After configuration, just type  
  
$ make
+
$ make
  
 
to build all the ALPS Applications and  
 
to build all the ALPS Applications and  
  
$ make install
+
$ make install
  
 
to install them into the same location as the ALPS library files. If you want to use the binary files from all locations in your file-system, then you will need to '''set the path''' to that directory.
 
to install them into the same location as the ALPS library files. If you want to use the binary files from all locations in your file-system, then you will need to '''set the path''' to that directory.
 
<br/>
 
  
 
==  Trouble  ==
 
==  Trouble  ==
  
Trouble? go to the [http://alps.comp-phys.org/bugzillaALPS bugzilla] and file a bug!
+
Trouble? Please file a bug in our trac system [https://alps.comp-phys.org/trac ALPS trac] or send an e-mail to the general mailing list!

Latest revision as of 09:55, 21 January 2010

ALPS Library setup and installation documentation is on the separate page " Setup and Installation (Library)".

Download

For downloading the ALPS-applications please visit our download page.


Prerequisites

The ALPS Applications require the ALPS library, so you need to download and install a matching version of the library first. Additional libraries required by some of the applications are (some of them may be already installed on your system):

the linear programming package lp_solve (version 3.2 or 4.0 can be installed from ALPS server). The latest versions of lp_solve and more information can be found here. Just do

$ wget http://alps.comp-phys.org/static/software/lp_solve_4.0.tar.gz
$ tar zxf lp_solve_4.0.tar.gz
$ cd lp_solve_4.0
$ make

to setup lp_solve. Otherwise the SSE-binary will not be generated when using make. Depending on your version of bison there might be a compilation problem with lp_solve. To solve this problem, comment out line 17 "extern int yyleng" of the file lpglob.h.


  • For the sparse diagonalization program:

The BLAS (or ATLAS) and LAPACK libraries

  • For the full diagonalization program:

The BLAS (or ATLAS) and LAPACK libraries

  • For the classical Monte Carlo program:

The BLAS (or ATLAS) and LAPACK libraries

  • For the loop QMC program:

The BLAS (or ATLAS) and LAPACK libraries

Installation


For easy setup, we have provided a configure script. Just type

$ ./configure --prefix=ALPS_DIR

in the top directory of the ALPS Applications. ALPS_DIR is the directory where the ALPS library is installed (default is $HOME/ALPS). For all directories, please give the full path, starting from root. If certain libraries cannot be found, then the relevant applications will not be created when doing make (next step). The output of the configure scipt, indicates which applications will be generated when doing make (next step). Additional options can be specified to provide the paths to additonal optional libraries if the configure script cannot find them automatically:

--with-lp_solve=DIR specifies the place where the lp_solve library is installed. This is needed for the sse application.
--with-blas=-lLIB

The BLAS and LAPACK libraries need to be specified when configuring the ALPS libraries.


After configuration, just type

$ make

to build all the ALPS Applications and

$ make install

to install them into the same location as the ALPS library files. If you want to use the binary files from all locations in your file-system, then you will need to set the path to that directory.

Trouble

Trouble? Please file a bug in our trac system ALPS trac or send an e-mail to the general mailing list!