Re: [RFC][PATCH] Scalable Scheduling

[Hubertus Franke]

Obviously not. Take a look at the presentation from Ottawa Linux Symposium.
We ran this on 2-way for various applications from kernel builds to TPC-H
apps
and it actually slightly improved performance.
It has detailed overviews of lock contention and wait times.

start at http://lse.sourceforge.net/scheduling/ols2001/img41.htm if you are
lazy, but
the MQ description really is a good overview and describes some of the data
structures

At 2-ways we see it at worst to be a wash for the benchmarks we used.

Hubertus Franke
Enterprise Linux Group (Mgr),  Linux Technology Center (Member Scalability)
, OS-PIC (Chair)
email: [e-mail]
(w) 914-945-2003    (fax) 914-945-4425   TL: 862-2003

Victor Yodaiken <[e-mail]> on 08/08/2001 03:51:43 PM

To:   Hubertus Franke/Watson/IBM@IBMUS
cc:   Victor Yodaiken <[e-mail]>, [e-mail]
Subject:  Re: [RFC][PATCH] Scalable Scheduling

On Wed, Aug 08, 2001 at 03:40:00PM -0400, Hubertus Franke wrote:
>
> We did not modify the UP code at all. There will be NO effects (positive
> nor negative) what so ever.

Cool. So the obvious next question is
     How does it compare on a dual to the current Linux scheduler?

Obviously:
     Performance sucks on two processors but scales well

would not be a good thing.

>
> Hubertus Franke
> Enterprise Linux Group (Mgr),  Linux Technology Center (Member
Scalability)
> , OS-PIC (Chair)
> email: [e-mail]
> (w) 914-945-2003    (fax) 914-945-4425   TL: 862-2003
>
>
>
> Victor Yodaiken <[e-mail]> on 08/08/2001 03:27:55 PM
>
> To:   Mike Kravetz <[e-mail]>
> cc:   Linus Torvalds <[e-mail]>, Hubertus
>       Franke/Watson/IBM@IBMUS, [e-mail]
> Subject:  Re: [RFC][PATCH] Scalable Scheduling
>
>
>
> On Wed, Aug 08, 2001 at 11:28:00AM -0700, Mike Kravetz wrote:
> > One challenge will be maintaining the same level of performance
> > for UP as in the current code.  The current code has #ifdefs to
>
> How does the "current code" compare to the current Linux UP code?
>
>
>
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
in
> the body of a message to [e-mail]
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/