For multi-programmed workloads, I have a clean solution now to the problem
referred to in the thread:
http://www.mail-archive.com/[email protected]/msg02832.html
We need to ensure each core completes atleast max_insts_all_thread
instructions. The default implementation exits as soon as any 1 core reaches
that count. I created a new exit event called which counts the total system
events and puts an exit event on the main Q only when all cores have completed
the required instructions. Further, it can dump stats whenever a particular
core has reached that count, giving the effect of "freezing" stats for that
core for max_insts instructions.
I can send a patch for this if interested. Has anyone else implemented this?
Sujay
From: Lisa Hsu
Sent: Tuesday, June 29, 2010 12:36 PM
To: M5 users mailing list
Subject: Re: [m5-users] running multi-program workloads with
individualcheckpoints for each program
Hi Susie,
The code for that has been checked into the development repository. It's in
the util/ directory and is called checkpoint-aggregator.py.
Lisa
On Sun, Jun 20, 2010 at 3:27 PM, Susie Sally <[email protected]> wrote:
Hi,
I would like to run simulations with multiple workloads with restoring an
individual checkpoint for each program.
I read this thread
http://www.mail-archive.com/[email protected]/msg02870.html from the archive.
I will very much appreciate if you can point me to the mentioned patch or
give me pointers for the required modifications to do this.
thank you very much.
regards,
Susie
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
--------------------------------------------------------------------------------
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
