Hi Tim,

When I originally wrote the branch predictor code, I had hoped to make  
it a flexible object that could be used across multiple CPUs. At the  
same time I had the exact same conflict you're mentioning -- it seemed  
to be overkill for a small amount of code. But I think if your goal is  
to have the branch predictor be warmed up in SMARTS-style sampling, at  
that point it might be worthwhile to try to make it a simulation object.

I know it's not a definite yes or no, but hopefully it'll help sway  
you one way or another.

Kevin

Quoting Timothy M Jones <[email protected]>:

> Hi all,
>
> As I've previously mentioned, I'd like to run simulations using SMARTS
> methodology.  To do this, I will switch between O3CPU and AtomicSimpleCPU
> repeatedly.  I need to be able to keep the branch predictor warm when
> using Atomic, but at the moment this can't be done.
>
> Does anyone have any ideas on the best way to do this?  I was wondering if
> I should make the branch predictor a new simulation object, like the
> caches, that can be attached to any type of CPU.  Then I could group all
> the code together in the cpu/pred directory instead of having duplicated
> code in cpu/o3/bpred_unit.cc and cpu/inorder/resources/bpred_unit.cc.
> Would that make sense, or am I thinking of using a sledge hammer to crack
> a nut?
>
> Cheers
> Tim
>
>
> --
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
> _______________________________________________
> m5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/m5-dev
>
>
>


_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to