* niced tasks on SMT system
@ 2017-12-28 12:10 Pavel Machek
2017-12-28 13:23 ` Mike Galbraith
0 siblings, 1 reply; 6+ messages in thread
From: Pavel Machek @ 2017-12-28 12:10 UTC (permalink / raw)
To: Thomas Gleixner, kernel list; +Cc: mingo, hpa, x86, peterz
[-- Attachment #1: Type: text/plain, Size: 734 bytes --]
Hi!
Ok, so I'm compiling, and I'd like to run a flight simulator.
Flightgear normally does 20fps on my system... kinda low but playable.
I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
one of CPUs to Flightgear (good), but the smt sibling is used by the
compilation, and I'm down to 9 fps. Not good.
Even with single-threaded make, I have 10-13fps.
Is there way to learn which CPUs are SMT siblings?
Is there way to disable SMT during runtime?
Can I do something to improve Flightgear performance and still do
compilation?
Thanks,
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: niced tasks on SMT system
2017-12-28 12:10 niced tasks on SMT system Pavel Machek
@ 2017-12-28 13:23 ` Mike Galbraith
2017-12-28 18:53 ` Pavel Machek
0 siblings, 1 reply; 6+ messages in thread
From: Mike Galbraith @ 2017-12-28 13:23 UTC (permalink / raw)
To: Pavel Machek, Thomas Gleixner, kernel list; +Cc: mingo, hpa, x86, peterz
On Thu, 2017-12-28 at 13:10 +0100, Pavel Machek wrote:
> Hi!
>
> Ok, so I'm compiling, and I'd like to run a flight simulator.
>
> Flightgear normally does 20fps on my system... kinda low but playable.
>
> I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
> one of CPUs to Flightgear (good), but the smt sibling is used by the
> compilation, and I'm down to 9 fps. Not good.
>
> Even with single-threaded make, I have 10-13fps.
>
> Is there way to learn which CPUs are SMT siblings?
cat /sys/devices/system/cpu/cpu{N}/topology/thread_siblings
> Is there way to disable SMT during runtime?
You could offline them, but wouldn't it be better to tell each which
CPUs they can use, or perhaps partition your box with cpusets?
> Can I do something to improve Flightgear performance and still do
> compilation?
Sure, run everything associated with your game as RT, and everything
not game gets the leftover cycles. If there are none, box will
throttle RT to save itself from it's psycho :) driver and you'll know
that you need a bigger box. (it's likely your phone)
Oh yeah, echo NO_RT_RUNTIME_SHARE > /sys/kernel/debug/sched_features
before you try that, otherwise the throttle won't help.
-Mike
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: niced tasks on SMT system
2017-12-28 13:23 ` Mike Galbraith
@ 2017-12-28 18:53 ` Pavel Machek
2017-12-28 18:58 ` Ozgur
0 siblings, 1 reply; 6+ messages in thread
From: Pavel Machek @ 2017-12-28 18:53 UTC (permalink / raw)
To: Mike Galbraith; +Cc: Thomas Gleixner, kernel list, mingo, hpa, x86, peterz
[-- Attachment #1: Type: text/plain, Size: 2523 bytes --]
Hi!
> > Ok, so I'm compiling, and I'd like to run a flight simulator.
> >
> > Flightgear normally does 20fps on my system... kinda low but playable.
> >
> > I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
> > one of CPUs to Flightgear (good), but the smt sibling is used by the
> > compilation, and I'm down to 9 fps. Not good.
> >
> > Even with single-threaded make, I have 10-13fps.
> >
> > Is there way to learn which CPUs are SMT siblings?
>
> cat /sys/devices/system/cpu/cpu{N}/topology/thread_siblings
Thanks for a hint.
Well, something is definitely wrong there:
pavel@duo:/data/l/linux-n900$ cat
/sys/devices/system/cpu/cpu0/topology/thread_siblings
03
I believe that means that cpu0 & cpu1 are sharing physical cpu. Yet,
when I run flightgear and "nice while1". flightgear goes to CPU#0 and
while1 to CPU#1. Would not it be nice if while1 went preferably to
CPUs #2 and #3? Ok. after a while while1 moved to cpu#2, good.
> > Is there way to disable SMT during runtime?
>
> You could offline them, but wouldn't it be better to tell each which
> CPUs they can use, or perhaps partition your box with cpusets?
Let me try offlining first... and yes, SMT seems to be problem here.
| CPU0 | CPU1 | CPU2 | CPU3 | |
| flightgear | off | off | off | 21 fps |
| flightgear | while1 | off | off | 17 fps |
| flightgear+while1 | off | off | off | 10 fps |
| flightgear | off | while1 | off | 21 fps |
> > Can I do something to improve Flightgear performance and still do
> > compilation?
>
> Sure, run everything associated with your game as RT, and everything
> not game gets the leftover cycles. If there are none, box will
> throttle RT to save itself from it's psycho :) driver and you'll know
> that you need a bigger box. (it's likely your phone)
> Oh yeah, echo NO_RT_RUNTIME_SHARE > /sys/kernel/debug/sched_features
> before you try that, otherwise the throttle won't help.
Sure, but that is not a problem I have.
Scheduler already (correctly) decides that the game is important, and
gives game all the time it wants (100% of one CPU). That is not a problem.
Problem seems to be that it schedules other tasks to SMT sibling, and
that slows down the game quite significantly.
Best regards,
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: niced tasks on SMT system
2017-12-28 18:53 ` Pavel Machek
@ 2017-12-28 18:58 ` Ozgur
2017-12-28 20:59 ` Pavel Machek
0 siblings, 1 reply; 6+ messages in thread
From: Ozgur @ 2017-12-28 18:58 UTC (permalink / raw)
To: Pavel Machek, Mike Galbraith
Cc: Thomas Gleixner, kernel list, mingo@redhat.com, hpa@zytor.com,
x86@kernel.org, peterz@infradead.org
28.12.2017, 21:54, "Pavel Machek" <pavel@ucw.cz>:
> Hi!
>
>> > Ok, so I'm compiling, and I'd like to run a flight simulator.
>> >
>> > Flightgear normally does 20fps on my system... kinda low but playable.
>> >
>> > I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
>> > one of CPUs to Flightgear (good), but the smt sibling is used by the
>> > compilation, and I'm down to 9 fps. Not good.
>> >
>> > Even with single-threaded make, I have 10-13fps.
>> >
>> > Is there way to learn which CPUs are SMT siblings?
>>
>> cat /sys/devices/system/cpu/cpu{N}/topology/thread_siblings
>
> Thanks for a hint.
>
> Well, something is definitely wrong there:
>
> pavel@duo:/data/l/linux-n900$ cat
> /sys/devices/system/cpu/cpu0/topology/thread_siblings
> 03
Ops,
do you fly on N900 via flightgear? :)
I have two N900 and one run gentoo.
> I believe that means that cpu0 & cpu1 are sharing physical cpu. Yet,
> when I run flightgear and "nice while1". flightgear goes to CPU#0 and
> while1 to CPU#1. Would not it be nice if while1 went preferably to
> CPUs #2 and #3? Ok. after a while while1 moved to cpu#2, good.
>
>> > Is there way to disable SMT during runtime?
>>
>> You could offline them, but wouldn't it be better to tell each which
>> CPUs they can use, or perhaps partition your box with cpusets?
>
> Let me try offlining first... and yes, SMT seems to be problem here.
>
> | CPU0 | CPU1 | CPU2 | CPU3 | |
> | flightgear | off | off | off | 21 fps |
> | flightgear | while1 | off | off | 17 fps |
> | flightgear+while1 | off | off | off | 10 fps |
> | flightgear | off | while1 | off | 21 fps |
>
>> > Can I do something to improve Flightgear performance and still do
>> > compilation?
>>
>> Sure, run everything associated with your game as RT, and everything
>> not game gets the leftover cycles. If there are none, box will
>> throttle RT to save itself from it's psycho :) driver and you'll know
>> that you need a bigger box. (it's likely your phone)
>
>> Oh yeah, echo NO_RT_RUNTIME_SHARE > /sys/kernel/debug/sched_features
>> before you try that, otherwise the throttle won't help.
>
> Sure, but that is not a problem I have.
>
> Scheduler already (correctly) decides that the game is important, and
> gives game all the time it wants (100% of one CPU). That is not a problem.
>
> Problem seems to be that it schedules other tasks to SMT sibling, and
> that slows down the game quite significantly.
>
> Best regards,
>
> Pavel
> --
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: niced tasks on SMT system
2017-12-28 18:58 ` Ozgur
@ 2017-12-28 20:59 ` Pavel Machek
2017-12-28 21:07 ` Ozgur
0 siblings, 1 reply; 6+ messages in thread
From: Pavel Machek @ 2017-12-28 20:59 UTC (permalink / raw)
To: Ozgur
Cc: Mike Galbraith, Thomas Gleixner, kernel list, mingo@redhat.com,
hpa@zytor.com, x86@kernel.org, peterz@infradead.org
[-- Attachment #1: Type: text/plain, Size: 1306 bytes --]
Hi!
> >> > Ok, so I'm compiling, and I'd like to run a flight simulator.
> >> >
> >> > Flightgear normally does 20fps on my system... kinda low but playable.
> >> >
> >> > I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
> >> > one of CPUs to Flightgear (good), but the smt sibling is used by the
> >> > compilation, and I'm down to 9 fps. Not good.
> >> >
> >> > Even with single-threaded make, I have 10-13fps.
> >> >
> >> > Is there way to learn which CPUs are SMT siblings?
> >>
> >> cat /sys/devices/system/cpu/cpu{N}/topology/thread_siblings
> >
> > Thanks for a hint.
> >
> > Well, something is definitely wrong there:
> >
> > pavel@duo:/data/l/linux-n900$ cat
> > /sys/devices/system/cpu/cpu0/topology/thread_siblings
> > 03
>
> Ops,
>
> do you fly on N900 via flightgear? :)
> I have two N900 and one run gentoo.
No, no flightgear on N900. It was preparing kernels for my N900 that
interfered with flying.
BTW there's lot of fun to be had with
N900. https://wiki.postmarketos.org/wiki/Nokia_N900_(nokia-rx51) I
even have voice calls working on Debian, help welcome :-).
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: niced tasks on SMT system
2017-12-28 20:59 ` Pavel Machek
@ 2017-12-28 21:07 ` Ozgur
0 siblings, 0 replies; 6+ messages in thread
From: Ozgur @ 2017-12-28 21:07 UTC (permalink / raw)
To: Pavel Machek
Cc: Mike Galbraith, Thomas Gleixner, kernel list, mingo@redhat.com,
hpa@zytor.com, x86@kernel.org, peterz@infradead.org
28.12.2017, 23:59, "Pavel Machek" <pavel@ucw.cz>:
> Hi!
>
>> >> > Ok, so I'm compiling, and I'd like to run a flight simulator.
>> >> >
>> >> > Flightgear normally does 20fps on my system... kinda low but playable.
>> >> >
>> >> > I have reniced make -j 5 fo kernel running. Scheduler gives 100% of
>> >> > one of CPUs to Flightgear (good), but the smt sibling is used by the
>> >> > compilation, and I'm down to 9 fps. Not good.
>> >> >
>> >> > Even with single-threaded make, I have 10-13fps.
>> >> >
>> >> > Is there way to learn which CPUs are SMT siblings?
>> >>
>> >> cat /sys/devices/system/cpu/cpu{N}/topology/thread_siblings
>> >
>> > Thanks for a hint.
>> >
>> > Well, something is definitely wrong there:
>> >
>> > pavel@duo:/data/l/linux-n900$ cat
>> > /sys/devices/system/cpu/cpu0/topology/thread_siblings
>> > 03
>>
>> Ops,
>>
>> do you fly on N900 via flightgear? :)
>> I have two N900 and one run gentoo.
>
> No, no flightgear on N900. It was preparing kernels for my N900 that
> interfered with flying.
I would like to be involved in the team and share the job. firstly I will read the project documents.
> BTW there's lot of fun to be had with
> N900. https://wiki.postmarketos.org/wiki/Nokia_N900_(nokia-rx51) I
> even have voice calls working on Debian, help welcome :-).
I use default Maemo, the n900 is already debian, gentoo run a chroot :)
> Pavel
Ozgur
> --
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-12-28 21:07 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-12-28 12:10 niced tasks on SMT system Pavel Machek
2017-12-28 13:23 ` Mike Galbraith
2017-12-28 18:53 ` Pavel Machek
2017-12-28 18:58 ` Ozgur
2017-12-28 20:59 ` Pavel Machek
2017-12-28 21:07 ` Ozgur
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox