[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <b7e5ce7837acad87b54fee5e30e0e6d0ea0b35ac.camel@gmx.de>
Date: Sun, 04 Jul 2021 07:37:53 +0200
From: Mike Galbraith <efault@....de>
To: Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
Vlastimil Babka <vbabka@...e.cz>
Cc: linux-mm@...ck.org, linux-kernel@...r.kernel.org,
Christoph Lameter <cl@...ux.com>,
David Rientjes <rientjes@...gle.com>,
Pekka Enberg <penberg@...nel.org>,
Joonsoo Kim <iamjoonsoo.kim@....com>,
Thomas Gleixner <tglx@...utronix.de>,
Mel Gorman <mgorman@...hsingularity.net>,
Jesper Dangaard Brouer <brouer@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Jann Horn <jannh@...gle.com>
Subject: Re: [RFC v2 00/34] SLUB: reduce irq disabled scope and make it RT
compatible
On Sat, 2021-07-03 at 17:47 +0200, Mike Galbraith wrote:
> On Sat, 2021-07-03 at 09:24 +0200, Mike Galbraith wrote:
> > It also appears to be saying that there's something RT specific to
> > stare at in addition to the list_lock business.
>
> The what is ___slab_alloc() consuming 3.9% CPU in tip-rt-slub whereas
> it consumes < 1% in both tip-rt (sans slub patches) and tip-slub.
>
> The why remains to ponder.
Ignoring odd distribution in the profile (red herring whackable via
prefetch games), and removing perf/debug options from the picture, the
remaining RT specific cost is just that of forced slow path. For
hackbench in my little box, it adds up to right at 3%.
-Mike
Powered by blists - more mailing lists