Re: [Gems-users] DNUCA cache

[Bradford Beckmann <beckmann@xxxxxxxxxxx>]

Daniele,

You are correct.  The MSI_MOSI_CMP_directory protocol does not implement
DNUCA memory.  The actual protocol that implements DNUCA was not included
with the GEMS release.  The reason being is the CMP-DNUCA protocol is
extremely complicated and relies on some undocumented SLICC features.

I can send you the CMP-DNUCA protocol files, but they protocol won't work
directly in the released version of GEMS.  However, they will provide you
a good understanding of how GEMS implemented CMP-DNUCA.

I hope that helps,

Brad


On Tue, 4 Oct 2005, Daniele Mangano wrote:

> Hello,
>
>
>
> We are evaluating the possibility to use your Ruby memory system in order to
> explore CMPs with a shared L2 cache memory. In the paper "Multifacet's
> General Execution-driven Multiprocessors Simulator (GEMS) Toolset" you wrote
> that networks such as CMP-DNUCA can be create by using the file-specified
> mode. Although we are able to create CMPs with shared L2 NUCA memories, we
> are not sure that your MSI_MOSI_CMP_directory SLICC model allows to use also
> a DynamicNUCA (DNUCA).
>
> How can we create a shared L2 DynamicNUCA memory?
>
> Thanks,
>
> ________________________________
> Daniele Mangano, Ph.D Student
> Dept. of Information Engineering
>
> University of Pisa
>
> Via Diotisalvi, 2 - 56122 Pisa, Italy
>
> Tel  +39 050 2217 458
>
> Fax +39 050 2217 600
>
> E-mail: daniele.mangano@xxxxxxxxxxxx
>
>
>
>

-----------------------------------------------------------------
 Department of Computer Science         Residence
 University of Wisconsin
 1210 W. Dayton St. #6366               608 Eagle Heights Apt. L
 Madison, WI 53706                      Madison, WI 53705
 (608)265-2702				(608)852-6133
-----------------------------------------------------------------