[lkml]   [2019]   [Sep]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [RFC PATCH 1/9] sched,cgroup: Add interface for latency-nice

On Thu, Sep 05, 2019 at 12:30:02 +0100, Peter Zijlstra wrote...

> On Thu, Sep 05, 2019 at 12:13:47PM +0100, Qais Yousef wrote:
>> On 09/05/19 12:46, Peter Zijlstra wrote:
>> > This is important because we want to be able to bias towards less
>> > importance to (tail) latency as well as more importantance to (tail)
>> > latency.
>> >
>> > Specifically, Oracle wants to sacrifice (some) latency for throughput.
>> > Facebook OTOH seems to want to sacrifice (some) throughput for latency.
>> Another use case I'm considering is using latency-nice to prefer an idle CPU if
>> latency-nice is set otherwise go for the most energy efficient CPU.
>> Ie: sacrifice (some) energy for latency.
>> The way I see interpreting latency-nice here as a binary switch. But
>> maybe we can use the range to select what (some) energy to sacrifice
>> mean here. Hmmm.
> It cannot be binary, per definition is must be ternary, that is, <0, ==0
> and >0 (or middle value if you're of that persuasion).
> In your case, I'm thinking you mean >0, we want to lower the latency.
> Anyway; there were a number of things mentioned at OSPM that we could
> tie into this thing and finding sensible mappings is going to be a bit
> of trial and error I suppose.
> But as patrick said; we're very much exporting a BIAS knob, not a set of
> behaviours.

Right, although I think behaviours could still be exported but via a
different and configurable interface, using thresholds.

Either at compile time or via procfs maybe we can expose and properly
document what happen in the scheduler if/when a task has a "latency
niceness" crossing a given threshold.

For example, by setting something like:

/proc/sys/kernel/sched_cfs_latency_idle = 1000

we state that the task is going to be scheduled according to the
SCHED_IDLE policy.

( ( (tomatoes target here) ) )

Not sure also if we wanna commit to user-space APIs how we internally
map/translate a "latency niceness" value into a scheduler behaviour
bias. Maybe better not at least at the very beginning.


#include <best/regards.h>

Patrick Bellasi

 \ /
  Last update: 2019-09-05 13:41    [W:0.103 / U:0.776 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site