On 08/14/11 21:56, Korey Sewell wrote: > On Mon, Aug 15, 2011 at 12:48 AM, Gabe Black <[email protected]> wrote: > >> The skid buffer doesn't really have anything to do with it. Instructions >> are on the wire from fetch when the squash comes. Because they aren't >> actually in one stage or another they keep coming. By the time they >> arrive, decode isn't processing any more instructions since it's waiting >> for the squash to finish. > Which is why you should always sort the incoming instructions before you > actually process them, no? That sort, will either block them (putting them > in the skidBuffer) or let them be, right? >
I don't know. As I said, I didn't have time to dig around in decode to figure it out. If that's what we should do then go ahead. Gabe _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
