On Tue, Jul 15, 2003 at 09:13:46PM -0500, Billy Biggs wrote: > Andrew Suffield (asuffield@debian.org): > > Here's the problem. You don't want to have multiple implementations at > > this level - but you _do_ want the generated code to be different at > > this level. And you need to bridge this gap. > > > > Like most problems with programming, it can be solved by introducing > > another layer of abstraction. The preprocessor is a good one. > > > > [...] > > > > And you're done. Run-time CPU detection with no significant cost. > > Sure, but there is still a cost. Careful, consider the law of limiting factors. Anything that isn't a limiting factor doesn't cost you anything (yet) - and a couple of conditionals, high enough up the tree, won't even be measurable. > You're just advocating pushing it > higher if performance is an issue. Excluding methods that change the problem (usually by skipping irrelevant data), that's essentially all you ever do to optimise code by hand, in one form or another. -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | Dept. of Computing, `. `' | Imperial College, `- -><- | London, UK
Attachment:
pgpXmxLry540I.pgp
Description: PGP signature