Thanks for the quick reply!

Based on the last mail, the stop command received at the start of ROI
is used for creating the checkpoint.
After this, we can restart the simulation with the newly created
checkpoint and continue the simulation.

Am I right?

PS. I have noticed that only some of the PARSEC benchmarks will stop
right after entering ROI, is this normal?

Thanks again

Dennis



2011/12/17 avadh patel <[email protected]>:
>
> On Fri, Dec 16, 2011 at 2:35 AM, Teng-Feng Yang <[email protected]>
> wrote:
>>
>> Hi,
>>
>> I am trying to run the parsec benchmark in the parsec-ROI  image on marss
>> 0.2.1.
>> I have tried several benchmarks, but it seems that the simulation will
>> receive a stop and end quickly in nearly all benchmarks (approx 10
>> sec).
>> However, the console of marss shows that the benchmark has not yet
>> finished, is this normal?
>>
>> FYI
>> The command I use to start marss is
>> qemu/qemu-system-x86_64 -vnc :0 -hda parsecROI.img
>>
>> The command to start benchmark is
>> parsecmgmt -a run -x roi -c gcc-hooks -n 8 -p (BENCH)
>>
> Normally we create a checkpoint at start of ROI by setting 'CHECKPOINT_NAME'
> environment variable.  Once the checkpoints are created then we use those
> checkpoints to run simulations.  Take a look at 'create_checkpoints.py'
> script that is useful in creating checkpoints.  Also take a look at
> 'run_bench.py' script for running benchmarks from checkpoints.  Both scripts
> are in 'util' folder.
>
> - Avadh
>>
>> Thanks
>>
>> Dennis
>>
>> _______________________________________________
>> http://www.marss86.org
>> Marss86-Devel mailing list
>> [email protected]
>> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
>



-- 
Teng Feng Yang
Research Assistant of Director. P.C. Yew
Parallel Processing Laboratory
Institute of Information Science
Academia Sinica, Taiwan
Tel: 886-2-27883799#1676
E-mail:[email protected]

_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to