Re: [Gems-users] IPC and Number of Instructions

[Muhammad abid Mughal <mabidm_pieas@xxxxxxxxx>]

AOA:-

    when you ask opal to run for 100 million instructions,it means you want opal to run each core for 100 million instructions. But actually if you take a look of system_t::simulate(),  you will come to know that opal makes sure that core0 must run for 100 million instructions, other cores can execute less/more instructions.

In your case , i guess core0 more often fails when it tries to get a lock (ie CASA/CASXA/swap/LDSTUB), causing core0 leaves behind in instruction executions, whereas other cores can go past 100 million inst.

i hope it helps

Cheers,

Muhammad abid

---

From: Dan Gibson <degibson@xxxxxxxx>
To: Gems Users <gems-users@xxxxxxxxxxx>
Sent: Wednesday, December 9, 2009 7:15:00
Subject: Re: [Gems-users] IPC and Number of Instructions

The other processors are probably spinning. Spinning tends to have great IPC.

On Tue, Dec 8, 2009 at 2:53 PM, <ubaid001@xxxxxxx> wrote:

Hi all,

I simulated a four core CMP system with both Opal and Ruby. In the
results.opal file the runtime stats for processor [0],[1],[2],[3], only
processor [0] executed 100 million instructions specified by
"opal0.sim-step 100000000", but processor [1] and [2] have more total
number of instructions than what I had specified.

Even though only one core runs the required number of instructions, how do
the other cores run so many more instructions (almost double the number of
instructions specified). The configuration for all the processors are the
same. This leads to opal reporting a very high IPC for those processors.
e.g : P0 1.6, P1 3.96 , P2 3.97 etc.

This is due to the fact the the number of cycles remain the same for all
processors.

Am wondering if the high value of IPC reported is spurious. Can someone
shed more light on this?

Suhail




_______________________________________________
Gems-users mailing list
Gems-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/gems-users
Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.

--
http://www.cs.wisc.edu/~gibson [esc]:wq!

---

  Get your new Email address!  
Grab the Email name you've always wanted before someone else does!