Date: Thu, 15 May 2003 23:14:43 -0700 From: Terry Lambert <tlambert2@mindspring.com> To: Wes Peters <wes@softweyr.com> Cc: freebsd-hackers@freebsd.org Subject: Re: Optimizations. Message-ID: <3EC481D3.124F61B8@mindspring.com> References: <20030515123033.GP45118@garage.freebsd.pl> <FHEMJMOKKMJDGKFOHHEPEEILGLAA.mooneer@translator.cx> <200305151407.28717.wes@softweyr.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Wes Peters wrote: > On Thursday 15 May 2003 08:19, Pawel Jakub Dawidek wrote: > > Messy code? Why? People skilled in optimization know which parts of > > code should be really optimized and which aren't important from > > performance point of view at all (code profiling?). Good optimized > > code shouldn't be ugly, ehh. > > People skilled in optimization know to profile the code and find what is > slowing real, actual applications down before attempting to profile > anything. You can't imagine how uninterested any FreeBSD developer is > in optimizing getuid() so it will out-perform Linux on some completely > braindead benchmark. Show me an application that is getuid() > performance bound and I'll show you an application that is so poorly > written as to be meaningless. What about "gettimeofday()"? There are a number of applications, particularly HTTP protocol applications, with strict timestamp logging requirements, mandated by standards. The tend to call gettimeofday() up to six times per transaction. Examples will include HTTP proxies, L7 load balancers, and proxy caches. Any box competing in the Cisco CSS/F5 falls into this category. Personally, I think the discussion about compiler option in this thread is useless (want a good compiler? Join the FSF GCC project, and do your work in that context, instead of FreeBSD), but there are valid targets of opportunity for optimizing things, above and beyond "good micro benchmark scores". -- Terry
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3EC481D3.124F61B8>