[LLVMdev] LLVM capability question.

Ralph Corderoy ralph at inputplus.co.uk
Thu Dec 21 06:18:10 PST 2006


Michael T. Richter wrote:
> 4. GHC relies an awful lot on having many, many, many "little stacks".
> Is it possible to have lots of tiny stacks in LLVM's architecture, to
> switch between them easily, to check when they overflow and to
> automatically grow them when they exceed their bounds?

I was also wondering about this from the point of view of using channels
to communicate between tasks, where only one task runs at a time and it
runs until it blocks.  (Yes, I'm being influenced by CSP and Bell Labs'
Squeak, Newsqueak, Alef, Limbo, etc.)



More information about the llvm-dev mailing list