Re: [Gems-users] SMP vs CMP in GEMS

[Edward Lee <edwl202@xxxxxxxxx>]

Thanks Greg for your prompt and clear response.

Ed

On Wed, May 13, 2009 at 11:10 AM, Greg Byrd <gbyrd@xxxxxxxx> wrote:

Yes, that's correct -- it's all smoke and mirrors!

The CMP protocols tend to have shared L2 caches.  If you want private L2
only, you can use the SMP protocols and adjust the parameters to match
your expected CMP configuration.

GEMS has the notion of Chips and a certain number of processors per
chip.  This is orthogonal to what Simics (and the OS) believe is the
configuration.

...Greg

Edward Lee wrote:
> Hi all,
>
> I have a very general question to clarify some of my thoughts.
>
> Since GEMS is mainly using serengeti targets which are based on single
> UltraSparc-III Cu processors. I assume multi-core/CMP concept is more
> of an emulation of such a chip utilizing these single core processors
> and changing system parameters accordingly. Is this true?
>
> Accordingly, if we look at the protocols in GEMS there are SMP based
> and CMP based protocols. So, can we think of SMPs also as CMPs with
> private caches in GEMS maybe after modifying the latencies to be smaller?
>
> Thanks in advance for any input on this.
>
> Regards,
>
> Ed

> ------------------------------------------------------------------------
>
> _______________________________________________
> Gems-users mailing list
> Gems-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.
>
>

_______________________________________________
Gems-users mailing list
Gems-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/gems-users
Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.