"dataflow rewriting engine"

[Simon Peyton-Jones]

| > Has there been any thought about working with the LLVM project? I
| > didn't find anything on the wiki along those lines.
|
| I have only had a rather brief look at LLVM, but my understanding at
| the moment is that LLVM would not be able to support one of GHC's
| current code layout optimisations.

With the new, modular code generation framework, described here
http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/NewCodeGen
I think it'll be much easier [for someone else] to consider using LLVM.  I don't see the "data next to code" thing as a show-stopper; it used to be a performance win of a handful of percent, but I don't think we have recent data.  And an LLVM back end might be willing to give up that few percent.

There may well be other difficulties, of course.  If anyone is interested, the best thing is to wait for the new stuff to be committed to the HEAD (which will happen in Sept), and look at the new Cmm data types (which define the language you'd have to convert to LLVM).

Simon