Mac OS X works* for me if I build with python 2.6 but then run with 2.7. There 
is one message on startup complaining about a version mismatch but everything 
appears to run fine. I will write this up someday, really I will.


-          Kevin

* so far as I can tell, though I haven't done rigorous testing**. I only tested 
on hotgym and I only looked at the text output (slightly tweaked from Scott's 
original code. The output isn't an exact match to the corresponding output from 
python 2.6 on a CentOS VM, but so far as I could tell the difference was just 
that 2.7 was doing a better job of rounding when converting floating point to 
text.

** numerical regression tests would be really nice, maybe criticial if there is 
to be an ARM port. This is nontrivial because, well, floating point 
arithmetic***. I'm not even sure what the right comparison model is, but 
probably one would want to compare means and variances at least.

*** an example of the nightmare that is validation of floating point 
computations: 
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0038234

From: nupic [mailto:[email protected]] On Behalf Of Erik Blas
Sent: Monday, July 15, 2013 11:50 AM
To: NuPIC general mailing list.
Subject: Re: [nupic-dev] Linux Version/Distribution

On linux, pyenv and pythonbrew seem to solve the build problem, regardless of 
python version. Only platform that's giving me trouble still is OS X.

On Mon, Jul 15, 2013 at 9:44 AM, Scott Purdy 
<[email protected]<mailto:[email protected]>> wrote:
Sounds similar to https://issues.numenta.org/browse/NPC-158 (if you look at the 
pastebin, there is a multiarray import error).

On Thu, Jun 27, 2013 at 6:09 AM, Laurent Julliard 
<[email protected]<mailto:[email protected]>> wrote:
I do confirm that it's much more convenient to target  a distro with Python 
2.6. I run Fedora 16 myself which comes standard with Python 2.7. I had to 
recompile Python 2.6 by myself and although everythig semmed to work well I was 
unable to make it work with CLA (error on undefined symbol in multiarray.so). 
HAs anybody experienced the same problem ?

Laurent

On 06/26/2013 12:40 PM, Marek Otahal wrote:
Hi Patric,
or otherwise any linux distro where python 2.6 (hope 2.7 soon too) is easily 
available will do for you. The requirements are quite basic.

On Wed, Jun 26, 2013 at 4:46 AM, Patrick Higgins 
<[email protected]<mailto:[email protected]>> wrote:
I'm setting up a Linux box for working
with the CLA and we're wondering what
distribution of Linux is recommended.
We'd like build the same environment used
for NuPIC to avoid complications.

Once I get that going I'd like build it in
OS X, but want first to get it going as it
was developed.


Thanks!



Patrick




_______________________________________________
nupic mailing list
[email protected]<mailto:[email protected]>
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org



--
Marek Otahal :o)


_______________________________________________

nupic mailing list

[email protected]<mailto:[email protected]>

http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org


_______________________________________________
nupic mailing list
[email protected]<mailto:[email protected]>
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org


_______________________________________________
nupic mailing list
[email protected]<mailto:[email protected]>
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org


________________________________

The material in this message is private and may contain Protected Healthcare 
Information (PHI). If you are not the intended recipient, be advised that any 
unauthorized use, disclosure, copying or the taking of any action in reliance 
on the contents of this information is strictly prohibited. If you have 
received this email in error, please immediately notify the sender via 
telephone or return mail.
_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to