Skip to content

run4flat/perl_nvcc

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Intro
=====

ExtUtils::nvcc is a CUDA compiler and linker wrapper for Perl's toolchain.

Modern compilers take a lot of options. Perl has learned to use many of these 
options. The problem is that nvcc, nVidia's compiler wrapper---which must be 
used to compile CUDA code---doesn't like most of those extra flags and options.
It has a special set of command-line arguments that allow you to pass compiler 
and linker flags on to the actual compiler and linker, but it does not send 
those options along by default. That makes life harder than it ought to be.

Hence ExtUtils::nvcc. The main goal of this project is to provide a compiler and 
linker acceptable to the common Perl build tools, Inline::C, ExtUtils::MakeMaker,
and Module::Build, which allows for the compilation of CUDA code without having
to modify the myriad of compiler and linker flags. It also provides some super-
simple functions for configuring the common build tools to use nvcc in their
build chain.


Installation
============

Thorough installation instructions can be found on the github wiki:

http://github.com/run4flat/perl_nvcc/wiki/Installing


Windows Restrictions
====================

First off, ExtUtils::nvcc doesn't yet work for Windows. However, if it ever
does (and I hope it does, actually), it will likely on work with Visual
Studio's C++ compiler.

The people at nVidia decided to only support gcc on Linux machines, and cl.exe
(MS Visual Studio's compiler) on Windows machines. In practice, this means that
ExtUtils::nvcc will not work properly on Strawberry Perl or Cygwin's perl. I'm
afraid you'll have to get a version of Perl compiled with Visual Studio, and
you'll need a copy of Visual Studio to compile nVidia code. The good news is
that Microsoft allows you to download a free copy of any single componenet of
the Visual Studio suite: you can download Visual C/C++ for free. You will likely
need to compile your own perl, but once that's in place, you'll have a fully
operating CPAN and ExtUtils::nvcc will (some day) work correctly.


Dependencies
============

Although you can use ExtUtils::nvcc for whatever purposes you may have, the 
underlying motivation for this work is for integration with the Perl toolchain.
You can use this module outside of the Perl toolchain (by wrapping the awkward
command-line invocation, perhaps), but it's really meant to operate with
ExtUtils::MakeMaker, Inline::C, and Module::Build.

Although not necessary for use, the installation requires Module::Build.


Test Reports
============

In the few months that ExtUtils::nvcc has been quietly sitting on CPAN, it
has received zero test reports. That's because there isn't a single automated
tester that has nvcc installed. If you try to install this module, please
report your success or failure at cpantesters. The process to set everything
up is a bit involved, but once in place it should be pretty easy to use and
your test reports will help to turn this piddly little thing into a useful
tool! You can read more here: L<http://wiki.cpantesters.org/wiki/QuickStart>.


Usage
=====

See InlineCUDA.pl for an example script, as well as the synopsis in
ExtUtils::nvcc


Bugs
====

I used to have an issue on Ubuntu with segmentation faults (hence the END block
in the example Inline script) but that seems to have gone.

The only outstanding bug at the moment is that Windows is not yet supported,
not even with cl.exe, in spite of great efforst. I am working on it and help
here would be appreciated.


Feedback Welcome
================

For ideas, suggestions, or anything else, feel free to write me. I've mangled
my email address, so remove the language that would seem inappropriate for a
perl developer:

David Mertens <[email protected]>

I'll be happy to move discussion of this project onto an appropriate list, such 
as PDL, Inline, Module::Build, or something else. I am open to suggestions.


Copyright
=========

All of this material is copyright David Mertens, 2010-2011, and is distributed
under the Artistic License. For more information, see 
http://dev.perl.org/licenses/artistic.html

About

A CUDA compiler and linker wrapper for Perl's toolchain.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published