[jruby-dev] [ANN] New Truffle/Graal Backend

[Chris Seaton]

I’m very happy to announce that today there is a new backend in JRuby,
based on two new JVM technologies being researched by Oracle Labs. It’s
called the Truffle backend as it uses a framework for implementing
languages called Truffle, and is designed to work with a new JVM JIT
compiler called Graal. Documentation linked below gives the technical
details and pointers to more information, but in general the goal is that
the Truffle backend is significantly faster and simpler than the current
backends as it uses a more expressive interface to the JVM than is possible
by emitting bytecode.

We’ve been talking to Charlie Nutter and Thomas Enebo about this work for
some time now, and we’re creating this new branch after they’ve told us
that it’s welcome in the repository.

The code is in a branch called ‘truffle’ -
https://github.com/jruby/jruby/tree/truffle.

There is a blog post with more general background information at
http://blog.jruby.org/2014/01/truffle_graal_high_performance_backend/, and
there is a wiki page with some details of how to get started with the code
at https://github.com/jruby/jruby/wiki/Truffle.

I’m Chris Seaton - I work for Oracle Labs part-time alongside my PhD at the
University of Manchester in the UK. I’m now going to be working on the
Truffle backend within JRuby, so you can consider me a new paid contributor
to JRuby! I hope to be a valuable member of the community, so please feel
more than free to get in touch with me.

Chris

@ChrisGSeaton
chris-***@public.gmane.org
http://www.chrisseaton.com/

[Uwe Kubosch]

Welcome to JRuby, Chris! This sounds very interesting! Looking forward to trying it :-)

I’m very happy to announce that today there is a new backend in JRuby, based on two new JVM technologies being researched by Oracle Labs. It’s called the Truffle backend as it uses a framework for implementing languages called Truffle, and is designed to work with a new JVM JIT compiler called Graal. Documentation linked below gives the technical details and pointers to more information, but in general the goal is that the Truffle backend is significantly faster and simpler than the current backends as it uses a more expressive interface to the JVM than is possible by emitting bytecode.
We’ve been talking to Charlie Nutter and Thomas Enebo about this work for some time now, and we’re creating this new branch after they’ve told us that it’s welcome in the repository.
The code is in a branch called ‘truffle’ - https://github.com/jruby/jruby/tree/truffle.
There is a blog post with more general background information at http://blog.jruby.org/2014/01/truffle_graal_high_performance_backend/, and there is a wiki page with some details of how to get started with the code at https://github.com/jruby/jruby/wiki/Truffle.
I’m Chris Seaton - I work for Oracle Labs part-time alongside my PhD at the University of Manchester in the UK. I’m now going to be working on the Truffle backend within JRuby, so you can consider me a new paid contributor to JRuby! I hope to be a valuable member of the community, so please feel more than free to get in touch with me.

--
Uwe Kubosch
uwe-***@public.gmane.org
http://kubosch.no/





---------------------------------------------------------------------
To unsubscribe from this list, please visit:

http://xircles.codehaus.org/manage_email