ffc team mailing list archive
-
ffc team
-
Mailing list archive
-
Message #00463
Re: Restructure computation of reference tensor to use less memory. [...]
-
To:
Discussion of FFC development <ffc-dev@xxxxxxxxxx>
-
From:
Matthew Knepley <knepley@xxxxxxxxxxx>
-
Date:
Tue, 06 Dec 2005 10:02:35 -0600
-
Cc:
Discussion of FIAT development <fiat-dev@xxxxxxxxxx>
-
In-reply-to:
<20051206155522.GA31303@galerkin> (Anders Logg's message of "Tue, 6 Dec 2005 09:55:22 -0600")
-
User-agent:
Gnus/5.1007 (Gnus v5.10.7) XEmacs/21.4 (Jumbo Shrimp, linux)
Anders Logg <logg@xxxxxxxxx> writes:
> On Tue, Dec 06, 2005 at 09:04:06AM -0600, Anders Logg wrote:
>
>> > Seems like Numarray is also going to cease in favour of SciPy.
>>
>> Yes, see http://numeric.scipy.org/.
>>
>> It looks like the array interface is already included in the SciPy
>> package that I have installed on my machine (Debian packages
>> python2.4-scipy 0.3.2-8 and python2.4-scipy-core 0.3.2-4).
>>
>> Perhaps I will try to update the next release of FFC (which might be
>> 0.3.0) to use SciPy instead of Numeric. Any objections to this?
>
> It's probably good if a switch to SciPy is coordinated with a similar
> move in FIAT, or otherwise there would still be dependencies to
> Numeric (and LinearAlgebra).
>
> Rob, would it be much work to switch FIAT to SciPy from Numeric and
> LinearAlgebra?
Actually, I think this would be a waste of time. At the last SciPy conference,
Travis announced he was submitting a PEP to try and get Numeric 3.0 directly
into the core. I think this is likely. Also, it is a separate package from
SciPy, which is quite large.
Matt
--
"Failure has a thousand explanations. Success doesn't need one" -- Sir Alec Guiness
Follow ups
References