On Monday 19 May 2008 15:52:22 Felipe Sateler wrote:
> I'll run a test here and see if it makes a difference. I have a Core 2 Duo
> with Linux 64 bit installed. I'll be using trapped.csd to benchmark.

OK, I've done a few tests, using 5.08. They are done without optimization, 
with gcc4opt=generic and gcc4opt=native. The result is that without 
optimization there is not much difference (2.1% slower with doubles), 
presumably because there is too much time spent dealing with memory and other 
things.
With optimization, however, there is a lot of difference. doubles are around 
30% faster than floats. 

An interesting thing I saw is that gcc4opt=native and gcc4opt=generic did not 
make any difference (0.2%). Possibly this will be different in i386 since it 
has been extended a lot since the original i386, meanwhile the amd64 arch has 
not changed much yet.

-- 
Felipe Sateler