Re: [Gems-users] different number of cycles, Simics and GEMS

[tian <tian@xxxxxxxxxxxxxxxxxxxxxxx>]

i guess when you use ptime-s in a multi-core system
it will return the value of the core currently running at.

you might try $start=(cpu0.ptime -s) + (cpu1.ptime -s) +...




On 11/03/2010 08:54 PM, Andrew Green wrote:
> Dear all,
> I ran Freqmine (a benchmark from PARSEC) from the beginning of ROI to
> the end of ROI with simics (ipc=1) and got the total cycles equal to
> 49,206,002,380. So about 50b instructions where executed in ROI. I
> calculated that number in this way:
>
> simics> magic-break-enable
> simics> $start = (ptime -s)
> .....
> simics> c
> ...... got magic
> simics> $finish = (ptime -s)
> ....
> simics> $finish - $start
>
> However with GEMS, I get a very different number:
>
> instruction_executed: 227,385,079,675 [ 52866792093 56921608271
> 59666582857 57930096454 ]
> simics_cycles_executed: 298,940,409,581 [ 74735102396 74735102395
> 74735102395 74735102395 ]
> cycles_per_instruction: 0.328672 [ 0.353412 0.328237 0.313136 0.322523 ]
>
> How does GEMS calculate "simics_cycles_executed"?
> _______________________________________________
> 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.
>