linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Jesper Dangaard Brouer <brouer@redhat.com>
To: linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>
Cc: netdev@vger.kernel.org,
	Jesper Dangaard Brouer <brouer@redhat.com>,
	Alexander Duyck <alexander.duyck@gmail.com>,
	Pekka Enberg <penberg@kernel.org>,
	David Rientjes <rientjes@google.com>,
	Christoph Lameter <cl@linux.com>,
	Joonsoo Kim <iamjoonsoo.kim@lge.com>
Subject: [PATCH 7/7] slub: do prefetching in kmem_cache_alloc_bulk()
Date: Mon, 28 Sep 2015 14:26:39 +0200	[thread overview]
Message-ID: <20150928122639.15409.21583.stgit@canyon> (raw)
In-Reply-To: <20150928122444.15409.10498.stgit@canyon>

For practical use-cases it is beneficial to prefetch the next freelist
object in bulk allocation loop.

Micro benchmarking show approx 1 cycle change:

bulk -  prev-patch     -  this patch
   1 -  49 cycles(tsc) - 49 cycles(tsc) - increase in cycles:0
   2 -  30 cycles(tsc) - 31 cycles(tsc) - increase in cycles:1
   3 -  23 cycles(tsc) - 25 cycles(tsc) - increase in cycles:2
   4 -  20 cycles(tsc) - 22 cycles(tsc) - increase in cycles:2
   8 -  18 cycles(tsc) - 19 cycles(tsc) - increase in cycles:1
  16 -  17 cycles(tsc) - 18 cycles(tsc) - increase in cycles:1
  30 -  18 cycles(tsc) - 17 cycles(tsc) - increase in cycles:-1
  32 -  18 cycles(tsc) - 19 cycles(tsc) - increase in cycles:1
  34 -  23 cycles(tsc) - 24 cycles(tsc) - increase in cycles:1
  48 -  21 cycles(tsc) - 22 cycles(tsc) - increase in cycles:1
  64 -  20 cycles(tsc) - 21 cycles(tsc) - increase in cycles:1
 128 -  27 cycles(tsc) - 27 cycles(tsc) - increase in cycles:0
 158 -  30 cycles(tsc) - 30 cycles(tsc) - increase in cycles:0
 250 -  37 cycles(tsc) - 37 cycles(tsc) - increase in cycles:0

Note, benchmark done with slab_nomerge to keep it stable enough
for accurate comparison.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
---
 mm/slub.c |    2 ++
 1 file changed, 2 insertions(+)

diff --git a/mm/slub.c b/mm/slub.c
index c25717ab3b5a..5af75a618b91 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -2951,6 +2951,7 @@ bool kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size,
 				goto error;
 
 			c = this_cpu_ptr(s->cpu_slab);
+			prefetch_freepointer(s, c->freelist);
 			continue; /* goto for-loop */
 		}
 
@@ -2960,6 +2961,7 @@ bool kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size,
 			goto error;
 
 		c->freelist = get_freepointer(s, object);
+		prefetch_freepointer(s, c->freelist);
 		p[i] = object;
 
 		/* kmem_cache debug support */

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2015-09-28 12:26 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-09-28 12:26 [PATCH 0/7] Further optimizing SLAB/SLUB bulking Jesper Dangaard Brouer
2015-09-28 12:26 ` [PATCH 1/7] slub: create new ___slab_alloc function that can be called with irqs disabled Jesper Dangaard Brouer
2015-09-28 12:26 ` [PATCH 2/7] slub: Avoid irqoff/on in bulk allocation Jesper Dangaard Brouer
2015-09-28 12:26 ` [PATCH 3/7] slub: mark the dangling ifdef #else of CONFIG_SLUB_DEBUG Jesper Dangaard Brouer
2015-09-28 13:49   ` Christoph Lameter
2015-09-28 12:26 ` [PATCH 4/7] slab: implement bulking for SLAB allocator Jesper Dangaard Brouer
2015-09-28 15:11   ` Christoph Lameter
2015-09-28 12:26 ` [PATCH 5/7] slub: support for bulk free with SLUB freelists Jesper Dangaard Brouer
2015-09-28 15:16   ` Christoph Lameter
2015-09-28 15:51     ` Jesper Dangaard Brouer
2015-09-28 16:28       ` Christoph Lameter
2015-09-29  7:32         ` Jesper Dangaard Brouer
2015-09-28 16:30       ` Christoph Lameter
2015-09-29  7:12         ` Jesper Dangaard Brouer
2015-09-28 12:26 ` [PATCH 6/7] slub: optimize bulk slowpath free by detached freelist Jesper Dangaard Brouer
2015-09-28 15:22   ` Christoph Lameter
2015-09-28 12:26 ` Jesper Dangaard Brouer [this message]
2015-09-28 14:53   ` [PATCH 7/7] slub: do prefetching in kmem_cache_alloc_bulk() Alexander Duyck
2015-09-28 15:59     ` Jesper Dangaard Brouer
2015-09-29 15:46 ` [MM PATCH V4 0/6] Further optimizing SLAB/SLUB bulking Jesper Dangaard Brouer
2015-09-29 15:47   ` [MM PATCH V4 1/6] slub: create new ___slab_alloc function that can be called with irqs disabled Jesper Dangaard Brouer
2015-09-29 15:47   ` [MM PATCH V4 2/6] slub: Avoid irqoff/on in bulk allocation Jesper Dangaard Brouer
2015-09-29 15:47   ` [MM PATCH V4 3/6] slub: mark the dangling ifdef #else of CONFIG_SLUB_DEBUG Jesper Dangaard Brouer
2015-09-29 15:48   ` [MM PATCH V4 4/6] slab: implement bulking for SLAB allocator Jesper Dangaard Brouer
2015-09-29 15:48   ` [MM PATCH V4 5/6] slub: support for bulk free with SLUB freelists Jesper Dangaard Brouer
2015-09-29 16:38     ` Alexander Duyck
2015-09-29 17:00       ` Jesper Dangaard Brouer
2015-09-29 17:20         ` Alexander Duyck
2015-09-29 18:16           ` Jesper Dangaard Brouer
2015-09-30 11:44       ` [MM PATCH V4.1 " Jesper Dangaard Brouer
2015-09-30 16:03         ` Christoph Lameter
2015-10-01 22:10         ` Andrew Morton
2015-10-02  9:41           ` Jesper Dangaard Brouer
2015-10-02 10:10             ` Christoph Lameter
2015-10-02 10:40               ` Jesper Dangaard Brouer
2015-10-02 13:40             ` Jesper Dangaard Brouer
2015-10-02 21:50               ` Andrew Morton
2015-10-05 19:26                 ` Jesper Dangaard Brouer
2015-10-05 21:20                   ` Andi Kleen
2015-10-05 23:07                     ` Jesper Dangaard Brouer
2015-10-07 12:31                       ` Jesper Dangaard Brouer
2015-10-07 13:36                         ` Arnaldo Carvalho de Melo
2015-10-07 15:44                           ` Andi Kleen
2015-10-07 16:06                         ` Andi Kleen
2015-10-05 23:53                   ` Jesper Dangaard Brouer
2015-10-07 10:39                   ` Jesper Dangaard Brouer
2015-09-29 15:48   ` [MM PATCH V4 6/6] slub: optimize bulk slowpath free by detached freelist Jesper Dangaard Brouer
2015-10-14  5:15     ` Joonsoo Kim
2015-10-21  7:57       ` Jesper Dangaard Brouer
2015-11-05  5:09         ` Joonsoo Kim

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150928122639.15409.21583.stgit@canyon \
    --to=brouer@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=alexander.duyck@gmail.com \
    --cc=cl@linux.com \
    --cc=iamjoonsoo.kim@lge.com \
    --cc=linux-mm@kvack.org \
    --cc=netdev@vger.kernel.org \
    --cc=penberg@kernel.org \
    --cc=rientjes@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox