simulation@uflorida.cis.ufl.edu (Moderator: Paul Fishwick) (05/27/88)
Volume: 2, Issue: 9, Fri May 27 10:12:49 EDT 1988 +----------------+ | TODAY'S TOPICS | +----------------+ (1) Debugging Distributed Simulations Moderator: Paul Fishwick, Univ. of Florida Send topical mail to: simulation@uflorida.cis.ufl.edu ----------------------------------------------------------------------------- Date: Wed, 25 May 88 15:21:23 PDT From: unni@sm.unisys.com (Unni Warrier) To: lomow@cpsc.calgary.cdn, simulation@ufl.edu Subject: Your comments in comp.simulation >> But my heart goes out to someone trying to follow 10 threads of >> logic WITH the possibility of rollback! > > From lomow@cpsc.calgary.cdn... > >This is a non-issue. Someone using TW does not see processes >being rolled back and does not have to take rollback into >consideration when writing the simulation program. I was talking about debugging a dist. program, not writing one here. We all agree that writing a DP is easier than debugging one...... > >Going back to my first point, if we can provide programming >abstractions that hide the underlying synchronisation mechanism, >we should be able to provide debugging abstractions that hide >the underlying synchronisation mechanism. But do we know of any work being done in this area? >One of the things necessary for debugging distributed >simulations is "determinism". >If you don't have determinism it is very difficult (almost >impossible) to re-create some bugs and this can be very >frustrating. > This is precisely the point that was being made by Jerry. And my concern is that the TW mechanism is less "deterministic" than BCM, because of the possibility of rollback. unni@cs.ucla.edu +--------------------------+ | END OF SIMULATION DIGEST | +--------------------------+