[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Pine.LNX.4.64.0801051206580.14866@schroedinger.engr.sgi.com>
Date: Sat, 5 Jan 2008 12:16:22 -0800 (PST)
From: Christoph Lameter <clameter@....com>
To: Matt Mackall <mpm@...enic.com>
cc: Ingo Molnar <mingo@...e.hu>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Pekka Enberg <penberg@...helsinki.fi>,
Hugh Dickins <hugh@...itas.com>,
Andi Kleen <andi@...stfloor.org>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] procfs: provide slub's /proc/slabinfo
On Fri, 4 Jan 2008, Matt Mackall wrote:
> > SLUB 32 (all memory of the 4k page is used for 128 byte objects)
> > SLAB 29/30 (management structure occupies first two/three objects)
> > SLOB 30(?) (Alignment results in object being 136 byte of effective size,
> > we have 16 bytes leftover that could be used for a
> > very small allocation. Right?)
>
> Don't know how you got to 136, the minimum alignment is 4 on x86. But I
Right I am thinking about 64 bit systems where the alignment is 8 bytes.
> already said in my last email that SLUB would win for the special case
> of power of two allocations. But as long as we're looking at worst
> cases, let's consider an alloc of 257 bytes..
Yup that hits it by forcing a rounding up to a size of 512 bytes because
there is no intermediate cache size before 1024. The rounding up is
a pretty weak spot in terms of memory use.
> SLUB 8 (1016 bytes wasted)
> SLOB 15 (105 bytes wasted, with 136 bytes still usable)
Well we can actually turn this around. What I gave was not actually the
worst case for SLOB. The worst case is an 8 byte allocation where SLOB
needs double the memory of SLUB.
SLUB 512 (Nothing wasted)
SLOB 256 (Half of the page wasted for metadata)
SLAB 119 (32 byte mininum alloc size + management struct needs)
But these are all extreme cases. Depends on the mix of allocs who wins and
from what I can tell the avoiding of rounding up to a power of two gives
SLOB a key advantage. If we would find the worst offenders there and use
kmem_cache_alloc instead then we may be able to offset that advantage.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists