Re: [PATCH 0/6] sched: Expose writable interface for preferred CPUs mask

[Mete Durlu <meted@linux.ibm.com>]

Hi Shrikanth,

>>> I've been trying to do this with crutches in user-space for a long 
>>> time.  Thanks for implementing this in the kernel, and with a 
>>> convenient user interface.
>>>

>>>> A user can mark a CPU non-preferred and pin a workload to it. Fair 
>>>> tasks
>>>> whose affinity includes preferred CPUs are steered away, while tasks
>>>> whose allowed CPUs contain no preferred CPU can still run there. The 
>>>> CPU
>>>> stays online, task affinity is preserved, and the selection can be
>>>> reversed without rebooting:
>>>>
>>>>     # rmmod steal_governor
>>>>     # echo 0 > /sys/devices/system/cpu/cpu3/preferred
>>>>     # taskset -c 3 ./my_workload
>>>>     # echo 1 > /sys/devices/system/cpu/cpu3/preferred
>>>
>>> My use-case (being a compiler optimization engineer) is to pin 
>>> benchmarks to specific cores, and stop the system from using those 
>>> cores even when the cores are idle.  This prevents L1/L2 cache 
>>> pollution and makes performance numbers more stable.
>>
>> Hi Maxim,
>>
>> Thank you for the feedback.
>>
>> I agree, this trick with pinning a workload to an non-preferred CPU
>> looks promising for benchmarking purposes, and probably for more.
>>
> 
> Sorry, i don't follow.
> 
> Did you mean pin to preferred CPUs? Why would anyone want to run their
> workload only on non-preferred CPUs? By definition, they are not meant to
> be used as much as possible. and It is benchmarking.

I think they want to prevent anything else *besides* their workload
from running on these CPUs by marking them non-preferred and then
pinning their workload there.
Sort of like like dynamically configurable isolated CPUs.

IMO this fits 1-1 with what Yury is describing with this patchset.

Thanks,
-Mete