I think you should make modifications to the Python script that makes
GEM5 execute, such that every N cycles you dump statistics to a new
file, reset statistics and get back to execution. Probably some hints
can be found in the "configs/example" directory under GEM5, which
contain execution scripts.

On Fri, Feb 17, 2012 at 9:37 AM, Chun Zhang <[email protected]> wrote:
> Hi,
>
> In my project I need to dump out the power profile (with time) for some
> benchmarks. In that case, I think I need to dump out the GEM5 statistics to
> McPAT for power calculation every a few cycles.
>
> As I'm very new to GEM5, could anybody let me know where to start for
> dumping GEM5 statistics for every 100 cycles?
>
> Thanks a lot.
>
>
> On Thu, Feb 16, 2012 at 3:21 AM, Simone Corbetta <[email protected]>
> wrote:
>>
>> As Patrick said, to enable DVFS with GEM5 simulation, you have to plug
>> a power model. McPAT can be used to generate power estimates according
>> to the statistics collected from a cycle-accurate simulation with GEM5
>> (you need also to translate GEM5 output statistics to XML format as
>> required by McPAT). For Dynamic VFS you need to reiterate the loop and
>> simulate changing DVFS after desired number of iterations; GEM5 allows
>> you to stop simulation and then start back again, so you can for sure
>> loop on it.
>>
>> Simone Corbetta
>>
>> On Wed, Feb 15, 2012 at 7:11 PM, Patrick <[email protected]> wrote:
>> > Xiao,
>> >
>> > One of my colleagues at Notre Dame, Sheng Li, was working on a tool
>> > called
>> > McPAT for power analysis. I believe that it has been shown to work with
>> > gem5. This might be useful for you. Last I heard, Li and McPAT had moved
>> > to HP: http://www.hpl.hp.com/research/mcpat/.
>> >
>> > Patrick La Fratta
>> > On Fri, Feb 10, 2012 at 2:42 PM, Xiao Guo <[email protected]> wrote:
>> >>
>> >> Hi,
>> >>
>> >> Is there any DVFS support for gem5? Anybody has done this before?
>> >> Any help would be appreciated!
>> >>
>> >> Thanks!
>> >>
>> >> Xiao
>> >>
>> >>
>> >>
>> >> _______________________________________________
>> >> gem5-users mailing list
>> >> [email protected]
>> >> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>> >
>> >
>> >
>> > _______________________________________________
>> > gem5-users mailing list
>> > [email protected]
>> > http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>>
>>
>> --
>> Simone Corbetta, Ph.D. student
>>
>> Politecnico di Milano
>> Dipartimento di Elettronica e Informazione
>> Viale delle Rimembranze di Lambrate 14
>> 20134 Milano, ITALY
>>
>> Office: +39.02.2399.9623
>> Home: http://home.dei.polimi.it/scorbetta
>> Linkedin: http://www.linkedin.com/in/thelibrarian
>> _______________________________________________
>> gem5-users mailing list
>> [email protected]
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>
>
>
>
> --
> Best Regards,
>
> ======================
> Chun Zhang
> [email protected]
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users



-- 
Simone Corbetta, Ph.D. student

Politecnico di Milano
Dipartimento di Elettronica e Informazione
Viale delle Rimembranze di Lambrate 14
20134 Milano, ITALY

Office: +39.02.2399.9623
Home: http://home.dei.polimi.it/scorbetta
Linkedin: http://www.linkedin.com/in/thelibrarian
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to