user-pic

David Mertens

Subscribe to feed Recent Actions from David Mertens

  • Aristotle commented on PDL features I'd like to see in Perl 6

    I think the new refaliasing feature (experimental in 5.22) will allow you to alias slices like so:

    \( @slice[ 0 .. 4 ] ) = \( @data[ 2, 3, 5, 7, 11 ] );

    Obviously it’s ugly to have to count out the LHS slice by hand:

    \( @slice[ 0 .. $#$_ ] ) = \( @data[ @$_ ] )
        for [2, 3, 5, 7, 11];

    That’s quite a mouthful, but at least you can package it up in a function:

    sub alias_slice {
        my ( $src, @dst ) = shift;
        \( @dst[ 0 .. $#_ ] ) = \( @$src[ @_ ] );
        \@dst;
    }
    
    

    \my @slice = al…

  • therflabs commented on PDL features I'd like to see in Perl 6

    Call me crazy, but I had a freelance gig converting Python numpy/scipy to Perl PDL... And barely knowing Python it was a lot easier to understand than PDL.

    I can see why numpy/scipy has gained a lot of traction.

    I'd like to see P6 PDL to be like numpy/scipy to ease crossover to Perl 6 and lower the entry level for beginners. In my opinion usability and clear documentation for the non-math scholar has hurt what PDL could have been and could be a lesson for PDL Perl 6.

  • PetaMem commented on What?! CUDA::Minimal... works?

    And here's this - with Perl 5.22, GCC 5.3.0, CUDA 7.5 anno domini 2016:

    dev@sol.petamem.com:/data/soft/perl-CUDA-Minimal
    # ./Build test
    t/00_load.t ............... ok   
    t/Index-Manipulation.t .... ok   
    t/Memory.t ................ ok   
    t/Transfer.t .............. ok     
    t/z_PDL.t ................. ok     
    t/z_kernel_invocations.t .. ok    
    All tests successful.
    Files=6, Tests=76,  1 wallclock secs ( 0.02 usr  0.01 sys +  0.21 cusr  0.74 csys =  0.98 CPU)
    Result: PASS
    

    You have to do some hacking

    /opt/cuda/host_config.h
    to remain silent about gcc …

  • Yuki Kimoto commented on SPVM is released! Perl maybe become much fast.

    SPVM is not C compiler.

    SPVM is compiler and runtime of SPVM bytecode.

    SPVM don't need any C compiler to run SPVM module.

    SPVM module load is fast and easy.

    If I need more performance, I maybe use libjit to compile bytecodes to machine code.


  • Yuki Kimoto commented on SPVM is released! Perl maybe become much fast.

    > you're making it very easy to call from Perl.

    I'm happy you have found the advantage of bytecode.
    Easy to use is one of the big goals.

    >Have you toyed around with any benchmarks compared with pure Perl? Could I try writing some C::Blocks versions to compare?

    I don't do any benchmark yet because SPVM specification has yet something bugs. I must fix it.

    I'm happy if you do benchmark in current status of SPVM. Maybe array loop is very fast than Perl because
    arrays are present in consecutive memory areas.

Subscribe to feed Responses to Comments from David Mertens

About blogs.perl.org

blogs.perl.org is a common blogging platform for the Perl community. Written in Perl and offering the modern features you’ve come to expect in blog platforms, the site is hosted by Dave Cross and Aaron Crane, with a design donated by Six Apart, Ltd.