[Shootout-list] Ray tracer

John Skaller skaller@users.sourceforge.net
Mon, 13 Jun 2005 17:11:12 +1000


On Mon, 2005-06-13 at 00:15 +0100, Jon Harrop wrote:

> I've noticed that gcc 4.0 produces substantially slower code than gcc 3.4 on 
> this benchmark (on both x86 and AMD64) so I'm wondering if the switch to 4.0 
> isn't too early? Have other benchmarks improved?

That's weird. I noticed a *significant* improvement from gcc-4.0
in my tests. Unfortunately my testing code is temporarily broken,
or I'd post the graphs to back up my claim (the tests runs, but
the output format is completely different). Here is an example:

~/flx> time speed/exes/gccopt_3_4/ack 12
Ack(3,12): 32765

real    0m3.038s
user    0m3.031s
sys     0m0.002s
~/flx> time speed/exes/gccopt_4_0/ack 12
Ack(3,12): 32765

real    0m2.758s
user    0m2.757s
sys     0m0.000s


-- 
John Skaller, skaller at users.sf.net
PO Box 401 Glebe, NSW 2037, Australia Ph:61-2-96600850 
Download Felix here: http://felix.sf.net