On Thu, 15 Mar 2012 21:17:08 +1100, Vasily Chekalkin <[email protected]>
wrote:
Hello.
I have some idea about reordering of how PCC works with CallContexts.
At the moment CallContext created by Parrot_pcc_build_signature on
caller side and stored inside current CallContext. This is most
unnatural way of calling Subs and lead as to quite few issues:
1. We don't always need fresh CallContext on callee side. For example
C functions/methods doesn't have to have any.
2. Corollary of 1 is increased pressure on GC.
3. Corollary of 1 that Sub PMC (and inherited PMCs such as Coroutine,
Continuation, etc) can create CallContext and storage for registers
in
single call. Which is "suboptimal" (i.e. crappy slow).
I want to change it other way around:
1. Caller fills params in current CallContext.
2. Callee creates new CallContext (if needed).
3. Callee unpacks params from parent CallContext.
After implementing this we can optimize CallContext creation to
allocate storage for registers in single call.
All of this should give us
1. Some small performance boost (about 5% I think).
2. Will definitely speed up C METHODs calls. Which will help with
something like 6model.
3. But most important will make PCC much more sensible in terms of
allocating stuff. Which lead us to various optimizations like "Scheme
style" CallContext stack allocations with trivial bump pointer
allocation and bit of magic for capturing CallContext for
Continuations
Any comments, objections, volunteers to help with it?
Right now calling and returning from sub are symmetric - callee fills
the return values in callContext and jumps to return address. Then the
caller unpacks return values. How the proposed change would affect
this?
Best regards
--
Luben Karavelov
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev
