linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Michel Lespinasse <walken@google.com>
To: riel@redhat.com, peterz@infradead.org, daniel.santos@pobox.com,
	aarcange@redhat.com, dwmw2@infradead.org,
	akpm@linux-foundation.org
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	torvalds@linux-foundation.org
Subject: [PATCH v2 0/9] faster augmented rbtree interface
Date: Thu,  2 Aug 2012 15:34:09 -0700	[thread overview]
Message-ID: <1343946858-8170-1-git-send-email-walken@google.com> (raw)

These are my proposed changes for a faster augmented rbtree interface.
They are implemented on top of a previous patch series that is already
in Andrew's -mm tree, and I feel they are ready to join it.

Patch 1 is a trivial fix for a sparse warning.

Patch 2 is a small optimization I already sent as part of my previous RFC.
Rik had ACKed it.

Patches 3-4 are small cleanups, mainly intended to make the code more readable.

Patches 5-6 are new, based on something George Spelvin observed in my
previous RFC. It turns out that in rb_erase(), recoloring is trivial for
nodes that have exactly 1 child. We can shave a few cycles by handling it
locally, and changing rb_erase_color() to only deal with the no-childs case.

Patch 7 adds a performance test for the augmented rbtree support.

Patch 8 introduces my proposed API for augmented rbtree support.
rb_insert_augmented() and rb_erase_augmented() are augmented versions of
rb_insert_color() and rb_erase(). They take an additional argument
(struct rb_augment_callbacks) to specify callbacks to be used to maintain
the augmented rbtree information. users have to specify 3 callbacks
through that structure. Non-augmented rbtree support is provided by
inlining dummy callbacks, so that the non-augmented case is not affected
(either in speed or in compiled size) by the new augmented rbtree API.
For augmented rbtree users, no inlining takes place at this point (I may
propose this later, but feel this shouldn't go with the initial proposal). 

Patch 9 removes the old augmented rbtree interface and converts its
only user to the new interface.


Overall, this series improves non-augmented rbtree speed by ~5%. For
augmented rbtree users, the new interface is ~2.5 times faster than the old.

Michel Lespinasse (9):
  rbtree test: fix sparse warning about 64-bit constant
  rbtree: optimize fetching of sibling node
  rbtree: add __rb_change_child() helper function
  rbtree: place easiest case first in rb_erase()
  rbtree: handle 1-child recoloring in rb_erase() instead of
    rb_erase_color()
  rbtree: low level optimizations in rb_erase()
  rbtree: augmented rbtree test
  rbtree: faster augmented rbtree manipulation
  rbtree: remove prior augmented rbtree implementation

 Documentation/rbtree.txt |  190 ++++++++++++++++++++----
 arch/x86/mm/pat_rbtree.c |   65 ++++++---
 include/linux/rbtree.h   |   23 ++-
 lib/rbtree.c             |  370 +++++++++++++++++++++++++---------------------
 lib/rbtree_test.c        |  135 ++++++++++++++++-
 5 files changed, 557 insertions(+), 226 deletions(-)

-- 
1.7.7.3

--
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>

             reply	other threads:[~2012-08-02 22:34 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-02 22:34 Michel Lespinasse [this message]
2012-08-02 22:34 ` [PATCH v2 1/9] rbtree test: fix sparse warning about 64-bit constant Michel Lespinasse
2012-08-05 20:47   ` Rik van Riel
2012-08-02 22:34 ` [PATCH v2 2/9] rbtree: optimize fetching of sibling node Michel Lespinasse
2012-08-05 23:20   ` Rik van Riel
2012-08-06 14:23   ` Peter Zijlstra
2012-08-06 20:46     ` Michel Lespinasse
2012-08-02 22:34 ` [PATCH v2 3/9] rbtree: add __rb_change_child() helper function Michel Lespinasse
2012-08-06  1:00   ` Rik van Riel
2012-08-06 14:22   ` Peter Zijlstra
2012-08-07  0:04     ` Michel Lespinasse
2012-08-02 22:34 ` [PATCH v2 4/9] rbtree: place easiest case first in rb_erase() Michel Lespinasse
2012-08-06  1:13   ` Rik van Riel
2012-08-02 22:34 ` [PATCH v2 5/9] rbtree: handle 1-child recoloring in rb_erase() instead of rb_erase_color() Michel Lespinasse
2012-08-06  1:27   ` Rik van Riel
2012-08-02 22:34 ` [PATCH v2 6/9] rbtree: low level optimizations in rb_erase() Michel Lespinasse
2012-08-06  1:41   ` Rik van Riel
2012-08-06 14:21   ` Peter Zijlstra
2012-08-06 20:50     ` Michel Lespinasse
2012-08-06 20:58       ` Peter Zijlstra
2012-08-06 21:20         ` Michel Lespinasse
2012-08-06 21:21           ` Michel Lespinasse
2012-08-02 22:34 ` [PATCH v2 7/9] rbtree: augmented rbtree test Michel Lespinasse
2012-08-06  2:11   ` Rik van Riel
2012-08-02 22:34 ` [PATCH v2 8/9] rbtree: faster augmented rbtree manipulation Michel Lespinasse
2012-08-06  2:12   ` Rik van Riel
2012-08-06 14:17   ` Peter Zijlstra
2012-08-06 15:38     ` Peter Zijlstra
2012-08-06 21:55       ` Michel Lespinasse
2012-08-07  7:12         ` [PATCH] rbtree: add RB_DECLARE_CALLBACKS() macro Michel Lespinasse
2012-08-06 14:25   ` [PATCH v2 8/9] rbtree: faster augmented rbtree manipulation Peter Zijlstra
2012-08-06 21:34     ` Michel Lespinasse
2012-08-06 21:36       ` Peter Zijlstra
2012-08-06 14:29   ` Peter Zijlstra
2012-08-06 21:38     ` Michel Lespinasse
2012-08-02 22:34 ` [PATCH v2 9/9] rbtree: remove prior augmented rbtree implementation Michel Lespinasse
2012-08-02 22:41   ` Joe Perches
2012-08-06  2:13   ` Rik van Riel

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=1343946858-8170-1-git-send-email-walken@google.com \
    --to=walken@google.com \
    --cc=aarcange@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=daniel.santos@pobox.com \
    --cc=dwmw2@infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=peterz@infradead.org \
    --cc=riel@redhat.com \
    --cc=torvalds@linux-foundation.org \
    /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;
as well as URLs for NNTP newsgroup(s).