wmap 3yr likelihood llibraries
-
- Posts: 21
- Joined: December 10 2004
- Affiliation: Buenos Aires University
wmap 3yr likelihood llibraries
Hi eveyone !
It seems the SDSL Library or LAPACK library are not installed in my system because
while compiling the wmap 3yr likelihood code, I get the following message:
ifort -g -O1 -o test read_archive_map.o read_fits.o WMAP_3yr_options.o WMAP_3yr_util.o WMAP_3yr_tt_pixlike.o WMAP_3yr_tt_beam_and_ptsrc_corr.o WMAP_3yr_teeebb_pixlike.o WMAP_3yr_likelihood.o test.o -L. -L/usr/local/Healpix_2.01/lib -L/usr/lib -lhealpix -lcfitsio -lscs
ld: cannot find -lscs
or when trying with lapack
ifort -g -O1 -o test read_archive_map.o read_fits.o WMAP_3yr_options.o WMAP_3yr_util.o WMAP_3yr_tt_pixlike.o WMAP_3yr_tt_beam_and_ptsrc_corr.o WMAP_3yr_teeebb_pixlike.o WMAP_3yr_likelihood.o test.o -L. -L/usr/local/Healpix_2.01/lib -L/usr/lib -lhealpix -lcfitsio -llapack
ld: cannot find -llapack
I am using intel compiler v8.1 in a Mandrake Linux 10.0 system
I try to install the lapack library but I have several problems with installation.
I did not found any place to download the SCSL Science Libary,
any ideas ?
thank you very much in advanced
It seems the SDSL Library or LAPACK library are not installed in my system because
while compiling the wmap 3yr likelihood code, I get the following message:
ifort -g -O1 -o test read_archive_map.o read_fits.o WMAP_3yr_options.o WMAP_3yr_util.o WMAP_3yr_tt_pixlike.o WMAP_3yr_tt_beam_and_ptsrc_corr.o WMAP_3yr_teeebb_pixlike.o WMAP_3yr_likelihood.o test.o -L. -L/usr/local/Healpix_2.01/lib -L/usr/lib -lhealpix -lcfitsio -lscs
ld: cannot find -lscs
or when trying with lapack
ifort -g -O1 -o test read_archive_map.o read_fits.o WMAP_3yr_options.o WMAP_3yr_util.o WMAP_3yr_tt_pixlike.o WMAP_3yr_tt_beam_and_ptsrc_corr.o WMAP_3yr_teeebb_pixlike.o WMAP_3yr_likelihood.o test.o -L. -L/usr/local/Healpix_2.01/lib -L/usr/lib -lhealpix -lcfitsio -llapack
ld: cannot find -llapack
I am using intel compiler v8.1 in a Mandrake Linux 10.0 system
I try to install the lapack library but I have several problems with installation.
I did not found any place to download the SCSL Science Libary,
any ideas ?
thank you very much in advanced
-
- Posts: 9
- Joined: March 10 2005
- Affiliation: Berkeley Center for Cosmological Physics
wmap 3yr likelihood llibraries
Did you locate the Lapack libraries on your system? On my Linux machine using
-llapack -lblas
in place of
-lscs
did the job.
Now am having other problems with the compilation of cosmomc though...
-llapack -lblas
in place of
-lscs
did the job.
Now am having other problems with the compilation of cosmomc though...
-
- Posts: 21
- Joined: December 10 2004
- Affiliation: Buenos Aires University
wmap 3yr likelihood llibraries
I also tried compiling with -llapack -lblas and it does not work.
I also tried to install the lapack library downloading it form
http://www.netlib.org/lapack/
but I have several problems with this installation,
do you think I have to install both LAPACK and BLAS?
any idea how to check if LAPACK and/or BLAS are installed on my system?
I also tried to install the lapack library downloading it form
http://www.netlib.org/lapack/
but I have several problems with this installation,
do you think I have to install both LAPACK and BLAS?
any idea how to check if LAPACK and/or BLAS are installed on my system?
-
- Posts: 9
- Joined: March 10 2005
- Affiliation: Berkeley Center for Cosmological Physics
wmap 3yr likelihood llibraries
Try "locate liblapack.a" and "locate libblas.a". If it doesn't return a path to a directory containing those libraries, you probably don't have them.
-
- Posts: 5
- Joined: October 06 2004
- Affiliation: Helsinki Institute of Physics
- Contact:
Re: wmap 3yr likelihood llibraries
Hi!Susana Landau wrote: I am using intel compiler v8.1 in a Mandrake Linux 10.0 system
I try to install the lapack library but I have several problems with installation.
I did not found any place to download the SCSL Science Libary,
any ideas ?
I managed to solve this by installing the Intel Math Kernel Library for Linux. Then with the following options
I got the compilation go through without problems:
Code: Select all
INCS = -I. -I/usr/local/src/Healpix_2.01/include_ifort -I/usr/local/include
LIBS = -L. -L/usr/local/src/Healpix_2.01/lib_ifort -L/usr/local/lib
-L/opt/intel/mkl/8.0.2/lib/32 -lhealpix -lcfitsio -lmkl -lmkl_lapack
actually does the right thing. There was a difference in the 4th decimal in the total likelihood, so it's
not exactly the same result, but probably not catastrophically wrong either...
-
- Posts: 21
- Joined: December 10 2004
- Affiliation: Buenos Aires University
wmap 3yr likelihood llibraries
Thank you very much Oliver and Vesa.
Installing the Intel Math Kernel Library and modifying the makefile like Vesa suggested did work!!!! I also find a fouth decimal difference checking the test routine.
and a message
WARNING: obsolete interface to MAP2ALM:
cos_theta_cut (6th argument) currently a DP scalar with value
-1.00000
shoud now be replaced with a 2-element vector with values:
0.00000 0.00000
See documentation for details.
this comes from some problems with the library routines because it is not in any part of de wmap likelihood code. any ideas?
Installing the Intel Math Kernel Library and modifying the makefile like Vesa suggested did work!!!! I also find a fouth decimal difference checking the test routine.
and a message
WARNING: obsolete interface to MAP2ALM:
cos_theta_cut (6th argument) currently a DP scalar with value
-1.00000
shoud now be replaced with a 2-element vector with values:
0.00000 0.00000
See documentation for details.
this comes from some problems with the library routines because it is not in any part of de wmap likelihood code. any ideas?
-
- Posts: 5
- Joined: October 06 2004
- Affiliation: Helsinki Institute of Physics
- Contact:
Re: wmap 3yr likelihood llibraries
David Parkinson actually has already commented on this in this post, that the warning is due to using the latest versionSusana Landau wrote: WARNING: obsolete interface to MAP2ALM:
cos_theta_cut (6th argument) currently a DP scalar with value
-1.00000
shoud now be replaced with a 2-element vector with values:
0.00000 0.00000
See documentation for details.
this comes from some problems with the library routines because it is not in any part of de wmap likelihood code. any ideas?
of Healpix. However, this seems not to be a problen since he got the same results with an older version of Healpix.