[Gems-users] how to run different benchmarks on different cores in a CMP system

[shanshuchang@xxxxxxxxx]

Hi all,

    I have boot 16-core CMP system, also I have install simics3.0.31 +
GEMS2.1 as the simulation platform.
   I want to use different cores to run different benchmarks.For
example,assuming two multicore benchmarks called P1 and P2,and now I
want to use the first 8 cores(cpu0-cpu7) to run P1 and the remaining
cores run P2.
   Here is what I supposed: I need to write a new file which contains the
two benchmarks, do the specified work about the processor set and
processor binding in the file,then I compile the work,copy the binaries
to the simulated 16-core machine,load modules and run the simulation.
Is it right?
   I have found from the previous questions in this forum that benchmark
binding work has been handled by guys here.I wander if anyone can pass
me some instances about how to finish the benchmark-binding work...

any help will be appreciated!