pysparse-users

Re: [Pysparse-users] Enquiries.

[Guyer, J. E. Dr. <jon...@ni...>]

I can't speak for Bashiru, but when I `pip install pysparse`, it fails with:

```
gcc: pysparse/sparse/src/spmatrixmodule.c

pysparse/sparse/src/spmatrixmodule.c:4:10: fatal error: 'spmatrix.h' file not found

#include "spmatrix.h"

         ^

1 error generated.

pysparse/sparse/src/spmatrixmodule.c:4:10: fatal error: 'spmatrix.h' file not found

#include "spmatrix.h"

         ^

1 error generated.
```


On the other hand, 
```
pip install -e git+git://git.code.sf.net/p/pysparse/git#egg=PySparse
```
works for me. 

It looks like a new tarball, corresponding to git HEAD, needs to be uploaded to PyPI.





On Oct 14, 2014, at 7:00 PM, Dominique Orban <dom...@gm...> wrote:

> Hi,
> 
> Could you please outline the steps you took to install Pysparse? I'm also running Mavericks, so I can probably help. Please indicate what Python you are using (and whether it is system Python or another).
> 
> 
> Dominique
> 
> On Tue, Oct 14, 2014 at 5:40 PM, Bashiru Yahaya <b.y...@ai...> wrote:
> Hi 
> I am a new pysparse user. I have tried all means possible to install pyspares 1.3, but was not successful. I am using apple mac book running on os mavericks. I have an updated Xcode installed on it. Please can someone help me out. I need to work with it.
> Thank you
> 
> -- 
> Bashiru Yahaya
> African Institute For Mathematical Science- Ghana(AIMS-Ghana)
> 23A Awulen Kojo Street
> +233(0) 262525315
> +233(0) 242525315
> www.aims.edu.gh
> 
> 
> DISCLAIMER:  The contents of this email and any attachments are confidential. They are intended for the named recipient(s) only. If you have received this email by mistake, please notify the sender immediately and you are herewith notified that the contents are legally privileged and that you do not have permission to disclose the contents to anyone, make copies thereof, retain or distribute or act upon it by any means, electronically or in print. The views expressed in this communication may be not be those of AIMS-Ghana or AIMS-NEI.
> 
> ------------------------------------------------------------------------------
> Comprehensive Server Monitoring with Site24x7.
> Monitor 10 servers for $9/Month.
> Get alerted through email, SMS, voice calls or mobile push notifications.
> Take corrective actions from your mobile device.
> http://p.sf.net/sfu/Zoho
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
> 
> 
> ------------------------------------------------------------------------------
> Comprehensive Server Monitoring with Site24x7.
> Monitor 10 servers for $9/Month.
> Get alerted through email, SMS, voice calls or mobile push notifications.
> Take corrective actions from your mobile device.
> http://p.sf.net/sfu/Zoho_______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users

Re: [Pysparse-users] Enquiries.

[Dominique O. <dom...@gm...>]

Hi,

Could you please outline the steps you took to install Pysparse? I'm also
running Mavericks, so I can probably help. Please indicate what Python you
are using (and whether it is system Python or another).


Dominique

On Tue, Oct 14, 2014 at 5:40 PM, Bashiru Yahaya <b.y...@ai...>
wrote:

> Hi
> I am a new pysparse user. I have tried all means possible to install
> pyspares 1.3, but was not successful. I am using apple mac book running on
> os mavericks. I have an updated Xcode installed on it. Please can someone
> help me out. I need to work with it.
> Thank you
>
> --
> Bashiru Yahaya
> African Institute For Mathematical Science- Ghana(AIMS-Ghana)
> 23A Awulen Kojo Street
> +233(0) 262525315
> +233(0) 242525315
> www.aims.edu.gh
>
>
> *DISCLAIMER:*  The contents of this email and any attachments are
> confidential. They are intended for the named recipient(s) only. If you
> have received this email by mistake, please notify the sender immediately
> and you are herewith notified that the contents are legally privileged and
> that you do not have permission to disclose the contents to anyone, make
> copies thereof, retain or distribute or act upon it by any means,
> electronically or in print. The views expressed in this communication may
> be not be those of *AIMS-Ghana* or *AIMS-NEI*.
>
>
> ------------------------------------------------------------------------------
> Comprehensive Server Monitoring with Site24x7.
> Monitor 10 servers for $9/Month.
> Get alerted through email, SMS, voice calls or mobile push notifications.
> Take corrective actions from your mobile device.
> http://p.sf.net/sfu/Zoho
> _______________________________________________
> Pysparse-users mailing list
> Pys...@li...
> https://lists.sourceforge.net/lists/listinfo/pysparse-users
>
>

[Pysparse-users] Enquiries.

[Bashiru Y. <b.y...@ai...>]

Hi
I am a new pysparse user. I have tried all means possible to install
pyspares 1.3, but was not successful. I am using apple mac book running on
os mavericks. I have an updated Xcode installed on it. Please can someone
help me out. I need to work with it.
Thank you

-- 
Bashiru Yahaya
African Institute For Mathematical Science- Ghana(AIMS-Ghana)
23A Awulen Kojo Street
+233(0) 262525315
+233(0) 242525315
www.aims.edu.gh

-- 

*DISCLAIMER:*  The contents of this email and any attachments are 
confidential. They are intended for the named recipient(s) only. If you 
have received this email by mistake, please notify the sender immediately 
and you are herewith notified that the contents are legally privileged and 
that you do not have permission to disclose the contents to anyone, make 
copies thereof, retain or distribute or act upon it by any means, 
electronically or in print. The views expressed in this communication may 
be not be those of *AIMS-Ghana* or *AIMS-NEI*.

Re: [Pysparse-users] install help on OSX 10.9.1

[Dominique O. <dom...@gm...>]

On 4 February, 2014 at 12:04:10 AM, Gavan McGrath (gav...@uw...) wrote:
I get the following error when installing. Help appreciated.


pysparse-1.1.1 gavanmcgrath$ python setup.py install
running install
running build
running build_py
creating build
creating build/lib.macosx-10.6-x86_64-2.7
creating build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/__init__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/__version__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/directSolver.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/itsolvers_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/poisson.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/poisson_vec.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseSuperLU.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseUmfpack.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/sparray.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/sparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/spmatrix_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
running build_ext
building 'pysparse.spmatrix' extension
creating build/temp.macosx-10.6-x86_64-2.7
creating build/temp.macosx-10.6-x86_64-2.7/Src
gcc -fno-strict-aliasing -fno-common -dynamic -arch x86_64 -DNDEBUG -g -O3 -I/tmp/_py/libraries/usr/local/include -arch x86_64 -DNUMPY=1 -DLENFUNC_OK=1 -IInclude -I/Users/gavanmcgrath/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/numpy/core/include -I/Applications/Canopy.app/appdata/canopy-1.3.0.1715.macosx-x86_64/Canopy.app/Contents/include/python2.7 -c Src/spmatrixmodule.c -o build/temp.macosx-10.6-x86_64-2.7/Src/spmatrixmodule.o
gcc: error: x86_64: No such file or directory
gcc: error: x86_64: No such file or directory
gcc: error: unrecognized option ‘-arch’
gcc: error: unrecognized option ‘-arch’
error: command 'gcc' failed with exit status 1


Suggestions?
I built PySparse on my Mac without problems, but I'm not on Mavericks. The compiler flags are managed by Numpy's distutils. Not all versions of gcc support the -arch flag. It seems that gcc-4.2, as installed by Homebrew, does. I would suggest that you either use that version of gcc, or temporarily remove gcc from your PATH so clang is used for compiling PySparse.

Let us know if that works.

-- 
Dominique 

Re: [Pysparse-users] install help on OSX 10.9.1

[Daniel W. <dan...@gm...>]

On Tue, Feb 4, 2014 at 11:47 AM, Gavan McGrath <gav...@uw...> wrote:
> Thanks,
> A bit of Googling suggests it is the gcc version used by Xcode. The solution apparently is to either remove the -arch option and/or use an an earlier version of gcc (4.2).

Don't feel shy about patching PySparse if you fix the issue. Thanks.

-- 
Daniel Wheeler

Re: [Pysparse-users] install help on OSX 10.9.1

[Daniel W. <dan...@gm...>]

On Tue, Feb 4, 2014 at 11:35 AM, Daniel Wheeler
<dan...@gm...> wrote:
> On Tue, Feb 4, 2014 at 3:03 AM, Gavan McGrath <gav...@uw...> wrote:
>> I get the following error when installing. Help appreciated.
>>
>>
>> pysparse-1.1.1 gavanmcgrath$ python setup.py install
>> running install
>> running build
>> running build_py
>> creating build
>> creating build/lib.macosx-10.6-x86_64-2.7
>> creating build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/__init__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/__version__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/directSolver.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/itsolvers_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/poisson.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/poisson_vec.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/pysparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/pysparseSuperLU.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/pysparseUmfpack.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/sparray.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/sparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> copying Lib/spmatrix_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
>> running build_ext
>> building 'pysparse.spmatrix' extension
>> creating build/temp.macosx-10.6-x86_64-2.7
>> creating build/temp.macosx-10.6-x86_64-2.7/Src
>> gcc -fno-strict-aliasing -fno-common -dynamic -arch x86_64 -DNDEBUG -g -O3
>> -I/tmp/_py/libraries/usr/local/include -arch x86_64 -DNUMPY=1 -DLENFUNC_OK=1
>> -IInclude
>> -I/Users/gavanmcgrath/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/numpy/core/include
>> -I/Applications/Canopy.app/appdata/canopy-1.3.0.1715.macosx-x86_64/Canopy.app/Contents/include/python2.7
>> -c Src/spmatrixmodule.c -o
>> build/temp.macosx-10.6-x86_64-2.7/Src/spmatrixmodule.o
>> gcc: error: x86_64: No such file or directory
>> gcc: error: x86_64: No such file or directory
>> gcc: error: unrecognized option '-arch'
>> gcc: error: unrecognized option '-arch'
>> error: command 'gcc' failed with exit status 1
>>
>>
>> Suggestions?
>
> Gavin, it seems like you are trying to build PySparse on a mac. I
> don't really know why gcc seems to be complaining about the '-arch'
> argument.
>
> Just a suggestion, but maybe it would be better using Anaconda's
> scientific python distribution on a Mac. It is totally self contained
> and keep stuff up to date and might just deal with whatever is causing
> this issue.
>
> --
> Daniel Wheeler



-- 
Daniel Wheeler

[Pysparse-users] install help on OSX 10.9.1

[Gavan M. <gav...@uw...>]

I get the following error when installing. Help appreciated.


pysparse-1.1.1 gavanmcgrath$ python setup.py install
running install
running build
running build_py
creating build
creating build/lib.macosx-10.6-x86_64-2.7
creating build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/__init__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/__version__.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/directSolver.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/itsolvers_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/poisson.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/poisson_vec.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseSuperLU.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/pysparseUmfpack.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/sparray.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/sparseMatrix.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
copying Lib/spmatrix_util.py -> build/lib.macosx-10.6-x86_64-2.7/pysparse
running build_ext
building 'pysparse.spmatrix' extension
creating build/temp.macosx-10.6-x86_64-2.7
creating build/temp.macosx-10.6-x86_64-2.7/Src
gcc -fno-strict-aliasing -fno-common -dynamic -arch x86_64 -DNDEBUG -g -O3 -I/tmp/_py/libraries/usr/local/include -arch x86_64 -DNUMPY=1 -DLENFUNC_OK=1 -IInclude -I/Users/gavanmcgrath/Library/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/numpy/core/include -I/Applications/Canopy.app/appdata/canopy-1.3.0.1715.macosx-x86_64/Canopy.app/Contents/include/python2.7 -c Src/spmatrixmodule.c -o build/temp.macosx-10.6-x86_64-2.7/Src/spmatrixmodule.o
gcc: error: x86_64: No such file or directory
gcc: error: x86_64: No such file or directory
gcc: error: unrecognized option ‘-arch’
gcc: error: unrecognized option ‘-arch’
error: command 'gcc' failed with exit status 1


Suggestions?



Gavan McGrath
Soil Science
School of Earth and Environment
The University of Western Australia
Crawley, 6009, Australia
Ph: +618 64883735
Mob: +61458559765
gav...@uw...<mailto:gav...@uw...>

Re: [Pysparse-users] CSR slicing

[Dominique O. <dom...@gm...>]

On Saturday, December 14, 2013, Dinesh Vadhia wrote:

>  Long-time scipy sparse user and looking at pysparse wrt improving CSR
> performance.  But, need CSR slicing which isn't mentioned except that
> CSR/SSS matrices cannot be manipulated after creation.  Thanks.
>

Currently, there is no way to slice a CSR matrix, though that's probably
not too hard to implement. I'm not sure any of us has the time to do that
right now but we do welcome contributions.

Cheers.


-- 
Dominique

[Pysparse-users] CSR slicing

[Dinesh V. <din...@ou...>]

Long-time scipy sparse user and looking at pysparse wrt improving CSR performance.  But, need CSR slicing which isn't mentioned except that CSR/SSS matrices cannot be manipulated after creation.  Thanks.

Re: [Pysparse-users] Need your input: Major modernization; dropping legacy Python support?

[Oz N. T. <na...@gm...>]

Hi Dwayne and everyone,

>1. How many of you would really care if PyCrypto 2.6 was that last
>    version to support legacy versions of Python?  By "legacy", I mean
>   all versions of Python that are NOT one of these:

>     - Python 2.6.x
>     - Python 2.7.x
>    - Python 3.3 and above.

>    I'd continue to make bugfix releases of PyCrypto 2.6.x, but add no
>    more substantial new features.

+1 for that, assuming it will make other efforts in PyCrypto easier.

>2. I'm thinking of pulling in additional dependencies (e.g. cffi),
>    requiring setuptools, and basically joining what the rest of the
>   Python community is doing in 2013.

Like people said before me, it seems very immature, and Cython is not a bad
alternative at all.
There is a lot of know how out there with Cython. I would be willing to
help with the efforts.


>3. What if src/*.c were removed, and any relevant C code moved into an
>    independent library, which could be loaded using cffi?  (This is
>   basically what we need to do to support PyPy properly.)

An independent library, sounds good. But it seems that Cython supports pypy
too.

>4. What if Crypto.* became a wrapper around some other crypto library?

Why ? what is the benefit here?

>5. The Apache License 2.0.  What if PyCrypto were licensed under it, or
>    included dependencies that are licensed under it?

As the maintainer of a small project and humble project (pwman3) that uses
pycrypto, and it is
licensed unde GPL I would have a problem with it. I think the better choice
would be LGPL, as said
also before me.

I would love to have support for pypi, pip and setuptools, until now
installing pycrypto
was always a pain.

Best Regards,
Oz

[Pysparse-users] Mag Gam

[Mag G. <mag...@gm...>]

http://houtwormbestrijding-houtwormbestrijding.nl/mlwoeh/ibivodpmj.ikuklorzxzgycwtj





Mag Gam


7/21/2013 7:24:11 AM

Re: [Pysparse-users] PySparse HEAD does not build with SuiteSparse 4.2.0

[Jonathan G. <gu...@ni...>]

On May 3, 2013, at 10:15 AM, Jonathan Guyer <gu...@ni...> wrote:

> I am attempting to 
> 
>  pip install git+git://pysparse.git.sourceforge.net/gitroot/pysparse/pysparse#egg=pysparse
> 
> into a homebrew environment under Mac OS X Mountain Lion. (`pip install pysparse` is no better).
> 
> 
> I get three errors:


I can now successfully pip install from the git repository. I was actually able to get it working a few days after I posted by using:

  CFLAGS="" pip install git+git://pysparse.git.sourceforge.net/gitroot/pysparse/pysparse#egg=pysparse

This issue (at the time) was that some of the headers from suite-sparse were being found instead of the ones included with pysparse.

After several reconstructions of my Homebrew installation (for other reasons) I find I no longer need to pass `CFLAGS=""`. I don't know why it used to be necessary and now is not.

----

What does *not* work is 

  pip install pysparse

which raises

    cc: pysparse/sparse/src/spmatrixmodule.c
    pysparse/sparse/src/spmatrixmodule.c:4:10: fatal error: 'spmatrix.h' file not found
    #include "spmatrix.h"

The tarball at https://pypi.python.org/pypi/pysparse is incomplete (missing pysparse/includes is the one that raises the errors, but there is a bunch of other stuff missing) and the tarball does not seem to correspond to any revision in the git repository, e.g., pysparse/direct/setup.py has a number of uses of

  os.path.join(cwd,...

that do not appear (specifically the inclusion of `cwd` in the join) anywhere in the git history.

[Pysparse-users] PySparse HEAD does not build with SuiteSparse 4.2.0

[Jonathan G. <gu...@ni...>]

I am attempting to 

  pip install git+git://pysparse.git.sourceforge.net/gitroot/pysparse/pysparse#egg=pysparse

into a homebrew environment under Mac OS X Mountain Lion. (`pip install pysparse` is no better).


I get three errors:

pysparse/direct/superlu/src/colamd.c:547:12: error: conflicting types for 'colamd_recommended'
PUBLIC int colamd_recommended   /* returns recommended value of Alen. */
           ^
/Users/guyer/.homebrew/include/colamd.h:140:8: note: previous declaration is here
size_t colamd_recommended       /* returns recommended value of Alen, */
       ^
pysparse/direct/superlu/src/colamd.c:708:12: error: conflicting types for 'colamd'
PUBLIC int colamd               /* returns TRUE if successful */
           ^
/Users/guyer/.homebrew/include/colamd.h:166:5: note: previous declaration is here
int colamd                      /* returns (1) if successful, (0) otherwise*/
    ^
pysparse/direct/superlu/src/colamd.c:815:8: error: use of undeclared identifier 'COLAMD_JUMBLED_COLS'
    A [COLAMD_JUMBLED_COLS] = init_result ;
       ^


I have SuiteSparse 4.2.0 [*], but got the same errors before I upgraded from SuiteSparse 3.6.1.





[*] requirements.txt says 'OSX users using homebrew can simply say "brew install suitesparse".', but this does not appear to (any longer) be true. Homebrew presently has no suitesparse formula. What does work is:

  brew tap homebrew/science
  brew install suite-sparse

Re: [Pysparse-users] Memory leak

[Daniel W. <dan...@gm...>]

On Thu, Apr 18, 2013 at 4:23 AM, Nicola Creati <nc...@in...> wrote:

> Hello,
> I found a memory leak using pysparse if I create and fill an array
> several times. At each step the memory increases. I attached an example
> script that reproduces the problem.
>

Thanks for this and for the example. I'll try and look into it at some
point and get back to you.

-- 
Daniel Wheeler

[Pysparse-users] Memory leak

[Nicola C. <nc...@in...>]

Hello,
I found a memory leak using pysparse if I create and fill an array 
several times. At each step the memory increases. I attached an example 
script that reproduces the problem.
Thanks.

Cheers.

Nicola


from pysparse import spmatrix

def do():
     A = spmatrix.ll_mat(256**2, 256**2)
     for x in range(256):
         for y in range(256):
             k = y*256+x
             A[k,k] = 1
     return A

if __name__ == '__main__':
     for i in range(1024):
         b= do()


-- 
_____________________________________________________________________
Nicola Creati
Istituto Nazionale di Oceanografia e di Geofisica Sperimentale - OGS
IRI (Ricerca Teconologica e Infrastrutture) Department
B.go Grotta Gigante - Brisciki 42/c
34010 Sgonico - Zgonik (TS) - Italy
Tel. +39-0402140213
Fax  +39-040327307

Re: [Pysparse-users] [Pysparse-developers] pysparse release

[Daniel W. <dan...@gm...>]

On Thu, May 17, 2012 at 12:33 PM, Dominique Orban
<dom...@gm...> wrote:
> On Thu, May 17, 2012 at 12:18 PM, Daniel Wheeler
> <dan...@gm...> wrote:
>> Is it time for a new release? Dominique has done a lot of work since
>> 1.1.1 if I remember rightly. I believe there was something holding up
>> a release a year or two back, but I can't remember what that was.
>
> The last thing I did is use the version of UMFPACK that comes with
> SuiteSparse. The code is in the suitesparse branch. My hope was to
> pave the way to interfacing other factorizations (CHOLMOD and
> SuiteSparseQR are my two targets), but time is short.
>
> Also I'm wondering if we should phase out SuperLU. I'm not having much
> success with it. Its only advatange is to run in parallel. Just a
> notion.

I assume you mean that you would remove pysparse/direct/superlu and
only have pysparse/direct/umfpack. I'll leave that up to you. The only
requirement that I have right now is that the code snippet at then end
of  pysparse/__init__.py still works in some way that doesn't break
everything in fipy. Probably, the best thing to do is to deprecate it
now and remove it sometime in the future (or make it work with UMFPACK
and give a warning that superlu is not using SuperLU but something
else to do LU). The code snippet is this:

{{{
class _superlu:
    @Deprecated('Use pysparse.direct.superlu.factorize instead.')
    def factorize(self, *args, **kwargs):
        import pysparse.direct.superlu
        self.factorizeFnc = pysparse.direct.superlu.factorize
        return self.factorizeFnc(*args, **kwargs)

superlu = _superlu()
}}}

so that

{{{
from pysparse import superlu
}}}

still works. Actually, I need to switch fipy over to the new import
system of pysparse. I'll do that today or tomorrow while it is on my
mind.

Cheers

-- 
Daniel Wheeler

Re: [Pysparse-users] [Pysparse-developers] pysparse release

[Dominique O. <dom...@gm...>]

On Thu, May 17, 2012 at 12:18 PM, Daniel Wheeler
<dan...@gm...> wrote:
> Is it time for a new release? Dominique has done a lot of work since
> 1.1.1 if I remember rightly. I believe there was something holding up
> a release a year or two back, but I can't remember what that was.

The last thing I did is use the version of UMFPACK that comes with
SuiteSparse. The code is in the suitesparse branch. My hope was to
pave the way to interfacing other factorizations (CHOLMOD and
SuiteSparseQR are my two targets), but time is short.

Also I'm wondering if we should phase out SuperLU. I'm not having much
success with it. Its only advatange is to run in parallel. Just a
notion.

Cheers,

-- 
Dominique

[Pysparse-users] pysparse release

[Daniel W. <dan...@gm...>]

Is it time for a new release? Dominique has done a lot of work since
1.1.1 if I remember rightly. I believe there was something holding up
a release a year or two back, but I can't remember what that was.

Cheers

-- 
Daniel Wheeler

Re: [Pysparse-users] reading the source code of pysparse

[Oz N. T. <na...@gm...>]

Hi Dominique,

Just one more comment the examples in
https://github.com/dpo/pykrylov don't work with the latest official
pysprase (1.1.1)

neither from source or from debian packages.
I does work with the latest pysparse from git on sf.net.  I think it
would be easier if
it was mentioned.

Cheers,
Oz

Re: [Pysparse-users] reading the source code of pysparse

[Oz N. T. <na...@gm...>]

>> On Sat, Feb 11, 2012 at 6:57 PM, Dominique Orban
>> <dom...@gm...> wrote:
>>

Hi Dominique,

>
> What's so confusing about it?
>

The confusing part is that I need to pass the matrix via lambda. Is there any
other way to do it?

Thanks,
Oz

Re: [Pysparse-users] reading the source code of pysparse

[Dominique O. <dom...@gm...>]

On Sat, Feb 11, 2012 at 19:15, Oz Nahum Tiram <na...@gm...> wrote:
> On Sat, Feb 11, 2012 at 6:57 PM, Dominique Orban
> <dom...@gm...> wrote:
>
> Hi Dominique,
> Thanks for the answer. A few more questions.
>
>
>> This is a macro that adjusts the call to various types of Fortran
>> compilers (this particular call is a call to the BLAS library, which
>> is written in Fortran). Some Fortran compilers add a trailing
>> underscore to symbols, some add two, some add none, etc.
>>
>
> So this macro is found in BLAS? not in pysprase itself ?

The function dnrm2 is in the BLAS. The macro F77 is in
pysparse/include/fortran.h

>> I wouldn't recommend coding bicgstab all over again. The only costly
>> operations in Bi-CGSTAB (and other Krylov-type methods) are vector
>> operations (most often, addition of vectors) and operator-vector
>> products (e.g., A*x or A.T*x). I would say that to speed things up,
>> you'll want to speed up your operator-vector operations; they are the
>> dominant cost.
>>
>> You can take a look at PyKrylov (https://github.com/dpo/pykrylov)
>> which contains a pure Python implementation of Bi-CGSTAB and allows
>> you to input your operator in different ways (a Pysparse matrix being
>> one of them). For instance, you could implement your operator in C or
>> in Cython and that should speed things up. I believe that is the way
>> to go.
>
> Why is pure python Bi-CGSTAB should be faster than the one writen in C
> contained in pysparse?

It shouldn't. I'm saying that there isn't much to gain by recoding it
in C because the dominant cost is in the operator-vector products.
That's what should be fast.

>> Of course, vector operations in PyKrylov could also be speeded up with
>> Cython. That's been on my list for a while.
>
> You mean doing stuff on numpy vectores instead of for loops? I didn't see
> so many of these in the code ...

Or simply declaring types.

> And one last thing, in pysparse solvers, I pass a Matrix, initial x vector,
> and b vector (representing my boundary conditions.) I was quite
> confused to see the example here:
> https://github.com/dpo/pykrylov/blob/master/examples/bmark.py
>
>    for KSolver in [CGS, TFQMR, BiCGSTAB]:
>        ks = KSolver( lambda v: A*v,
>                      #precon = dp,
>                      #verbose=False,
>                      reltol = 1.0e-8
>                      )
>        ks.solve(rhs, guess = 1+np.arange(n, dtype=np.float), matvec_max=2*n)
>
> you are not assigning the lambda v to anything? What does this do?

lambda is standard Python. The lambda function is stored inside the
object upon instantiation. This syntax lets you solve several systems
with the same operator and different right-hand sides, or from
different initial guesses.

> solveing is done only with rhs and guess ? is this because A is
> already contained
> in ks ? I'd be happy to see an example which does not confuse me so much :-)

What's so confusing about it?

-- 
Dominique

Re: [Pysparse-users] reading the source code of pysparse

[Oz N. T. <na...@gm...>]

On Sat, Feb 11, 2012 at 6:57 PM, Dominique Orban
<dom...@gm...> wrote:

Hi Dominique,
Thanks for the answer. A few more questions.


> This is a macro that adjusts the call to various types of Fortran
> compilers (this particular call is a call to the BLAS library, which
> is written in Fortran). Some Fortran compilers add a trailing
> underscore to symbols, some add two, some add none, etc.
>

So this macro is found in BLAS? not in pysprase itself ?


> I wouldn't recommend coding bicgstab all over again. The only costly
> operations in Bi-CGSTAB (and other Krylov-type methods) are vector
> operations (most often, addition of vectors) and operator-vector
> products (e.g., A*x or A.T*x). I would say that to speed things up,
> you'll want to speed up your operator-vector operations; they are the
> dominant cost.
>
> You can take a look at PyKrylov (https://github.com/dpo/pykrylov)
> which contains a pure Python implementation of Bi-CGSTAB and allows
> you to input your operator in different ways (a Pysparse matrix being
> one of them). For instance, you could implement your operator in C or
> in Cython and that should speed things up. I believe that is the way
> to go.

Why is pure python Bi-CGSTAB should be faster than the one writen in C
contained in pysparse?


> Of course, vector operations in PyKrylov could also be speeded up with
> Cython. That's been on my list for a while.

You mean doing stuff on numpy vectores instead of for loops? I didn't see
so many of these in the code ...

And one last thing, in pysparse solvers, I pass a Matrix, initial x vector,
and b vector (representing my boundary conditions.) I was quite
confused to see the example here:
https://github.com/dpo/pykrylov/blob/master/examples/bmark.py

    for KSolver in [CGS, TFQMR, BiCGSTAB]:
        ks = KSolver( lambda v: A*v,
                      #precon = dp,
                      #verbose=False,
                      reltol = 1.0e-8
                      )
        ks.solve(rhs, guess = 1+np.arange(n, dtype=np.float), matvec_max=2*n)

you are not assigning the lambda v to anything? What does this do?

solveing is done only with rhs and guess ? is this because A is
already contained
in ks ? I'd be happy to see an example which does not confuse me so much :-)

Thanks again for the answers and the effort invested in this!

Oz

Re: [Pysparse-users] reading the source code of pysparse

[Dominique O. <dom...@gm...>]

On Sat, Feb 11, 2012 at 04:30, Oz Nahum Tiram <na...@gm...> wrote:
> Hi,
>
> First, thanks for the great work on this peace of software. I have
> been using it for more than 3 years now and I very happy.
> I have a few question and I hope this is not
> I am still very new to C programming and I was wondering what does the
> following statement for expamle do:
>
> n2b = F77(dnrm2)(&n, b, &ONE);/* Norm of rhs vector, b */
>
> Is this some kind of type casting ?

This is a macro that adjusts the call to various types of Fortran
compilers (this particular call is a call to the BLAS library, which
is written in Fortran). Some Fortran compilers add a trailing
underscore to symbols, some add two, some add none, etc.

> The second question is, now that I have a working application with
> pysparse that I'd like to accelerate,
> I am thinking of rewriting my code in C. I know there are C solvers
> for bicgstab, but I was wondering how
> much work is it just to convert the bicgstab.c from pysparse to plain
> C without the python dependencies.
> Are there any hints here?

I wouldn't recommend coding bicgstab all over again. The only costly
operations in Bi-CGSTAB (and other Krylov-type methods) are vector
operations (most often, addition of vectors) and operator-vector
products (e.g., A*x or A.T*x). I would say that to speed things up,
you'll want to speed up your operator-vector operations; they are the
dominant cost.

You can take a look at PyKrylov (https://github.com/dpo/pykrylov)
which contains a pure Python implementation of Bi-CGSTAB and allows
you to input your operator in different ways (a Pysparse matrix being
one of them). For instance, you could implement your operator in C or
in Cython and that should speed things up. I believe that is the way
to go.

Of course, vector operations in PyKrylov could also be speeded up with
Cython. That's been on my list for a while.

I hope this helps.

-- 
Dominique

[Pysparse-users] reading the source code of pysparse

[Oz N. T. <na...@gm...>]

Hi,

First, thanks for the great work on this peace of software. I have
been using it for more than 3 years now and I very happy.
I have a few question and I hope this is not
I am still very new to C programming and I was wondering what does the
following statement for expamle do:

n2b = F77(dnrm2)(&n, b, &ONE);/* Norm of rhs vector, b */

Is this some kind of type casting ?


The second question is, now that I have a working application with
pysparse that I'd like to accelerate,
I am thinking of rewriting my code in C. I know there are C solvers
for bicgstab, but I was wondering how
much work is it just to convert the bicgstab.c from pysparse to plain
C without the python dependencies.
Are there any hints here?


Thanks in advance,

Oz

Re: [Pysparse-users] problem importing libifport.so.5

[Dominique O. <dom...@gm...>]

On Thu, Feb 9, 2012 at 20:39,  <av...@fa...> wrote:

> Traceback (most recent call last):
>  File "CourseA.py", line 2, in <module>
>    from ase import *
>  File "/usr/lib/python2.7/dist-packages/ase/__init__.py", line 7, in
> <module>
>    from ase.atom import Atom
>  File "/usr/lib/python2.7/dist-packages/ase/atom.py", line 3, in <module>
>    import numpy as np
>  File "/usr/local/lib/python2.7/dist-packages/numpy/__init__.py", line
> 137, in <module>
>    import add_newdocs
>  File "/usr/local/lib/python2.7/dist-packages/numpy/add_newdocs.py", line
> 9, in <module>
>    from numpy.lib import add_newdoc
>  File "/usr/local/lib/python2.7/dist-packages/numpy/lib/__init__.py",
> line 13, in <module>
>    from polynomial import *
>  File "/usr/local/lib/python2.7/dist-packages/numpy/lib/polynomial.py",
> line 17, in <module>
>    from numpy.linalg import eigvals, lstsq
>  File "/usr/local/lib/python2.7/dist-packages/numpy/linalg/__init__.py",
> line 48, in <module>
>    from linalg import *
>  File "/usr/local/lib/python2.7/dist-packages/numpy/linalg/linalg.py",
> line 23, in <module>
>    from numpy.linalg import lapack_lite
> ImportError: libifport.so.5: cannot open shared object file: No such file
> or directory

What makes you think that any of this has anything to do with
pysparse? It looks like it's your numpy configuration that needs to be
checked.

-- 
Dominique