Are you running a multithreaded benchmark?

On Tue, May 28, 2013 at 5:23 PM, Qiuling Zhu <[email protected]>wrote:

> Hi,
>
> I have the problem when I tried to simulate more than one core with
> "scons -Q c=4 config=./config/default.conf", and  I was expecting to see
> simulation results for each of the four cores. However, the reported stats
> sim result (including the area/power results from McPAT) is only for one
> core.  Really appreciate that if someone has any suggestions to solve the
> problem.
>
> Many thanks!
> Jolin
>
>
>
> ---------- Forwarded message ----------
> From: Qiuling Zhu <[email protected]>
> Date: Fri, May 24, 2013 at 4:35 PM
> Subject: Multi-core simulation problems
> To: [email protected]
>
>
>  Hi,
>
> I tried to simulate more than one core (e.g., 4 cores) by setting c=4 when
> compiling scons as below :
>
> scons -Q c=4 config=./config/default.conf
>
> ( I used the default.conf file without any changes. )
>
> (1) Then the first problem I have is that I can only set "-machine
> single_core" when running Marss in (qeum) simconfig. If I use other
> machines like  "-machine ooo_2_th", I will get the error as below:
> ---------
> server running on `127.0.0.1:5900'
> ::ERROR::Can't find 'ooo_2_th' machine generator.
> ---------
>
> Here is the whole command I use for running marss:
>
> simconfig -run -kill-after-run -stopinsns 100m -stats marss_result.txt
> -logfile marss_logfile.txt -machine single_core (or ooo_2_th)
>
>
> (2) For my second problem, since it works well if I set "-machine
> single_core", however, the resulting stats results are only for one core.
> I also tried to generate the power number with McPAT and it also only
> generates results for one core instead of for four cores.  Is it because
> the machine single_core can only be used for a single core simulation
> (scons -Q c=1) ?
>
> I really appreciate if anyone can help.  By the way, I use the
> ubuntu-natty.qcow2 disk image. Many thanks!!!
>
> Jolin
>
>
>
> --
> There will be miracles when you believe!
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> 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