public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Dodji Seketeli <dodji@redhat.com>
To: gcc-patches@gcc.gnu.org
Cc: tromey@redhat.com,	gdr@integrable-solutions.net,
	joseph@codesourcery.com,	burnus@net-b.de,	charlet@act-europe.fr,
	paolo@gnu.org,	jason@redhat.com
Subject: [PATCH 7/7] Reduce memory waste due to non-power-of-2 allocs
Date: Sat, 16 Jul 2011 15:34:00 -0000	[thread overview]
Message-ID: <411c73e26bd46ba176c22962851cacf21b013be5.1310824121.git.dodji@redhat.com> (raw)
In-Reply-To: <cover.1310812267.git.dodji@redhat.com>
In-Reply-To: <cover.1310824120.git.dodji@redhat.com>

This patch basically arranges for the allocation size of line_map
buffers to be as close as possible to a power of two.  This
*significantly* decreases peak memory consumption as (macro) maps are
numerous and stay live during all the compilation.

Ideally, I'd prefer some parts of this patch to be integrated into the
memory allocator.  That is, I'd like to see the memory allocator have
an interface that returns the actual size of memory it has allocated.
This would help client code like this one to avoid requesting memory
unnecessarily.  I am proposing this patch as is for now, at least to
see if the approach would be acceptable.

Tested on x86_64-unknown-linux-gnu against trunk.

libcpp/

	* line-map.c (next_largest_power_of_two): New.
	(new_linemap): Use it.
---
 libcpp/line-map.c |   50 +++++++++++++++++++++++++++++++++++++++++++-------
 1 files changed, 43 insertions(+), 7 deletions(-)

diff --git a/libcpp/line-map.c b/libcpp/line-map.c
index d69160c..327ce85 100644
--- a/libcpp/line-map.c
+++ b/libcpp/line-map.c
@@ -95,6 +95,18 @@ linemap_tracks_macro_expansion_locs_p (struct line_maps *set)
   return LINEMAPS_MACRO_MAPS (set) != NULL;
 }
 
+/* Return the smallest power of two that is greater than N.   */
+static unsigned
+next_largest_power_of_two (unsigned n)
+{
+  n |= (n >> 1);
+  n |= (n >> 2);
+  n |= (n >> 4);
+  n |= (n >> 8);
+  n |= (n >> 16);
+  return n+1;
+}
+
 /* Create a new line map in the line map set SET, and return it.
    REASON is the reason of creating the map. It determines the type
    of map created (ordinary or macro map). Note that ordinary maps and
@@ -109,19 +121,43 @@ new_linemap (struct line_maps *set,
   bool macro_map_p = (reason == LC_ENTER_MACRO);
   struct line_map *result;
 
-  if (LINEMAPS_USED (set, macro_map_p) == LINEMAPS_ALLOCATED (set, macro_map_p))
+  if (LINEMAPS_USED (set, macro_map_p)
+      == LINEMAPS_ALLOCATED (set, macro_map_p))
     {
       /* We ran out of allocated line maps. Let's allocate more.  */
+      unsigned alloc_size;
 
       line_map_realloc reallocator
 	= set->reallocator ? set->reallocator : xrealloc;
+
+      /* We are going to execute some dance to try to reduce the
+	 overhead of the memory allocator, in case we are using the
+	 ggc-page.c one.
+	 
+	 The actual size of memory we are going to get back from the
+	 allocator is the smallest power of 2 that is greater than the
+	 size we requested.  So let's consider that size then.  */
+
+      alloc_size =
+	(2 * LINEMAPS_ALLOCATED (set, macro_map_p) +  256)
+	* sizeof (struct line_map);
+
+      alloc_size = next_largest_power_of_two (alloc_size);
+
+      /* Now alloc_size contains the exact memory size we would get if
+	 we have asked for the initial alloc_size amount of memory.
+	 Let's get back to the number of macro map that amounts
+	 to.  */
       LINEMAPS_ALLOCATED (set, macro_map_p) =
-	2 * LINEMAPS_ALLOCATED (set, macro_map_p) + 256;
-      LINEMAPS_MAPS (set, macro_map_p)
-	= (struct line_map *) (*reallocator) (LINEMAPS_MAPS (set, macro_map_p),
-					      LINEMAPS_ALLOCATED (set,
-								  macro_map_p)
-					      * sizeof (struct line_map));
+	alloc_size / (sizeof (struct line_map));
+
+      /* And now let's really do the re-allocation.  */
+      LINEMAPS_MAPS (set, macro_map_p) =
+	(struct line_map *) (*reallocator)
+	(LINEMAPS_MAPS (set, macro_map_p),
+	 (LINEMAPS_ALLOCATED (set, macro_map_p)
+	  * sizeof (struct line_map)));
+
       result =
 	&LINEMAPS_MAPS (set, macro_map_p)[LINEMAPS_USED (set, macro_map_p)];
       memset (result, 0,
-- 
1.7.6

  parent reply	other threads:[~2011-07-16 14:38 UTC|newest]

Thread overview: 134+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-12-10 11:27 [PATCH 0/6] Tracking locations of tokens resulting from macro expansion Dodji Seketeli
2010-12-10 11:16 ` [PATCH 4/6] Support -fdebug-cpp option Dodji Seketeli
2010-12-10 11:16 ` [PATCH 0/6] *** SUBJECT HERE *** Dodji Seketeli
2010-12-10 12:56   ` Dave Korn
2010-12-10 11:27 ` [PATCH 5/6] Add line map statistics to -fmem-report output Dodji Seketeli
2010-12-21  7:30   ` Gabriel Dos Reis
2011-04-13 20:08     ` Dodji Seketeli
2010-12-10 11:27 ` [PATCH 3/6] Emit macro expansion related diagnostics Dodji Seketeli
2010-12-13 15:25   ` Paolo Bonzini
2010-12-13 15:38     ` Paolo Bonzini
2010-12-13 16:30     ` Manuel López-Ibáñez
2010-12-14  7:24     ` Dodji Seketeli
2010-12-14  7:28       ` Gabriel Dos Reis
2010-12-14  8:40         ` Dodji Seketeli
2010-12-14  9:38           ` Gabriel Dos Reis
2010-12-14  9:42             ` Dodji Seketeli
2010-12-14  9:48               ` Gabriel Dos Reis
2010-12-14  7:28     ` Dodji Seketeli
2010-12-14  8:19       ` Gabriel Dos Reis
2010-12-14  8:31         ` Paolo Bonzini
2010-12-14  9:23           ` Dodji Seketeli
2010-12-10 11:53 ` [PATCH 1/6] Linemap infrastructure for virtual locations Dodji Seketeli
2011-01-06 16:48   ` Tom Tromey
2011-04-12 14:39     ` Dodji Seketeli
2011-04-14 14:46       ` Tom Tromey
2010-12-10 12:27 ` [PATCH 2/6] Generate virtual locations for tokens Dodji Seketeli
2010-12-10 12:33 ` [PATCH 6/6] Kill pedantic warnings on system headers macros Dodji Seketeli
2010-12-10 12:52 ` [PATCH 0/6] Tracking locations of tokens resulting from macro expansion Gabriel Dos Reis
2010-12-10 18:22   ` Dodji Seketeli
2010-12-10 16:59 ` Jeff Law
2010-12-10 19:00   ` Dodji Seketeli
2010-12-13 15:10     ` Jeff Law
2010-12-13 16:35       ` Gabriel Dos Reis
2010-12-14  9:24         ` Dodji Seketeli
2010-12-14  9:40           ` Gabriel Dos Reis
2010-12-14  9:45             ` Dodji Seketeli
2011-07-16 14:38 ` [PATCH 0/7] " Dodji Seketeli
     [not found]   ` <cover.1310824120.git.dodji@redhat.com>
2011-07-16 14:38     ` [PATCH 3/7] Emit macro expansion related diagnostics Dodji Seketeli
2011-08-04 15:32       ` Dodji Seketeli
2011-09-12 21:54         ` Jason Merrill
2011-09-16  8:19           ` Dodji Seketeli
2011-09-17 21:27             ` Jason Merrill
2011-09-19 14:37               ` Dodji Seketeli
2011-09-19 19:47                 ` Jason Merrill
2011-09-19 21:27                   ` Jason Merrill
2011-09-20  8:47                     ` Dodji Seketeli
2011-09-20 14:12                       ` Jason Merrill
2011-09-20 14:12                         ` Dodji Seketeli
2011-09-21  2:51                           ` Jason Merrill
2011-09-21 19:09                             ` Dodji Seketeli
2011-09-22 15:32                               ` Jason Merrill
2011-09-26 21:11                                 ` Dodji Seketeli
2011-09-26 22:30                                   ` Jason Merrill
2011-09-27 18:43                                     ` Dodji Seketeli
2011-09-29  7:05                                       ` Jason Merrill
2011-09-29 19:20                                         ` Dodji Seketeli
2011-09-29 21:25                                           ` Jason Merrill
2011-09-29 23:33                                             ` Dodji Seketeli
2011-09-30 15:56                                               ` Jason Merrill
2011-09-30 21:04                                                 ` Jason Merrill
2011-10-03 22:50                                                   ` Dodji Seketeli
2011-10-04 19:59                                                     ` Jason Merrill
2011-10-04 20:35                                                       ` Dodji Seketeli
2011-10-03 20:09                                                 ` Dodji Seketeli
2011-10-04 20:03                                                   ` Jason Merrill
2011-10-04 20:28                                                     ` Dodji Seketeli
2011-09-20  0:08                   ` Dodji Seketeli
2011-07-16 14:38     ` [PATCH 6/7] Kill pedantic warnings on system headers macros Dodji Seketeli
2011-09-12 22:09       ` Jason Merrill
2011-09-16 11:25         ` Dodji Seketeli
2011-09-17 22:34           ` Jason Merrill
2011-09-18 18:59             ` Dodji Seketeli
2011-07-16 14:39     ` [PATCH 5/7] Add line map statistics to -fmem-report output Dodji Seketeli
2011-09-12 22:07       ` Jason Merrill
2011-09-16  8:29         ` Dodji Seketeli
2011-09-17 22:05           ` Jason Merrill
2011-09-20 12:10             ` Dodji Seketeli
2011-09-20 14:08               ` Jason Merrill
2011-07-16 14:39     ` [PATCH 4/7] Support -fdebug-cpp option Dodji Seketeli
2011-08-21 11:02       ` Alexandre Oliva
2011-08-21 11:40         ` Jakub Jelinek
2011-08-22 14:45           ` Tom Tromey
2011-08-22 15:22             ` Jakub Jelinek
2011-08-23 19:52         ` Dodji Seketeli
2011-08-24 15:11           ` Tom Tromey
2011-09-12 22:07       ` Jason Merrill
2011-09-16  8:23         ` Dodji Seketeli
2011-09-17 22:01           ` Jason Merrill
2011-07-16 15:25     ` [PATCH 2/7] Generate virtual locations for tokens Dodji Seketeli
2011-08-09 15:30       ` Dodji Seketeli
2011-09-12 21:15         ` Jason Merrill
2011-09-14 10:01           ` Dodji Seketeli
2011-09-14 22:56             ` Jason Merrill
2011-09-18 13:44               ` Dodji Seketeli
2011-09-19 22:31                 ` Jason Merrill
2011-09-21 14:55                   ` Dodji Seketeli
2011-09-22 17:10                     ` Jason Merrill
2011-09-26 14:47                       ` Dodji Seketeli
2011-09-26 20:39                         ` Jason Merrill
2011-09-28  3:23                           ` Dodji Seketeli
2011-09-28 14:49                             ` Jason Merrill
2011-09-28 21:24                               ` Dodji Seketeli
2011-09-28 21:45                                 ` Jason Merrill
2011-09-29  5:49                                   ` Dodji Seketeli
2011-07-16 15:28     ` [PATCH 1/7] Linemap infrastructure for virtual locations Dodji Seketeli
2011-07-18 22:06       ` Jason Merrill
2011-07-19 10:47         ` Dodji Seketeli
2011-07-19 17:26           ` Jason Merrill
2011-07-19 18:03             ` Dodji Seketeli
2011-07-19 23:37       ` Jason Merrill
2011-07-30  6:20       ` Jason Merrill
2011-08-01 18:54         ` Dodji Seketeli
2011-08-01  4:42       ` Jason Merrill
2011-08-02  4:48       ` Jason Merrill
2011-08-04 15:28         ` Dodji Seketeli
2011-08-04 21:30           ` Jason Merrill
2011-08-05 17:12             ` Dodji Seketeli
2011-08-05 17:31               ` Jason Merrill
2011-08-09 14:56                 ` Dodji Seketeli
2011-08-19  8:46                   ` Jason Merrill
2011-08-19 14:43                     ` Tom Tromey
2011-09-01 10:37                     ` Dodji Seketeli
2011-09-07 19:26                       ` Jason Merrill
2011-09-08 12:41                         ` Dodji Seketeli
2011-09-09  7:45                           ` Jason Merrill
2011-09-09  8:57                           ` Jason Merrill
2011-07-16 15:34     ` Dodji Seketeli [this message]
2011-09-12 22:25       ` [PATCH 7/7] Reduce memory waste due to non-power-of-2 allocs Jason Merrill
2011-09-17 13:41         ` Dodji Seketeli
2011-09-17 22:22           ` Jason Merrill
2011-09-18 22:30             ` Dodji Seketeli
2011-09-19  6:51           ` Laurynas Biveinis
2011-07-16 16:47   ` [PATCH 0/7] Tracking locations of tokens resulting from macro expansion Tobias Burnus
2011-07-16 17:57     ` Dodji Seketeli

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=411c73e26bd46ba176c22962851cacf21b013be5.1310824121.git.dodji@redhat.com \
    --to=dodji@redhat.com \
    --cc=burnus@net-b.de \
    --cc=charlet@act-europe.fr \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=gdr@integrable-solutions.net \
    --cc=jason@redhat.com \
    --cc=joseph@codesourcery.com \
    --cc=paolo@gnu.org \
    --cc=tromey@redhat.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).