[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: About octave performance...
From: |
Jaroslav Hajek |
Subject: |
Re: About octave performance... |
Date: |
Wed, 30 Jul 2008 14:26:39 +0200 |
On Wed, Jul 30, 2008 at 2:10 PM, Bill Denney <address@hidden> wrote:
> Jaroslav Hajek wrote:
>>
>> Matlab's interpreter now has a JIT compiler, which can
>> improve the speed of some m-file code a lot. However, I daresay that
>> Octave has significantly better support for compiled (C++) files via
>> its object-oriented internal API, so the demand for JIT in Octave is
>> not that high.
>>
>
> I would actually disagree with the above assertion. There are a rather
> large number of octave users who will never touch C++ and will only work via
> .m files, and for us, speed of interpreted code (via JIT or vectorization or
> any other mechanism) can be important.
>
I think speed of interpreted code is important for most users of
Octave. I was merely expressing my opinion that Octave's oct-files are
easier to use than mex (despite the lack of docs), and thus if you are
forced to meddle with compiled code, it will be less painful. The C++
code can be very similar to m-code decorated with type declarations.
> Have a good day,
>
> Bill
>
--
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz