> 7) Made various bug fixes and alterations to correct corner cases I found > when implementing this stuff. > > I'm not so sure about 1) and it would be good to get Kevin's input on that. > I hope to extend it to make one generic branch predictor so that the > in-order core can make use of it too, instead of having it's own, duplicated > code. > Merging the inorder branch predictor code would be good.
I recently posted some bug fixes for the branch predictor that should also apply to the branch predictor in O3 as well and at some point they may need to be considered for the general predictor (if they truly apply). Also for the inorder code, right now it's coded as a separate branch predictor "resource" then inside that resource there is some replicated generic branch predictor object code. There is another phase of indirection, because that generic branch predictor code is simply a wrapper for a specific branch predictor imlementation (tournament, two-bit local, etc.) So honestly, I think that that generic branch predictor code we have now, needs to broken up in some way such that the branch predictor is really a "plug-n-play" addition to the CPU models rather than just an object that refer to a specific implementation. However, in the interim, I'd be willing to help out such that the changes you made will apply to inorder as well. -- - Korey
_______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
