Subject: | |
From: | |
Reply To: | |
Date: | Mon, 3 Jul 2006 08:48:57 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Andy Buckley wrote:
> Paul F. Kunz wrote:
>
>>>>>>> On Fri, 30 Jun 2006 16:17:14 -0500, Troy Dawson <[log in to unmask]>
>>>>>>> said:
>
>
> [re. numarray, numeric, numpy]
>
>>> So it sounds like, I could put in all three, and whichever ones
>>> users want, they can install.
>>
>>
>> I think so. I've installed Numeric and numarray on the same
>> machine and there was no conflict. I imagine it is the same for
>> numpy but haven't yet tried.
>
>
> I'm not quite sure what incompatibilities there are, but numpy appears
> to be intended as the dominant version from now on. See
> http://scipy.org/NumPy for information. There's a module in NumPy which
> will apparently convert existing Python Numeric code to work with NumPy.
>
> Having NumPy, SciPy (www.scipy.org) and matplotlib (matplotlib.sf.net)
> in SL would be very useful: they're certainly becoming widespread
> numerical processing packages these days.
>
> Andy
Hi Andy,
I have tried my best to get scipy to recompile, but it depends on Atlas,
which I have NOT been able to get to recompile. Maybe I'm grabbing the
wrong version of Atlas, but scipy, I am quite sure I'm just getting the
generic version.
I've seen matplot go by a couple times in conversations. Any good
places for an rpm for that?
Troy
--
__________________________________________________
Troy Dawson [log in to unmask] (630)840-6468
Fermilab ComputingDivision/CSS CSI Group
__________________________________________________
|
|
|