Article View: gnu.gcc.help
Article #42495Re: performance measurements across 8 platforms: atoi()/sprintf("%d")
From: Axel Bauer
Date: Mon, 27 Nov 1995 00:00
Date: Mon, 27 Nov 1995 00:00
22 lines
1337 bytes
1337 bytes
Hello, > I timed the attached program on a number of platforms to compare the > relative performance of their system library versions of atoi() and > sprintf(str,"%d",i) against a hand-coded C version. In most cases, it > was easy to beat the system library-- perhaps losing some generality, > but I'm surprised how much better the simple hand-coded C version does > on some platforms. This generality shouldn't cost you more than a > factor of 2, should it? You didn't mention how you measured the performance. You didn't mention wether all the executables were linked statically. Maybe the time spent in the dynamic linkage process was included in some measurements ? Axel PS: if this is being sent twice, blame Netscape -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ a x e l b a u e r
Technical Details
Message-ID:
<gnusenet30B8F8D2.167E@netplace.com>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<DI229s.AA7@beaver.cs.washington.edu>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 11318c,41bf6a7c37c19df3
X-Google-Attributes: gid11318c,public
From: axel@netplace.com (Axel Bauer)
Subject: Re: performance measurements across 8 platforms: atoi()/sprintf("%d")
Date: 1995/11/27
Message-ID: <gnusenet30B8F8D2.167E@netplace.com>
X-Deja-AN: 120326623
distribution: gnu
sender: daemon@cis.ohio-state.edu
references: <DI229s.AA7@beaver.cs.washington.edu>
organization: Netplace Internet Consulting
newsgroups: gnu.gcc.help