From Steve Dyson · 1996-04-08 22:17 · #75252 in R25C
In reply to Speeding up C by Matthew Probert
-=> Quoting Matthew Probert to All <=- Matthew, An interesting contribution though I notice you don't mention the Pareto effect. Where usually 90% of the processing time is spent on 10% of the code (and 99% of misguided optimization efforts are only 1% effective! :-) One piece in you missive I would like to raise (or should that be rase)... MP> NB: Although it is normal practice for expressions to be evaluated MP> left to right, the ANSI standard on C does not specify an order of MP> preference for expression evaluation, and as such you should check MP> your compiler manual. IMHO such practice should be shunned by all programmers of any worth! To justify this remark I would point out that any such code can cause such havok and cost so dearly when the target hardware changes or when the development environment changes. If the order of evaluation is imperative and if it is not covered in the recognized standard then either the expression should be broken into smaller expressions or brackets should be used. I beleive that what you are advocating is a recipe for nervous breakdowns for maintenance staff years/months/days? hence. Regards, Steve. --- Blue Wave/RA v2.12 * Origin: From The Crystal Tower : London (0181-4478244) (2:254/220)