Re: [Gems-users] how to construct a CMP system with 32 or 64 cores


Date: Fri, 23 May 2008 15:26:53 -0700
From: "Derek Hower" <drh5@xxxxxxxxxxx>
Subject: Re: [Gems-users] how to construct a CMP system with 32 or 64 cores
Simics will access its own memory hierarchy; it has to to do a
functionally correct simulation.  However, that does not affect the
results Ruby or Opal present because they maintian their own image of
the memory hierarchy independent of whatever Simics thinks it looks
like.

-Derek

On Fri, May 23, 2008 at 1:53 PM, Guoqiang Yang <nkyangguoqiang@xxxxxxxxx> wrote:
> Hi Greg,
>
> Do you mean that simics won't access its memory hierarchy with ruby and
> opal both turned on?
>
> Guoqiang Yang
>
> Since memory timing is all done by Ruby, it doesn't matter where Simics
> thinks the memory is located.
>
>
> Guoqiang Yang wrote:
>
>    Hi Bruno,
>
>
>
>    I also have another question about the memory configuration. To my
>    understanding, in the clustered system, only the first board is
>    configured with memory and the other boards don't have any memory
>    since the megs_per_cpu is set to zero. Is this a reasonable system?
>    How much effect does this configuration have on the target system?
>
>    Many thanks,
>
>    Guoqiang Yang
>    _______________________________________________
>    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/"; <https://lists.cs.wisc.edu/archive/gems-users/%22>; 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.
>
>
[← Prev in Thread] Current Thread [Next in Thread→]