Mailman 3 June 2007 - NumPy-Discussion

Building NumPy 1.0.3 on SGI/IRIX system
by Atro Tossavainen 12 Jun '07

12 Jun '07

(Am not a list member, please cc copies to me explicitly.) Regarding recent conversation between Charles R Harris and Mary Haley: 1) The problem with Numpy is that *something* causes -L/usr/lib to be included explicitly. This should never be done! The compiler ABI switches -32, -n32, -64 automatically cause the inclusion of the library directories for the correct ABI. If you explicitly include one of /usr/lib{,32,64} with -L, the compiler ABI choice probably conflicts with it, and even if you don't explicitly specify an ABI for the compiler, /etc/compiler.defaults does it for you anyway. 2) Am willing to help with Numpy building/testing on IRIX with MIPS compilers, have considerable expertise building F/OSS software in general and under IRIX in particular. -- Atro Tossavainen (Mr.) / The Institute of Biotechnology at Systems Analyst, Techno-Amish & / the University of Helsinki, Finland, +358-9-19158939 UNIX Dinosaur / employs me, but my opinions are my own. < URL : http : / / www . helsinki . fi / %7E atossava / > NO FILE ATTACHMENTS

1 0

VMWare Virtual Appliance of Ubuntu with numpy, scipy, matplotlib, and ipython available
by Andrew Straw 12 Jun '07

12 Jun '07

This is a note to announce the availability of a VMWare Virtual Appliance with Ubuntu linux with numpy, scipy, matplotlib, and ipython installed. This should make it relatively easy to try out the software. The VMWare Player and VMWare Server are available for no cost from http://www.vmware.com/products/player/ and http://www.vmware.com/products/server/ The download URL is: http://mosca.caltech.edu/outgoing/Ubuntu%207.04%20for%20scientific%20comput… The username is "ubuntu" and the password is "abc123". The network will share the host's interface using NAT. The md5sum is 4191e13abda1154c94e685ffdc0f829b. I have updated http://scipy.org/Download with this information.

7 8

numpy 1.0.3 import error
by Kathy Frank 11 Jun '07

11 Jun '07

I receive the following error when I try to import numpy: $ python Python 2.5.1 (r251:54863, Jun 10 2007, 14:46:50) [GCC 3.4.6 [FreeBSD] 20060305] on freebsd6 Type "help", "copyright", "credits" or "license" for more information. >>> import numpy Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module> import linalg File "/usr/local/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module> from linalg import * File "/usr/local/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module> from numpy.linalg import lapack_lite ImportError: /usr/local/lib/python2.5/site-packages/numpy/linalg/lapack_lite.so: mmap returned wrong address: wanted 0x8048000, got 0x28cc8000 I am running python 2.5.1 on FreeBSD 6.2. I compiled numpy with gcc 3.4.6 and with ATLAS 3.6.0 with no errors. Any help on this error would be appreciated. Steve

1 0

Editing the SciPy wiki pages
by rex 10 Jun '07

10 Jun '07

While compiling Numpy using MKL9.1 is fresh in my mind, I'd like to update some things in the /Installing_SciPy/Linux page. I've registered as a user, but still am not allowed to edit the page. What's required? (I run a couple of Mediawiki sites, so it shouldn't take me long to do simple edits in MoinMoin.) -rex

2 2

tile() can't handle empty arrays
by dmitrey 10 Jun '07

10 Jun '07

in octave repmat([],2,3) works well: ans = [](0x0) I guess in MATLAB too but numpy tile() can't handle it correctly: >>> from numpy import * >>> tile(array([1]), (2,3)) array([[1, 1, 1], [1, 1, 1]]) >>> tile(array([]), (2,3)) Traceback (innermost last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.5/site-packages/numpy/lib/shape_base.py", line 623, in tile c = c.reshape(-1,n).repeat(nrep,0) ValueError: total size of new array must be unchanged (I have numpy 1.0.1) D.

1 0

Numpy-1.0.3 RPM install failure on SUSE 10.2
by rex 10 Jun '07

10 Jun '07

I don't know if this is the appropriate place for this, but thanks for any pointers to what the problem is. -rex deserv:/ # smart install python-numpy Loading cache... Updating cache... ########################################################## [100%] Computing transaction... Installing packages (3): blas-3.0-958@i586 lapack-3.0-958@i586 python-numpy-1.0.3-0.pm.1@i586 9.6MB of package files are needed. 30.5MB will be used. Confirm changes? (Y/n): y Fetching packages... -> http://download.opensuse.org/distribution/10.2/repo/oss/suse/i586/blas-3.0-… -> http://packman.inode.at/suse/10.2/i586/python-numpy-1.0.3-0.pm.1.i586.rpm blas-3.0-958.i586.rpm ########################################################## [ 33%] -> http://download.opensuse.org/distribution/10.2/repo/oss/suse/i586/lapack-3.… python-numpy-1.0.3-0.pm.1.i586.rpm ########################################################## [ 66%] lapack-3.0-958.i586.rpm ########################################################## [100%] Committing transaction... Preparing... ########################################################## [ 0%] 1:Installing blas ########################################################## [ 33%] 2:Installing lapack ########################################################## [ 66%] 3:Installing python-numpy ########################################################## [100%] deserv:/ # python Python 2.5 (r25:51908, Nov 27 2006, 19:14:46) [GCC 4.1.2 20061115 (prerelease) (SUSE Linux)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> from numpy import * Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.5/site-packages/numpy/__init__.py", line 43, in <module> import linalg File "/usr/lib/python2.5/site-packages/numpy/linalg/__init__.py", line 4, in <module> from linalg import * File "/usr/lib/python2.5/site-packages/numpy/linalg/linalg.py", line 25, in <module> from numpy.linalg import lapack_lite ImportError: /usr/lib/libblas.so.3: undefined symbol: _gfortran_st_write_done >>>

3 6

Compiling Numpy with Intel MKL9.1 on SUSE10.2
by rex 09 Jun '07

09 Jun '07

I changed the cc_exe line in numpy-1.0.3/numpy/distutils/intelccompiler.py to: cc_exe = 'icc -msse3 -xP -fast' #Core 2 Duo >From the numpy-1.03 directory executed: python setup.py config --compiler=intel build_clib --compiler=intel build_ext --compiler=intel install --prefix=/usr/local/ [much snipped] compiling C sources C compiler: icc -msse3 -xP -fast creating build/temp.linux-i686-2.5/numpy/fft compile options: '-Inumpy/core/include -Ibuild/src.linux-i686-2.5/numpy/core -Inumpy/core/src -Inumpy/core/include -I/usr/include/python2.5 +-c' icc: numpy/fft/fftpack.c icc: numpy/fft/fftpack_litemodule.c creating build/lib.linux-i686-2.5/numpy/fft icc -msse3 -xP -fast -shared build/temp.linux-i686-2.5/numpy/fft/fftpack_litemodule.o build/temp.linux-i686-2.5/numpy/fft/fftpack.o +-L/usr/lib/python2.5/config -lpython2.5 -o build/lib.linux-i686-2.5/numpy/fft/fftpack_lite.so ipo: remark #11000: performing multi-file optimizations ipo: remark #11005: generating object file /tmp/ipo_iccdXKswc.o numpy/fft/fftpack.c(1203): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(293): (col. 9) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(336): (col. 7) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(350): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1313): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1440): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(813): (col. 7) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(875): (col. 7) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(912): (col. 7) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(917): (col. 9) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1447): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(986): (col. 9) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1072): (col. 7) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1110): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1496): (col. 5) remark: LOOP WAS VECTORIZED. numpy/fft/fftpack.c(1433): (col. 5) remark: LOOP WAS VECTORIZED. building 'numpy.linalg.lapack_lite' extension compiling C sources building 'numpy.linalg.lapack_lite' extension compiling C sources C compiler: icc -msse3 -xP -fast creating build/temp.linux-i686-2.5/numpy/linalg compile options: '-DSCIPY_MKL_H -I/opt/intel/mkl/9.1/include -Inumpy/core/include -Ibuild/src.linux-i686-2.5/numpy/core -Inumpy/core/src -Inumpy/core/include -I/usr/include/python2.5 -c' icc: numpy/linalg/lapack_litemodule.c creating build/lib.linux-i686-2.5/numpy/linalg icc -msse3 -xP -fast -shared build/temp.linux-i686-2.5/numpy/linalg/lapack_litemodule.o -L/opt/intel/mkl/9.1/lib/32 -L/usr/lib/python2.5/config -lmkl_lapack32 -lmkl_lapack64 -lmkl -lvml -lguide -lpthread -lpython2.5 -o build/lib.linux-i686-2.5/numpy/ linalg/lapack_lite.so IPO link: can not find -lmkl_lapack32 icc: error #10014: problem during multi-file optimization compilation (code 1) IPO link: can not find -lmkl_lapack32 icc: error #10014: problem during multi-file optimization compilation (code 1) error: Command "icc -msse3 -xP -fast -shared build/temp.linux-i686-2.5/numpy/linalg/lapack_litemodule.o -L/opt/intel/mkl/9.1/lib/32 -L/usr/lib/python2.5/config -lmkl_lapack32 -lmkl_lapack64 -lmkl -lvml -lguide -lpthread -lpython2.5 -o build/lib.linux-i686-2.5/numpy/linalg/lapack_lite.so" failed with exit status 1 There is no mkl_lapack32 in /opt/intel/mkl/9.1/lib/32 There are only 'libmkl_lapack.a' and 'libmkl_lapack.so' Likewise, there is no mkl_lapack64 in/opt/intel/mkl/9.1/lib/64 I don't understand where the '-lmkl_lapack32' and '-lmkl_lapack64' are coming from. I looked in lapack_litemodule.c, but didn't see anything about lapack. Thanks for any pointers, -rex

1 1

I hate for loops
by Mathew Yeates 08 Jun '07

08 Jun '07

Hi I'm looking for a more elegant way of setting my array elements Using "for" loops it would be for i in range(rows): for j in range(cols): N[i,j] = N[i-1][j] + N[i][j-1] - N[i-1][j-1] It's sort of a combined 2d accumulate. Any ideas? Mathew

2 2

flatten() without copy - is this possible?
by dmitrey 08 Jun '07

08 Jun '07

hi all. in the numpy for matlab users I read y = x.flatten(1) turn array into vector (note that this forces a copy) Is there any way to do the trick wthout copying? What are the problems here? Just other way of array elements indexing... Thx, D.

4 5

numarray.fft causes fatal Python error
by Sebastien Maret 08 Jun '07

08 Jun '07

% python Python 2.5.1 (r251:54863, May 11 2007, 11:07:19) [GCC 4.0.1 (Apple Computer, Inc. build 5367)] on darwin Type "help", "copyright", "credits" or "license" for more information. from >>> from numarray import * >>> import numarray.fft >>> a = array([1., 0., 1., 0., 1., 0., 1., 0.]) + 10 >>> for i in range(5000): ... b = numarray.fft.fft(a).real ... Fatal Python error: deallocating None Abort trap I am using Numarray 1.5.2. See also: https://sourceforge.net/tracker/?func=detail&atid=450446&aid=1732413&group_… Sébastien

1 0