> On Nov. 14, 2012, 2:23 p.m., Nathan Binkert wrote: > > src/sim/core.hh, line 44 > > <http://reviews.gem5.org/r/1537/diff/1/?file=31015#file31015line44> > > > > This should go. > > Nilay Vaish wrote: > I think we need a global notion of time, which is separate from > the time maintained by an particular event queue.
I wasn't objecting to that. I was objecting to the fact that this is a reference to an internal value of another datastructore. > On Nov. 14, 2012, 2:23 p.m., Nathan Binkert wrote: > > src/sim/core.hh, line 47 > > <http://reviews.gem5.org/r/1537/diff/1/?file=31015#file31015line47> > > > > Is this still needed? > > Nilay Vaish wrote: > Ruby resets that current tick while doing warming up the caches. Ok. Then I think it should be call a setter function on the main event queue. > On Nov. 14, 2012, 2:23 p.m., Nathan Binkert wrote: > > src/sim/core.hh, line 46 > > <http://reviews.gem5.org/r/1537/diff/1/?file=31015#file31015line46> > > > > This should become: > > inline Tick curTick { return mainEventQueue.curTick(); } > > > > Nilay Vaish wrote: > This would require making mainEventQueue visible in sim/core.hh. This is > not possible > as of now given the dependency structure of the header files. Yuck. :/ Can you at least name the variable "mainEventQueueCurTick" and define it on the line below where mainEventQueue is defined in eventq.cc? I'd at least like it to be very clear what is going on. I'm hoping that some day it will go away, and if the two are right next to each other, that will be much easier to accomplish. (The location of the declaration in core.hh is fine). An alternative is to remove the curTick function from core.hh and move it into eventq.hh. This might require adding a some #includes to files, but it would be much better from an encapsulation standpoint. - Nathan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/1537/#review3697 ----------------------------------------------------------- On Nov. 14, 2012, 2:18 p.m., Nilay Vaish wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/1537/ > ----------------------------------------------------------- > > (Updated Nov. 14, 2012, 2:18 p.m.) > > > Review request for Default. > > > Description > ------- > > Changeset 9356:84a9103bd1b7 > --------------------------- > sim: have a curTick per eventq > This patch adds a _curTick variable to an eventq. This variable is updated > whenever an event is serviced in function serviceOne(), or all events upto > a particular time are processed in function serviceEvents(). This change > helps when there are eventqs that do not make use of curTick for scheduling > events. > > > Diffs > ----- > > src/sim/core.hh 0fea324c832c > src/sim/core.cc 0fea324c832c > src/sim/eventq.hh 0fea324c832c > src/sim/eventq.cc 0fea324c832c > src/sim/simulate.cc 0fea324c832c > > Diff: http://reviews.gem5.org/r/1537/diff/ > > > Testing > ------- > > > Thanks, > > Nilay Vaish > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
