Skip to Content.
Sympa Menu

coq-club - Re: [Coq-Club] Multi-core execution?

coq-club AT inria.fr

Subject: The Coq mailing list

List archive

Re: [Coq-Club] Multi-core execution?


Chronological Thread 
  • From: Adam Chlipala <adamc AT csail.mit.edu>
  • To: Pierre-Marie Pédrot <pierre-marie.pedrot AT inria.fr>
  • Cc: coq-club <coq-club AT inria.fr>
  • Subject: Re: [Coq-Club] Multi-core execution?
  • Date: Fri, 15 Mar 2013 10:50:15 -0400

On 03/11/2013 09:51 PM, Pierre-Marie Pédrot wrote:
I've quickly hacked a proof of concept using Unix.fork and darkest magic to bypass this type of problem. The tactic sends back the proof term through a pipe using OCaml Marshal module. This is ugly and really unsafe, but this should work.

Currently, the tactic only displays the list of proof-term as a proof-of-concept (no pun intended), so you're still having the problem of solving the goals. Nevertheless one should be able to instantiate a goal with this list, but I still don't know how to instantiate all evars representing a goal at once (any avice, Arnaud?).

Nifty! I'd love to see an evolution of this so that I can use it my scenario. I imagine it would be worth throttling the rate at which processes are forked, based on how many cores are available.



Archive powered by MHonArc 2.6.18.

Top of Page