Right, that part makes sense. The thing I'm asking about is the line you 
added here


   CPUClass.clock = '2GHz'
  +CPUClass.numThreads = len(workloads)

where it looks to me like you're scaling up the number of threads so the 
workloads will fit. What I'm saying is that if somebody has too many 
workloads, that's probably wrong and they should hit the fatal you've 
added rather than accommodating it. I could be misinterpreting the code 
since I'm not seeing it in context.

Gabe

Korey Sewell wrote:
> Basically, if you have a 2-thread SMT CPU then you cant give a >2 
> workload in SE mode especially since there is no concept of thread 
> management in SE mode as well.
>
> Otherwise, the parameter "numThreads" in the CPU is backwards because 
> you can pre-set it in the configuration but it just gets overwritten 
> every time you change the workload.
>
> With this change, I'm saying the "numThreads" parameter in the python 
> config needs to match the "actual_num_threads"  in the O3CPU.
>
> If you want a 4-thread CPU, then change the parameter to 
> "numThreads=4" I would think.
>
> On Sat, Jul 25, 2009 at 12:51 AM, Gabe Black <gbl...@eecs.umich.edu 
> <mailto:gbl...@eecs.umich.edu>> wrote:
>
>     Do we actually want to scale the number of threads on a CPU to
>     match the
>     workload assigned to it? That seems like something that should be
>     fixed
>     with the number of workloads conforming to it, not the other way
>     around.
>     Am I misunderstanding how that's supposed to work?
>
>     Gabe
>
>     Korey Sewell wrote:
>     > changeset 70172be3f986 in /z/repo/m5
>     > details: http://repo.m5sim.org/m5?cmd=changeset;node=70172be3f986
>     > description:
>     >       o3-smt: enforce numThreads parameter for SMT SE mode
>     >
>     > diffstat:
>     >
>     > 2 files changed, 8 insertions(+), 4 deletions(-)
>     > configs/example/se.py     |    1 +
>     > src/cpu/o3/cpu_builder.cc |   11 +++++++----
>     >
>     > diffs (36 lines):
>     >
>     > diff -r 82ee4a597908 -r 70172be3f986 configs/example/se.py
>     > --- a/configs/example/se.py   Wed Jul 22 20:28:32 2009 -0500
>     > +++ b/configs/example/se.py   Sat Jul 25 00:50:27 2009 -0400
>     > @@ -124,6 +124,7 @@
>     >  (CPUClass, test_mem_mode, FutureClass) =
>     Simulation.setCPUClass(options)
>     >
>     >  CPUClass.clock = '2GHz'
>     > +CPUClass.numThreads = len(workloads)
>     >
>     >  np = options.num_cpus
>     >
>     > diff -r 82ee4a597908 -r 70172be3f986 src/cpu/o3/cpu_builder.cc
>     > --- a/src/cpu/o3/cpu_builder.cc       Wed Jul 22 20:28:32 2009 -0500
>     > +++ b/src/cpu/o3/cpu_builder.cc       Sat Jul 25 00:50:27 2009 -0400
>     > @@ -51,14 +51,17 @@
>     >      // Full-system only supports a single thread for the moment.
>     >      ThreadID actual_num_threads = 1;
>     >  #else
>     > +    if (workload.size() > numThreads) {
>     > +        fatal("Workload Size (%i) > Max Supported Threads (%i)
>     on This CPU",
>     > +              workload.size(), numThreads);
>     > +    } else if (workload.size() == 0) {
>     > +        fatal("Must specify at least one workload!");
>     > +    }
>     > +
>     >      // In non-full-system mode, we infer the number of threads from
>     >      // the workload if it's not explicitly specified.
>     >      ThreadID actual_num_threads =
>     >          (numThreads >= workload.size()) ? numThreads :
>     workload.size();
>     > -
>     > -    if (workload.size() == 0) {
>     > -        fatal("Must specify at least one workload!");
>     > -    }
>     >  #endif
>     >
>     >      numThreads = actual_num_threads;
>     > _______________________________________________
>     > m5-dev mailing list
>     > m5-dev@m5sim.org <mailto:m5-dev@m5sim.org>
>     > http://m5sim.org/mailman/listinfo/m5-dev
>     >
>
>     _______________________________________________
>     m5-dev mailing list
>     m5-dev@m5sim.org <mailto:m5-dev@m5sim.org>
>     http://m5sim.org/mailman/listinfo/m5-dev
>
>
>
>
> -- 
> - Korey
> ------------------------------------------------------------------------
>
> _______________________________________________
> m5-dev mailing list
> m5-dev@m5sim.org
> http://m5sim.org/mailman/listinfo/m5-dev
>   

_______________________________________________
m5-dev mailing list
m5-dev@m5sim.org
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to