Re: ramblings about old hardware, gzip, bz2, and pentium opts
On Mon, Nov 29, 1999 at 05:35:07PM +0100, Tomasz Wegrzanowski wrote:
> Im very unsure about this :
> `pentium optimisations slows down 99% of all systems'
> You claim that Celeron / Cx 6x86 / PIII / K6 / Merced etc.
> are faster with 386-optimized soft than with Pentium-optimized soft
> Do you have any stats to prove it ???
I ran a povray benchmark recently on a number of systems with all
variety of -march= and -mcpu= flags, gcc 2.95.2. [evil grin] I'll email you
the results. }:}
For everyone else I'll just summarize: no statistitically significant
variation with architecture-specific compilation flags. Statistically
insignificant occurances where the *wrong* architecture appeared slightly
faster (eg -mcpu=k6 opts making a P-II run faster than with -mcpu=ppro).
-Drake (who has been waiting to pound some Pentiumize-Debian! poster over
the head... :)