On Mon, Nov 26, 2012 at 9:26 AM, Seyed <[email protected]> wrote:

> Dear all,
>
> I am trying to map a benchmark onto different cpu threads in a ooo_2_th
> machine
> configuration; however, I get 1041.28 ipc which is very strange. By setting
> PARSEC_CPU_NUM=1 and PARSEC_CPU_BASE=0 I get the following for
> "blackscholes"
> benchmark:
>
> These numbers are interesting. Can you look into more detail numbers like
each thread's total number of instructions committed and total number of
cycles etc.  There might be a small bug related to IPC or cycle calculation
for each thread in SMT mode.

- Avadh

Stopped after 63105458 cycles, 50000001 instructions and 171 seconds of sim
> time
> (cycle/sec: 369037 Hz, insns/sec: 292397, insns/cyc: 0.7923245086027265)
> Total Tags: ooo_2_th,....cs.....edu,2012-11-26,total
> Stats Summary:
> user.base_machine.ooo_0_0.thread0.commit.ipc = 0.785249
> user.base_machine.ooo_0_0.thread1.commit.ipc = 7.07377e-06
> kernel.base_machine.ooo_0_0.thread0.commit.ipc = 0.430886
> kernel.base_machine.ooo_0_0.thread1.commit.ipc = 0.939382
> total.base_machine.ooo_0_0.thread0.commit.ipc = 0.780967
> total.base_machine.ooo_0_0.thread1.commit.ipc = 0.0113571
> Received simulation kill signal, stopped the simulation and killing the VM
>
> by setting PARSEC_CPU_NUM=1 and PARSEC_CPU_BASE=1 I get this strange
> result:
>
> Stopped after 62507478 cycles, 50000001 instructions and 165 seconds of
> sim time
> (cycle/sec: 378833 Hz, insns/sec: 303030, insns/cyc: 0.7999043090492309)
> Total Tags: ooo_2_th,....cs.....edu,2012-11-25,total
> Stats Summary:
> user.base_machine.ooo_0_0.thread0.commit.ipc = 0.172486
> user.base_machine.ooo_0_0.thread1.commit.ipc = 1041.28
> kernel.base_machine.ooo_0_0.thread0.commit.ipc = 0.0133436
> kernel.base_machine.ooo_0_0.thread1.commit.ipc = 0.000193019
> total.base_machine.ooo_0_0.thread0.commit.ipc = 0.0134638
> total.base_machine.ooo_0_0.thread1.commit.ipc = 0.786441
> Received simulation kill signal, stopped the simulation and killing the VM
>
> Finally, by setting PARSEC_CPU_BASE=0 for "canneal" and PARSEC_CPU_BASE=1
> for
> "blackscholes" I get this:
>
> Stopped after 56453434 cycles, 50000001 instructions and 217 seconds of
> sim time
> (cycle/sec: 260154 Hz, insns/sec: 230414, insns/cyc: 0.8856857317129725)
> Total Tags: ooo_2_th,....cs.....edu,2012-11-26,total
> Stats Summary:
> user.base_machine.ooo_0_0.thread0.commit.ipc = 0.339402
> user.base_machine.ooo_0_0.thread1.commit.ipc = 0.549878
> kernel.base_machine.ooo_0_0.thread0.commit.ipc = 0.422622
> kernel.base_machine.ooo_0_0.thread1.commit.ipc = 0.156399
> total.base_machine.ooo_0_0.thread0.commit.ipc = 0.340366
> total.base_machine.ooo_0_0.thread1.commit.ipc = 0.545319
> Received simulation kill signal, stopped the simulation and killing the VM
>
>
> Any idea about this?
> Thank you in advance for your time.
>
> Best,
> Seyed
>
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to