[gecode-users] buggy parallel search

Christian Schulte cschulte at kth.se
Fri Feb 5 11:04:36 CET 2010


Hi Denys,

the assertion is correct: this function is called by a worker that has
successfully stolen new work for search. As there is at least one other busy
worker (the worker from which the work got stolen) it is really meant to be
> 0.

Alas, it looks as if there is a bug. Which engine? Which problem? How many
cores? Aaargh...

Cheers
Christian

--
Christian Schulte, www.ict.kth.se/~cschulte/


-----Original Message-----
From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Denys Duchier
Sent: Thursday, February 04, 2010 8:22 PM
To: users at gecode.org
Subject: [gecode-users] buggy parallel search

In gecode/search/parallel/engine.hh:

  Engine::busy(void) {
    m_search.acquire();
    assert(n_busy > 0);
    n_busy++;
    m_search.release();
  }

is the assert correct or should it be >= 0?

I am hitting this assertion with gecode compiled with --enable-debug.

Cheers,

--Denys

_______________________________________________
Gecode users mailing list
users at gecode.org
https://www.gecode.org/mailman/listinfo/gecode-users




More information about the users mailing list