boehmes [Tue, 03 Nov 2009 17:54:24 +0100] rev 33419
added HOL-Boogie
boehmes [Tue, 03 Nov 2009 14:51:55 +0100] rev 33418
added a specific SMT exception captured by smt_tac (prevents the SMT method from failing with an exception),
replaced unspecific 'error' invocations with raising of specific SMT exceptions,
added annotations to traced SMT problem and solver output
boehmes [Tue, 03 Nov 2009 14:07:38 +0100] rev 33417
ignore parsing errors, return empty assignment instead
wenzelm [Thu, 05 Nov 2009 13:16:22 +0100] rev 33416
scheduler: clarified interrupt attributes and handling;
wenzelm [Thu, 05 Nov 2009 13:01:11 +0100] rev 33415
worker_next: plain signalling via work_available only, not scheduler_event;
scheduler: tuned worker pool adjustments;
wenzelm [Thu, 05 Nov 2009 00:13:00 +0100] rev 33414
revert fulfill_proof_future tuning (actually a bit slower due to granularity issues?);
wenzelm [Wed, 04 Nov 2009 21:22:35 +0100] rev 33413
avoid broadcast work_available, use daisy-chained signal instead;
max_threads: allow as much as 4 * m, after extra delay;
wenzelm [Wed, 04 Nov 2009 21:21:05 +0100] rev 33412
fulfill_proof_future: tuned important special case of singleton promise;
wenzelm [Wed, 04 Nov 2009 20:31:36 +0100] rev 33411
worker_next: treat wait for work_available as Sleeping, not Waiting;
max_threads: simple adaptive scheme between m and 2 * m;
wenzelm [Wed, 04 Nov 2009 11:58:29 +0100] rev 33410
worker activity: distinguish between waiting (formerly active) and sleeping;
tuned;
wenzelm [Wed, 04 Nov 2009 11:37:06 +0100] rev 33409
tuned;
wenzelm [Wed, 04 Nov 2009 11:30:22 +0100] rev 33408
tuned thread data;
wenzelm [Wed, 04 Nov 2009 00:29:58 +0100] rev 33407
worker_next: ensure that work_available is passed on before sleeping (was occasionally lost when worker configuration changed, causing scheduler deadlock);
worker_start: back to non-self version;
scheduler: status output based on ticks;
wenzelm [Tue, 03 Nov 2009 19:52:09 +0100] rev 33406
slightly leaner and more direct control of worker activity etc.;