Beefy Boxes and Bandwidth Generously Provided by pair Networks
more useful options
 
PerlMonks  

Re^2: Putting Perl Back on Top in the Fields of Scientific and Financial Computing

by chm (Novice)
on Mar 07, 2011 at 19:22 UTC ( #891888=note: print w/replies, xml ) Need Help??


in reply to Re: Putting Perl Back on Top in the Fields of Scientific and Financial Computing
in thread Putting Perl Back on Top in the Fields of Scientific and Financial Computing

PDL is pretty slow, also, compared to other available software for doing what it doesn't provide the speed, ease of use of other pieces of software.

I've been a PDL developer for a few years now and am the current release manager. This is the first time I've heard any reports that "PDL is pretty slow". I would appreciate references and benchmarks for this.

Also, a weird thing is that PDL depends on OpenGL, so on a machine without X you can't actually do anything with PDL (or you can but it takes you some additional $amount_of_time). That is not normal since machines that only crunch numbers needn't have X on them.

OpenGL is not X and X is not OpenGL. We're moving the baseline graphics capability to one based on OpenGL because it is the best common denominator for 2D and 3D display across all the major perl platforms.

Also, how portable is PDL actually? Haven't tried it on windows.

See the CPAN Testers matrix for the PDL-2.4.7 stable release at http://matrix.cpantesters.org/?dist=PDL+2.4.7. For current and accurate information on PDL, I refer you to our web site at http://pdl.perl.org

  • Comment on Re^2: Putting Perl Back on Top in the Fields of Scientific and Financial Computing

Replies are listed 'Best First'.
Re^3: Putting Perl Back on Top in the Fields of Scientific and Financial Computing
by spx2 (Deacon) on Mar 07, 2011 at 20:41 UTC
    well, I would benchmark it against LAPACK, but I don't have the time, but if you want, you can do the benchmark and show me that I'm wrong :)

    As a matter of fact, I would like to see some benchmarks myself, who knows, maybe I'm completely wrong(I saw most of the PDL code is C code and the Perl is just calling that).

    the main problem is if people would use Perl+PDL for scientific/financial computing, but if they decide they won't use it, it doesn't matter if it's fast or slow. I think it is slower than LAPACK which doesn't necessarily make it very slow, just slower than LAPACK ..
      well, I would benchmark it against LAPACK, but I don't have the time, but if you want, you can do the benchmark and show me that I'm wrong :)
      So, you haven't really benchmarked it against LAPACK. Yet, you imply that PDL is slower than not just one (LAPACK) but many other similar tools. You want to be shown you are wrong. Why don't you make the effort and show that what you are asserting is indeed true.

      Fwiw, I do use Perl/PDL, and I don't use LAPACK, so it doesn't matter how fast LAPACK is.



      when small people start casting long shadows, it is time to go to bed
        I don't make the effort because I don't have time, I'll stick to assertions in the meantime :)

        But you can make some benchmarks, not very complicated.

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://891888]
help
Chatterbox?
[marto]: good morning all
[Corion]: Hi marto!
[Corion]: Oooh - finally, I can say no lazy; :)
[Corion]: But it's another The::Net incarnation, too sad ;)
[Corion]: WTF? There is a dev release of Danga::Socket?! I thought that module died a long time ago?! According to its changes, the last release was in 2008 ;)

How do I use this? | Other CB clients
Other Users?
Others scrutinizing the Monastery: (9)
As of 2018-07-17 08:10 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?
    It has been suggested to rename Perl 6 in order to boost its marketing potential. Which name would you prefer?















    Results (359 votes). Check out past polls.

    Notices?