[racket] Engineering Tradeoffs of ANF transforms and the Stateless Server

From: Neil Van Dyke (neil at neilvandyke.org)
Date: Sun Jan 1 17:33:48 EST 2012

Jay McCarthy wrote at 01/01/2012 04:34 PM:
> On Sat, 31 Dec 2011 18:37:03 -0500,Neil Van Dyke<neil at neilvandyke.org>  mumbled:
>> Maybe some student wants to see whether any big wins can be gotten for
>> keeping continuations in addressable virtual memory, if they can tweak
>> the app, Racket compiler, Racket VM and GC, Racket Web Server, and Linux.
> Emery Berger has already produced the "swap-sensitive" GC which would
> transparently improve things like this for Racket and many other
> languages. However, the Linux kernel maintainers didn't take the patch
> and I believe Emery moved.

Many people run production servers with patched kernels.  And distros 
also are known to provide alternative kernel builds, in addition to 
separating out some kernel modules into separate distro packages.

That said, if someone has a big win for GC'd languages that needs 
support in the Linux kernel, I might be able to put them in touch with 
someone who can advise on how to shepherd it in.  It might have been 
rejected for perceived quality, for perceived effectiveness, or because 
it was not ``marketed'' the best way.

The kernel certainly has numerous esoteric modules and features already.


Posted on the users mailing list.