GNU bug report logs - #17136
[PATCH] dfa: speed up memory allocation, port to IRIX

Previous Next

Full log

View this message in rfc822 format

From: help-debbugs <at> gnu.org (GNU bug Tracking System)
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: tracker <at> debbugs.gnu.org
Subject: bug#17136: closed ([PATCH] dfa: speed up memory allocation, port
 to IRIX)
Date: Fri, 18 Apr 2014 06:00:03 +0000

[Message part 1 (text/plain, inline)]

Your message dated Thu, 17 Apr 2014 22:59:17 -0700
with message-id <5350BF35.5040407 <at> cs.ucla.edu>
and subject line Re: bug#17136: [PATCH] dfa: speed up memory allocation, port to IRIX
has caused the debbugs.gnu.org bug report #17136,
regarding [PATCH] dfa: speed up memory allocation, port to IRIX
to be marked as done.

(If you believe you have received this mail in error, please contact
help-debbugs <at> gnu.org.)


-- 
17136: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=17136
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems

[Message part 2 (message/rfc822, inline)]

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: grep mailing list <bug-grep <at> gnu.org>
Subject: [PATCH] dfa: speed up memory allocation, port to IRIX
Date: Fri, 28 Mar 2014 18:28:57 -0700

[Message part 3 (text/plain, inline)]

Tags: patch

I'm not going to install the attached memory-management changes yet, as 
they're fairly intrusive and I want to give them a chance to cool off. 
The benefits are minor but are there, e.g., this shrinks the overall 
grep text size by 1.2% on my platform (Fedora 20 x86-64).  Mostly, 
though, I wanted the memory management to be clearer.

[0001-dfa-speed-up-memory-allocation-port-to-IRIX.patch (text/plain, attachment)]

[Message part 5 (message/rfc822, inline)]

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: 17136-done <at> debbugs.gnu.org
Subject: Re: bug#17136: [PATCH] dfa: speed up memory allocation, port to IRIX
Date: Thu, 17 Apr 2014 22:59:17 -0700

[Message part 6 (text/plain, inline)]

I merged that patch into the grep trunk and am closing this bug report. 
 While merging, I removed one part of the patch that had a small but 
measurable performance degradation, namely, the introduction of struct 
dfatab.  I found some other minor simplifications and threw them in. 
There seems to be no significant change in CPU performance.  The patch's 
causes the dfa module's text size to shrink by 8% on my platform (x86-64 
Fedora 20), which I suppose is a small win, but the main point of the 
patch is simplifying the source.  I'm attaching what I installed, broken 
into 12 little patches to ease review.

[0001-dfa-clarify-memory-allocation-and-port-to-IRIX.patch (text/plain, attachment)]

[0002-dfa-avoid-unnecessary-work-and-other-initialization.patch (text/plain, attachment)]

[0003-dfa-better-size-overflow-check.patch (text/plain, attachment)]

[0004-dfa-simplify-transition-table-allocation.patch (text/plain, attachment)]

[0005-dfa-simplify-range-char-allocation.patch (text/plain, attachment)]

[0006-dfa-simplify-multibyte_prop-allocation.patch (text/plain, attachment)]

[0007-dfa-simplify-position-set-and-element-count-allocati.patch (text/plain, attachment)]

[0008-dfa-simplify-memory-allocation.patch (text/plain, attachment)]

[0009-dfa-avoid-duplicate-strlen-when-allocating-memory.patch (text/plain, attachment)]

[0010-dfa-simplify-freelist.patch (text/plain, attachment)]

[0011-dfa-simplify-dfmust-initialization.patch (text/plain, attachment)]

[0012-dfa-trans-reallocation-microoptimization.patch (text/plain, attachment)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.