Review Board 2.0.15


se.py: Modify script to make multiprogramming much easier.

Review Request #598 - Created March 18, 2011 and submitted

Information
Lisa Hsu
gem5
Reviewers
Default
ali, gblack, nate, stever
se.py: Modify script to make multiprogramming much easier.
Now, instead of --bench benchname, you can do --bench bench1-bench2-bench3 and it will
set up a simulation that instantiates those three workloads.  Only caveat is that now,
for sanity checking, your -n X must match the number of benches in the list.

   
Posted (March 18, 2011, 11:30 a.m.)



  
configs/example/se.py (Diff revision 1)
 
 
Splitting on "," would be a little more standard.
  1. True.  I have a moderately good reason for this.  When creating an aggregated checkpoint, the aggregate script makes cpt.b1-b2-b3.XXXXX.  So, if you hand your argument over as --bench b1-b2-b3, no finagling is necessary to be able to find the appropriate checkpoint, it just takes the options.bench string and looks for it, exactly the same as a uniprocessor checkpoint restore situation.  Splitting on commas would mean later overwriting the options.bench param to hand over to Simulation.py to find an aggregated checkpoint or doing some other such thing to make things understood.  This just seemed easier, even if it is a little non-traditional.  I'd prefer to keep it this way.
  2. Eh, this seems like optimizing for a coincidence instead of making things work like people expect (and other parts of M5 already do). It would also be ambiguous if a benchmark had a "-" as part of its name some day.