Author: Eugene Nalimov
Date: 23:29:42 02/01/00
Go up one level in this thread
On February 02, 2000 at 02:15:02, Peter McKenzie wrote: >On February 01, 2000 at 23:25:44, Eugene Nalimov wrote: > >>*Always* search for the bug in your code first. In the 95% of cases it's there. >>not in the optimizer. > >If 5% of my bugs were actually compiler bugs, I'd be returning the compiler very >quickly! 5% of the "mysterious" bugs, that appears only when you turn on the maximum optimizations. Eugene >> >>Eugene >> >>On February 01, 2000 at 22:10:19, KarinsDad wrote: >> >>>On February 01, 2000 at 16:22:44, Jon Dart wrote: >>> >>>[snip] >>>> >>>>I have had a significant amount of trouble with the VC++ >>>>optimizer. Even with all the service packs installed, I have >>>>still seen cases where it generates bad code. It's not that >>>>you can't use it, but you may have to reduce the optimization >>>>level for parts of your code (with consequent loss in >>>>performance). You also get the fun of figuring out if your code >>>>is wrong or if the compiler is screwing it up :-). But you may >>>>or may not encounter this problem. >>> >>>One way to "attempt to" detect this problem is to use multiple compilers. There >>>should be enough "stability" in your program that an optimized compiler does not >>>make much difference when investigating problems over a non-optimized one. If >>>the optimized one has problems when the non-optimized one (or the same compiler >>>optimized versus non-optimized) does not, suspect the optimization before >>>suspecting bugs in your code. >>> >>>KarinsDad :)
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.