GNU bug report logs - #23474
Gzip and CFLAGS

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#23474: closed (Gzip and CFLAGS)
Date: Sat, 07 May 2016 02:53:02 +0000

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

Your message dated Fri, 6 May 2016 19:52:26 -0700
with message-id <572D586A.6050309 <at> cs.ucla.edu>
and subject line Re: bug#23474: Gzip and CFLAGS
has caused the debbugs.gnu.org bug report #23474,
regarding Gzip and CFLAGS
to be marked as done.

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


-- 
23474: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=23474
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems

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

From: Tom Smith <gnuacc2016 <at> gmail.com>
To: bug-gzip <at> gnu.org
Subject: Gzip and CFLAGS
Date: Fri, 6 May 2016 21:15:15 -0400

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

Hello there,


            After having compiled gzip, I had looked at the Makefile
“./configure” generated.

            Looking at CFLAGS=”” (system configuration section), not many
compiler flags are specified for optimization.

            The “Specific Targets” section sets different flags for
different systems, which are set accordingly.

            Currently, I am looking at “CFLAGS= -O”.


            I noticed that some optimization flags were not set (curious as
to why):

-         -march=native (pentium3, athlon64, etc) (produces code for the
specific architecture specified)

-         -O2 (recommended level to use to increase code performance )

-         -pipe (makes compilation process faster)

-         -fomit-frame-pointer (reduces generated code size)


Also, I found this article
http://lists.gnu.org/archive/html/bug-gzip/2013-02/msg00000.html and it
mentions setting –O2 and encountering an issue (it has since been fixed).

After running some tests after I recompiled (Linux), sometimes I noticed it
to be a little bit faster (at times) on x86_64.

Note: There are many different conditions that cause the numbers to
fluctuate, such as: # of processes running, ram usage, CPU power etc.




Thank You

[Message part 4 (text/html, inline)]

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

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Tom Smith <gnuacc2016 <at> gmail.com>, 23474-done <at> debbugs.gnu.org
Subject: Re: bug#23474: Gzip and CFLAGS
Date: Fri, 6 May 2016 19:52:26 -0700

Tom Smith wrote:

>             Currently, I am looking at “CFLAGS= -O”.

You don't mention which gzip version. If you haven't already, please try the
current version, gzip 1.8. Its 'configure' script should default to -O2
optimization for GCC, and to no optimization for other compilers. To some extent
this is arbitrary. Tweaking this further is up to you. For example,
-march=native is better for some usages but not for others, which is why it is
not the default.

Previous Next

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