You could certainly do it per core. MemoryRequest objects have a core ID in
there somewhere -- you could just capture the core ID at the memory
controller boundary after it comes in over the interconnect.

However, per thread is a bit trickier since I believe you'd have to get the
linux kernel involved somehow.

On Thu, Aug 23, 2012 at 11:42 AM, snataraj <[email protected]> wrote:

> Is it possible to study on-chip traffic and memory bandwidth utilization
> on per thread basis of a multithreaded program using MARSS?
>
> If yes, which part of statistics should I look at or which part of the
> simulator should I instrument to get these statistics?
>
> --
> Surya Narayanan,
> E301 - 12,
> Inria Rennes - Bretagne Atlantique Research Centre
> Campus universitaire de Beaulieu
> 35042 Rennes Cedex
>
>
> ______________________________**_________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected].**edu <[email protected]>
> https://www.cs.binghamton.edu/**mailman/listinfo/marss86-devel<https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel>
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to