Monday, June 3, 2013

NumPyPy status update

Hello everyone,

May was the first month I was paid to work on NumPyPy (thanks to all who donated!), here is what I worked on during this period :

  • It is now possible to use subarrays.
  • It is now possible to pickle ndarrays (including those using subarrays), dtypes and scalars, the pickling protocol is the same as numpy's.

For June, I plan to work on the nditer class, it seems that there's enough work for an entire month.

Cheers
Romain Guillebert
at

7 comments:

  1. AnonymousJune 3, 2013 at 7:49 PM

    What's numpypy's recommended way for a C/cffi extension to get a pointer to the data?

    Thanks,
    Andreas

    ReplyDelete
  2. AnonymousJune 4, 2013 at 9:37 AM

    Excellent work!

    ReplyDelete
  3. AnonymousJune 4, 2013 at 11:34 AM

    Thanks! But pickling sliced arrays doesn't work yet (tested with nightly build pypy-c-jit-64739-f556942951f9-linux):

    import cPickle as pickle
    import numpypy as numpy
    a = numpy.arange(10.)[::2]
    print a # [ 0. 2. 4. 6. 8.]
    p = pickle.dumps(a)
    print pickle.loads(p) # [ 0. 1. 2. 3. 4.] oops!

    ReplyDelete
  4. Romain GuillebertJune 4, 2013 at 8:55 PM

    @Anonymous

    Thanks for reporting it, it's fixed

    ReplyDelete
  5. AnonymousJune 4, 2013 at 10:16 PM

    Great to hear about the progress, keep up the good work!

    ReplyDelete
  6. AnonymousDecember 5, 2013 at 11:23 PM

    It is working very well for me, thanks!

    Now, is there any way to load the resulting pickle in cPython?

    numpy.save and numpy.load do work between pypy and cPython, but my arrays are embedded in larger data structures.

    The motivation is that I would like to run a numerical program and store some results, and then load the results and plot them with matplotlib (which does not work on pypy).

    Here is the error in cPython:

    >>> pickle.load(open('/tmp/x', 'r+b'))
    Traceback (most recent call last):
    File "", line 1, in
    File "/usr/lib/python2.7/pickle.py", line 1378, in load
    return Unpickler(file).load()
    File "/usr/lib/python2.7/pickle.py", line 858, in load
    dispatch[key](self)
    File "/usr/lib/python2.7/pickle.py", line 1090, in load_global
    klass = self.find_class(module, name)
    File "/usr/lib/python2.7/pickle.py", line 1124, in find_class
    __import__(module)
    ImportError: No module named _numpypy.multiarray

    ReplyDelete
  7. AnonymousDecember 5, 2013 at 11:28 PM

    It is working great, thanks!

    Now, is there any way to load the resulting pickle in cPython?

    numpy.save and numpy.load do work between pypy and cPython, but my arrays are embedded in larger data structures.

    The motivation is that I would like to run a numerical program and store some results, and then load the results and plot them with matplotlib (which does not work on pypy).

    Here is the error in cPython:

    >>> pickle.load(open('/tmp/x', 'r+b'))
    Traceback (most recent call last):
    File "", line 1, in
    File "/usr/lib/python2.7/pickle.py", line 1378, in load
    return Unpickler(file).load()
    File "/usr/lib/python2.7/pickle.py", line 858, in load
    dispatch[key](self)
    File "/usr/lib/python2.7/pickle.py", line 1090, in load_global
    klass = self.find_class(module, name)
    File "/usr/lib/python2.7/pickle.py", line 1124, in find_class
    __import__(module)
    ImportError: No module named _numpypy.multiarray

    ReplyDelete

See also PyPy's IRC channel: #pypy at freenode.net, or the pypy-dev mailing list.
If the blog post is old, it is pointless to ask questions here about it---you're unlikely to get an answer.