FYI:
the patch so syscall emulation threads dont stop after one thread finishes
should come soon. I have it in my tree now, but I just have to find the time
to generate the appropriate mercurial patch queue in a clean tree then email
the patch out. Since everything else in my tree isnt "patchqueued" there's a
little reverse engineering to do...
I'll try to post it in a day or two though...
On Tue, Apr 15, 2008 at 12:20 AM, nathan binkert <[EMAIL PROTECTED]> wrote:
> Yes, this should work. My guess is that in the 2GHz/1GHz case, the
> second CPU isn't getting far enough to print the value before the
> first CPU causes the simulation to exit. If you used more complicated
> programs, or hacked things so that the system didn't terminate until
> all threads were done, I'm pretty sure it will work. I run multiple
> frequencies all the time in full system mode.
>
> Nate
>
> > I am trying to run two simple workloads "Hello World" in two independent
> > CPUs which have different frequencies.
> > At first when I set CPU0'clock to "2GHz" and CPU1's clock to "3GHz",
> > everything seems to work fine.
> > But when I changed CPU1's clock from "3GHz" to "1GHz", there is only
> one
> > program has been executed.
> > If I changed CPU1's clock to "4GHz", it also only one program executed.
> >
> > Can different CPUs run at any different frequencies in m5? Any
> suggestion
> > would be highly appreciated.
> >
> > Regards,
> > Jerry
> >
> > ===================================================
> > More details about the config file and results:
> >
> > The changes from the original example file se.py:
> > --- se.py 2008-04-15 10:06:02.000000000 +0800
> > +++ se0.py 2008-03-01 01:28:51.000000000 +0800
> > @@ -86,8 +86,8 @@
> >
> > np = options.num_cpus
> >
> > -system = System(cpu = [CPUClass(cpu_id=i) for i in xrange(np)],
> > - physmem = PhysicalMemory(range=AddrRange("512MB")),
> > membus = Bus(), mem_mode = test_mem_mode)
> > +system = System(cpu = [CPUClass(cpu_id=0,clock="2GHz"),
> > CPUClass(cpu_id=1,clock="3GHz"))],
> > + physmem = PhysicalMemory(range=AddrRange("512MB")),
> >
> >
> > system.physmem.port = system.membus.port
> > @@ -106,7 +106,7 @@
> > system.cpu[i].connectMemPorts(system.tol2bus)
> > else:
> > system.cpu[i].connectMemPorts(system.membus)
> > - system.cpu[i].workload = process
> > + system.cpu[i].workload = process[i]
> >
> > if options.fastmem:
> > system.cpu[0].physmem_port = system.physmem.port
> >
> > Command line running m5 :
> > $ build/ALPHA_SE/m5.debug configs/example/se.py
> >
> >
> --cmd="/home/jerry/m5-2.0b5/tests/test-progs/hello/bin/alpha/linux/hello0;/home/jerry/m5-2.0b5/tests/test-progs/hello/bin/alpha/linux/hello1"
> > -d --cache -n2
> >
> > The first running result:
> > ...................
> > Global frequency set at 1000000000000 ticks per second
> > 0: system.remote_gdb.listener: listening for remote gdb #0 on port 7000
> > 0: system.remote_gdb.listener: listening for remote gdb #1 on port 7001
> > **** REAL SIMULATION ****
> > warn: Entering event queue @ 0. Starting simulation...
> > warn: Increasing stack size by one page.
> > warn: Increasing stack size by one page.
> > Hello World0!
> > Hello World1!
> > Exiting @ cycle 5536000 because target called exit()
> >
> > Change CPU1's clock from "3GHz" to "3GHz":
> > --- se0.py 2008-04-15 11:09:02.000000000 +0800
> > +++ se1.py 2008-04-15 11:07:06.000000000 +0800
> > @@ -86,7 +86,7 @@
> >
> > np = options.num_cpus
> >
> > -system = System(cpu = [CPUClass(cpu_id=0,clock="2GHz"),
> > CPUClass(cpu_id=1,clock="3GHz")],
> > +system = System(cpu = [CPUClass(cpu_id=0,clock="2GHz"),
> > CPUClass(cpu_id=1,clock="1GHz")],
> > physmem = PhysicalMemory(range=AddrRange("512MB")),
> > membus = Bus(), mem_mode = test_mem_mode)
> > The second running result:
> > .........................
> > Global frequency set at 1000000000000 ticks per second
> > 0: system.remote_gdb.listener: listening for remote gdb #0 on port 7000
> > 0: system.remote_gdb.listener: listening for remote gdb #1 on port 7001
> > **** REAL SIMULATION ****
> > warn: Entering event queue @ 0. Starting simulation...
> > warn: Increasing stack size by one page.
> > warn: Increasing stack size by one page.
> > Hello World1!
> > Exiting @ cycle 5502000 because target called exit()
> >
> > _______________________________________________
> > m5-users mailing list
> > [email protected]
> > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
> >
> _______________________________________________
> m5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>
--
----------
Korey L Sewell
Graduate Student - PhD Candidate
Computer Science & Engineering
University of Michigan
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
