> Nate, so what mechanism are you envisioning to change the eventq pointer? It > seems like we need a solution that modifies the Ruby objects' eventq pointers > to point to a supplemental eventq during cache warmup and the main eventq > during the main simulation loop.
There is no mechanism for changing the eventq pointer in place, but what you're talking about is what I was thinking. That said, I'm not sure that there's any clear advantage to this mechanism over Steve's, so whichever is easier. Nate _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
