Hi all, Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS. Do I need version 3 of gcc ? I am on SuSE Linux 10.0 with gcc v4.0.2 If I try to install ATLAS I get .. Looking for compilers (this may take a while): /usr/bin/gcc : v4.0.2 /usr/bin/gcc : v4.0.2 F77 = /usr/bin/g77 -fomit-frame-pointer -O -m64 CC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 MCC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 ATLAS has detected that you are running on an x86 architecture (eg., Athlon, P4, PIII, etc), but do not have a pre-2.96 gcc installed. Using gcc 3/2.96 will result in a performance loss of 10-20% IN THE BEST CASE over a gcc previous to 2.96. If your install goes slightly wrong, performance drops of 50% are more normal with these compilers. Full details of this problem can be found at: http://www.cs.utk.edu/~rwhaley/ATLAS/gcc30.html In the meantime, we suggest you abondon this ATLAS install, and install a gcc that delivers decent x86 performance, and restart the install then. You do not need to be root to install, and full instructions, along with all needed files, are given in the ATLAS errata file at: http://math-atlas.sourceforge.net/errata.html#UnixGccInst Stop ATLAS install? [y]: How can I continue ? Nils
Nils, The website you cite says that the problem was fixed with version 3.1 of gcc. Why don't you just ry it and see what happens, or does the compile break? Hanno Nils Wagner <nwagner@mecha.uni-stuttgart.de> said:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS.
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
If I try to install ATLAS I get ..
Looking for compilers (this may take a while): /usr/bin/gcc : v4.0.2 /usr/bin/gcc : v4.0.2 F77 = /usr/bin/g77 -fomit-frame-pointer -O -m64 CC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 MCC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64
ATLAS has detected that you are running on an x86 architecture (eg., Athlon, P4, PIII, etc), but do not have a pre-2.96 gcc installed. Using gcc 3/2.96 will result in a performance loss of 10-20% IN THE BEST CASE over a gcc previous to 2.96. If your install goes slightly wrong, performance drops of 50% are more normal with these compilers. Full details of this problem can be found at: http://www.cs.utk.edu/~rwhaley/ATLAS/gcc30.html In the meantime, we suggest you abondon this ATLAS install, and install a gcc that delivers decent x86 performance, and restart the install then. You do not need to be root to install, and full instructions, along with all needed files, are given in the ATLAS errata file at: http://math-atlas.sourceforge.net/errata.html#UnixGccInst
Stop ATLAS install? [y]:
How can I continue ?
Nils
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
-- Hanno Klemm klemm@phys.ethz.ch
On Tue, 2006-02-28 at 09:34 +0100, Nils Wagner wrote:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS.
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
If I try to install ATLAS I get ..
Looking for compilers (this may take a while): /usr/bin/gcc : v4.0.2 /usr/bin/gcc : v4.0.2 F77 = /usr/bin/g77 -fomit-frame-pointer -O -m64 CC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 MCC = /usr/bin/gcc -fomit-frame-pointer -O -mfpmath=387 -m64 On a different distribution (ubuntu), I install ATLAS package, and modifiy the static library to build a full LAPACK library (it should be somewhere on the wiki, I can give more details if you need help).
From my POV, the best solution is to use a pre-compiled ATLAS from your distribution: then, it should have been tested, and you are less likely to make a error which can appear later in a very subtle way. Compiling and installing gcc yourself is a PITA if you have never done it yourself before (it is easy to break many things). Even building atlas yourself is a bit long, and I think it is easy to mess things up, at least from a
If SUSE does not have atlas package, then try to find if a gcc-2.95 package is available. If neither of them are available, I would advise you to just build ATLAS with your current version of gcc. performance point of view. David
Hi,
ATLAS has detected that you are running on an x86 architecture (eg., Athlon, P4, PIII, etc), but do not have a pre-2.96 gcc installed. Using gcc 3/2.96 will result in a performance loss of 10-20% IN THE BEST CASE over a gcc previous to 2.96. If your install goes slightly wrong, performance drops of 50% are more normal with these compilers. Full details of this problem can be found at: http://www.cs.utk.edu/~rwhaley/ATLAS/gcc30.html
If you actually click on the link you can read that "all of these problems appear to be fixed in gcc 3.1". So the message appears to be outdated, and should not apply to gcc 4.0.2. Do not be stopped by a mere warning ! But I am not an atlas expert, just one of many people who is annoyed by the incomplete blas/lapack libraries, so that one cannot invert a matrix in scipy without compiling atlas, which can be a major pain. Of course I have not yet switched to the new scipy, so maybe things are better now. Or packagers have finally decided to ban incomplete blas/lapack libraries and therefore build a better world where the sky is blue and matrices invert. Vincent -- Vincent Favre-Nicolin Université Joseph Fourier http://v.favrenicolin.free.fr ObjCryst & Fox : http://objcryst.sourceforge.net
Nils Wagner wrote:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS.
Well, at least it provides a link to a page with install instructions: http://www.scipy.org/Installing_SciPy/Linux
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
In addition to what was already said, I noticed that its important, that the gfortran compiler is used and not g77. The easiest way to achieve that, is to make sure that compat-g77.rpm is not installed, but gfortran.rpm is. And yes, gcc4 works without problems though I can't say anything about performance. Eric gave instructions to compile f2py. That, however is not needed with numpy as it's already included. Regards, Christian
On Tue, 28 Feb 2006 18:29:44 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS.
Well, at least it provides a link to a page with install instructions:
http://www.scipy.org/Installing_SciPy/Linux
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
In addition to what was already said, I noticed that its important, that the gfortran compiler is used and not g77. The easiest way to achieve that, is to make sure that compat-g77.rpm is not installed, but gfortran.rpm is.
And yes, gcc4 works without problems though I can't say anything about performance.
Eric gave instructions to compile f2py. That, however is not needed with numpy as it's already included.
Regards, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
Hi Christian, SuSE 10.0 comes with gcc-fortran Name : gcc-fortran Relocations: (not relocatable) Version : 4.0.2_20050901 Vendor: SUSE LINUX Products GmbH, Nuernberg, Germany Release : 3 Build Date: Fri 09 Sep 2005 06:08:56 PM CEST Install date: Mon 27 Feb 2006 01:01:19 PM CET Build Host: fatou.suse.de Group : Development/Languages/Fortran Source RPM: gcc-4.0.2_20050901-3.src.rpm Size : 7490931 License: GPL Signature : DSA/SHA1, Fri 09 Sep 2005 07:04:43 PM CEST, Key ID a84edae89c800aca Packager : http://www.suse.de/feedback URL : http://gcc.gnu.org/ Summary : The GNU Fortran Compiler and Support Files Description : This is the Fortran compiler of the GNU Compiler Collection (GCC). Authors: -------- The GCC team. Distribution: SUSE LINUX 10.0 (X86-64) Nils
Nils Wagner wrote:
On Tue, 28 Feb 2006 18:29:44 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS. Well, at least it provides a link to a page with install instructions:
http://www.scipy.org/Installing_SciPy/Linux
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
In addition to what was already said, I noticed that its important, that the gfortran compiler is used and not g77. The easiest way to achieve that, is to make sure that compat-g77.rpm is not installed, but gfortran.rpm is.
And yes, gcc4 works without problems though I can't say anything about performance.
Eric gave instructions to compile f2py. That, however is not needed with numpy as it's already included.
Regards, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
Hi Christian,
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up. Christian
On Tue, 28 Feb 2006 18:58:03 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
On Tue, 28 Feb 2006 18:29:44 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
Hi all,
Where can I find some hints w.r.t. to ATLAS and scipy/numpy ? AFAIK the new website has no information on building ATLAS. Well, at least it provides a link to a page with install instructions:
http://www.scipy.org/Installing_SciPy/Linux
Do I need version 3 of gcc ?
I am on SuSE Linux 10.0 with gcc v4.0.2
In addition to what was already said, I noticed that its important, that the gfortran compiler is used and not g77. The easiest way to achieve that, is to make sure that compat-g77.rpm is not installed, but gfortran.rpm is.
And yes, gcc4 works without problems though I can't say anything about performance.
Eric gave instructions to compile f2py. That, however is not needed with numpy as it's already included.
Regards, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
Hi Christian,
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints. Can I use them or must I recompile blas and lapack with gfortran ? Nils
Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
Nils Wagner wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ?
I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0. Good luck, Christian
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ?
I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0.
Good luck, Christian
Christian, I have recompiled both blas and lapack. There is a difference in the size of the libraries -rw-r--r-- 1 root root 7887268 2006-02-28 09:19 libflapack.a_g77 -rw-r--r-- 1 root root 7833516 2006-02-28 11:20 libflapack.a_gfortran -rw-r--r-- 1 root root 629322 2006-02-28 11:06 libfblas.a_g77 -rw-r--r-- 1 root root 615722 2006-02-28 11:28 libfblas.a_gfortran Nils
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ?
I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack http://math-atlas.sourceforge.net/errata.html#completelp I am on SuSE 10.0 x86_64 GNU/Linux However python setup.py build failed Running from numpy source directory. Warning: not existing path in numpy/distutils: site.cfg No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE atlas_blas_threads_info: Setting PTATLAS=ATLAS /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE atlas_blas_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so /usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status /usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)] Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend' Any suggestion ? Nils
Nils Wagner wrote:
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations: You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ? I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in
Nils Wagner wrote: the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack http://math-atlas.sourceforge.net/errata.html#completelp
I am on SuSE 10.0 x86_64 GNU/Linux
However
python setup.py build failed
Running from numpy source directory. Warning: not existing path in numpy/distutils: site.cfg No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_threads_info: Setting PTATLAS=ATLAS /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c
running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so /usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status /usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)]
Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend'
I can't say more than that I followed exactly the instructions here: http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 (the full LAPACK issue is covered there, too) and it worked. Christian
Hi Christian, you output looks like a problem I had: On my 64bit architecture I had to modify the make file for atlas and append an -fPIC at the compiler flags (I tired that at the architecture flags, the g77 and the c flags. That did the job). Otherwise it seems not possible to use the libraries with scipy. Good luck, Hanno Christian Kristukat <ckkart@hoc.net> said:
Nils Wagner wrote:
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations: You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ? I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in
Nils Wagner wrote: the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack http://math-atlas.sourceforge.net/errata.html#completelp
I am on SuSE 10.0 x86_64 GNU/Linux
However
python setup.py build failed
Running from numpy source directory. Warning: not existing path in numpy/distutils: site.cfg No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_threads_info: Setting PTATLAS=ATLAS /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c
running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)]
Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend'
I can't say more than that I followed exactly the instructions here: http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 (the full LAPACK issue is covered there, too) and it worked.
Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
-- Hanno Klemm klemm@phys.ethz.ch
On Tue, 28 Feb 2006 18:49:06 +0100 "Hanno Klemm" <klemm@phys.ethz.ch> wrote:
Hi Christian,
you output looks like a problem I had:
On my 64bit architecture I had to modify the make file for atlas and append an -fPIC at the compiler flags (I tired that at the architecture flags, the g77 and the c flags. That did the job). Otherwise it seems not possible to use the libraries with scipy.
Good luck, Hanno
Hi Hanno, Thank you for your valuable comment. Did you also apply the patch to ATLAS/bin/uumtst.c ? The sanity_test failed here. http://math-atlas.sourceforge.net/errata.html#gcc4 Cheers, Nils
Christian Kristukat <ckkart@hoc.net> said:
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
> SuSE 10.0 comes with gcc-fortran > > Name : gcc-fortran Relocations: You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ? I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in
Nils Wagner wrote: the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all
Nils Wagner wrote: the
time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack
http://math-atlas.sourceforge.net/errata.html#completelp
I am on SuSE 10.0 x86_64 GNU/Linux
However
python setup.py build failed
Running from numpy source directory. Warning: not existing path in numpy/distutils:
site.cfg
No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info:
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_threads_info: Setting PTATLAS=ATLAS
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_info:
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c
running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)]
Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o line line line line line line line line
702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend'
I can't say more than that I followed exactly the instructions here: http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 (the full LAPACK issue is covered there, too) and it worked.
Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
-- Hanno Klemm klemm@phys.ethz.ch
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
Hi Nils, no, I compiled with gcc 3.x, therefore I didn't run into that problem. Hanno On Feb 28, 2006, at 8:28 PM, Nils Wagner wrote:
On Tue, 28 Feb 2006 18:49:06 +0100 "Hanno Klemm" <klemm@phys.ethz.ch> wrote:
Hi Christian,
you output looks like a problem I had:
On my 64bit architecture I had to modify the make file for atlas and append an -fPIC at the compiler flags (I tired that at the architecture flags, the g77 and the c flags. That did the job). Otherwise it seems not possible to use the libraries with scipy.
Good luck, Hanno
Hi Hanno,
Thank you for your valuable comment. Did you also apply the patch to ATLAS/bin/uumtst.c ? The sanity_test failed here. http://math-atlas.sourceforge.net/errata.html#gcc4
Cheers,
Nils
Christian Kristukat <ckkart@hoc.net> said:
On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
>> SuSE 10.0 comes with gcc-fortran >> >> Name : gcc-fortran Relocations: > You're right. But the executable is called 'gfortran'. I > mixed that up. > I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ? I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in
Nils Wagner wrote: the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all
Nils Wagner wrote: the
time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack
http://math-atlas.sourceforge.net/errata.html#completelp
I am on SuSE 10.0 x86_64 GNU/Linux
However
python setup.py build failed
Running from numpy source directory. Warning: not existing path in numpy/distutils:
site.cfg
No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info:
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_threads_info: Setting PTATLAS=ATLAS
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE
atlas_blas_info:
/usr/local/svn/numpy/numpy/distutils/system_info.py:531:
UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c
running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/ atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/ bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status
/usr/lib64/gcc/x86_64-suse-linux/4.0.2/../../../../x86_64-suse-linux/ bin/ld:
/usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared
build/temp.linux-x86_64-2.4/build/src/ atlas_version_0x5b360788cdd9eddc.o
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared
-L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)]
Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py",
build/temp.linux-x86_64-2.4/build/src/ atlas_version_0x5b360788cdd9eddc.o line line line line line line line line
702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend'
I can't say more than that I followed exactly the instructions here: http://pong.tamu.edu/tiki/tiki-view_blog_post.php?blogId=6&postId=97 (the full LAPACK issue is covered there, too) and it worked.
Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
-- Hanno Klemm klemm@phys.ethz.ch
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
-- Hanno Klemm klemm@itp.phys.ethz.ch ETH Zurich tel: +41-1-6332580 Institute for theoretical physics mobile: +41-79-4500428 http://www.mth.kcl.ac.uk/~klemm
My own preference would be to use g77 or some other Fortran 77 compiler, not gfortran, to compile BLAS and LAPACK. My reasons are (1) LAPACK and BLAS are written in Fortran 77; (2) the gcc-4 is generally considered to be somewhat less trustworthy than much more mature gcc-3; and gfortran is entirely new with gcc-4, while g77 is ancient and well-tested; and (3) gfortran is particularly new. Perhaps I express the obvious when I say that g77 and gfortran are frontends for gcc; g77 compiles Fortran 77; gfortran compiles a GNU implementation, somewhat partial and peculiar, so I understand, of Fortran 95. It is supposed to be compatible with Fortran 77, but so far as I know, nobody has really used it very much. My impression is that post-77 versions of Fortran are regarded with some degree of suspicion for pure, fast numerics, so why complicate the compilation of LAPACK and BLAS with these issues? I might be more willing to try gfortran if it had some sort of flag to optimize pure Fortran 77 code. Concerning slow compilation on the native Suse compiler: am I wrong, or is Suse compiled for the i386? If you are running on an i686, you would achieve faster compilation if you built gcc optimized for your machine's architecture. I have built gcc-3.4.5 from source both on Linux and AIX, and I didn't find it terribly painful in either case. However, if you don't do much compilation, it probably wouldn't be worth the trouble. Mark F. Morss Principal Analyst, Market Risk American Electric Power "Nils Wagner" <nwagner@mecha.un i-stuttgart.de> To Sent by: SciPy Users List scipy-user-bounce <scipy-user@scipy.net> s@scipy.net cc Subject 02/28/2006 06:32 Re: [SciPy-user] ATLAS and gcc AM v4.0.2 Please respond to SciPy Users List <scipy-user@scipy .net> On Tue, 28 Feb 2006 19:20:55 +0900 Christian Kristukat <ckkart@hoc.net> wrote:
Nils Wagner wrote:
SuSE 10.0 comes with gcc-fortran
Name : gcc-fortran Relocations:
You're right. But the executable is called 'gfortran'. I mixed that up.
I have already build blas and lapack libraries using g77 according to the installation hints.
Can I use them or must I recompile blas and lapack with gfortran ?
I'm not an expert on this subject, just sharing my experiences. If you're using ATLAS you don't need BLAS as far as I understood. I followed the instructions in the link I posted before. Compared to the compilation time of ATLAS, LAPACK builds almost instantaneously, so it doesn't hurt to rebuild LAPACK, I just don't know wether g77 built LAPACK works together with gfortran ATLAS. At least I know, that if you're using gfrotran all the time, everything works fine on SuSE 10.0.
Good luck, Christian
_______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
I followed your advice and rebuilt blas and lapack using gfortran. Compilation of ATLAS seems to work. Also I have build a complete lapack http://math-atlas.sourceforge.net/errata.html#completelp I am on SuSE 10.0 x86_64 GNU/Linux However python setup.py build failed Running from numpy source directory. Warning: not existing path in numpy/distutils: site.cfg No module named __svn_version__ F2PY Version 2_2179 blas_opt_info: blas_mkl_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['mkl', 'vml', 'guide'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE atlas_blas_threads_info: Setting PTATLAS=ATLAS /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['ptf77blas', 'ptcblas', 'atlas'] found_libs=['/usr/local/lib/atlas/libatlas.a'] warnings.warn("Library error: libs=%s found_libs=%s" % \ NOT AVAILABLE atlas_blas_info: /usr/local/svn/numpy/numpy/distutils/system_info.py:531: UserWarning: Library error: libs=['f77blas', 'cblas', 'atlas'] found_libs=[] warnings.warn("Library error: libs=%s found_libs=%s" % \ FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c running build_src building extension "atlas_version" sources creating build creating build/src adding 'build/src/atlas_version_0x5b360788cdd9eddc.c' to sources. running build_ext customize UnixCCompiler customize UnixCCompiler using build_ext building 'atlas_version' extension compiling C sources gcc options: '-pthread -fno-strict-aliasing -DNDEBUG -O2 -fmessage-length=0 -Wall -D_FORTIFY_SOURCE=2 -g -fPIC' creating build/temp.linux-x86_64-2.4 creating build/temp.linux-x86_64-2.4/build creating build/temp.linux-x86_64-2.4/build/src compile options: '-Inumpy/core/include -I/usr/include/python2.4 -c' gcc: build/src/atlas_version_0x5b360788cdd9eddc.c gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so /usr/lib64/gcc/x86_64-suse-linux/4.0.2 /../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status /usr/lib64/gcc/x86_64-suse-linux/4.0.2 /../../../../x86_64-suse-linux/bin/ld: /usr/local/lib/atlas/libatlas.a(ATL_buildinfo.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/atlas/libatlas.a: could not read symbols: Bad value collect2: ld returned 1 exit status ##### msg: error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 error: Command "gcc -pthread -shared build/temp.linux-x86_64-2.4/build/src/atlas_version_0x5b360788cdd9eddc.o -L/usr/local/lib/atlas -lf77blas -lcblas -latlas -o build/temp.linux-x86_64-2.4/atlas_version.so" failed with exit status 1 FOUND: libraries = ['f77blas', 'cblas', 'atlas'] library_dirs = ['/usr/local/lib/atlas'] language = c define_macros = [('NO_ATLAS_INFO', 2)] Warning: distutils distribution has been initialized, it may be too late to add an extension _dotblas Traceback (most recent call last): File "setup.py", line 76, in ?x86_64 GNU/Linux setup_package() File "setup.py", line 63, in setup_package config.add_subpackage('numpy') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "/usr/local/svn/numpy/numpy/setup.py", line 10, in configuration config.add_subpackage('core') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 592, in add_subpackage config_list = self.get_subpackage(subpackage_name,subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 582, in get_subpackage subpackage_path) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 539, in _get_configuration_from_setup_py config = setup_module.configuration(*args) File "numpy/core/setup.py", line 217, in configuration config.add_data_dir('tests') File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 636, in add_data_dir self.add_data_files((ds,filenames)) File "/usr/local/svn/numpy/numpy/distutils/misc_util.py", line 702, in add_data_files dist.data_files.extend(data_dict.items()) AttributeError: 'NoneType' object has no attribute 'extend' Any suggestion ? Nils _______________________________________________ SciPy-user mailing list SciPy-user@scipy.net http://www.scipy.net/mailman/listinfo/scipy-user
participants (6)
-
Christian Kristukat -
Cournapeau David -
Favre-Nicolin Vincent -
Hanno Klemm -
mfmorss＠aep.com -
Nils Wagner