Mail Archives: djgpp-workers/1999/06/29/13:32:39
Jan Hubicka wrote:
> > > so mregparm seems to be win.
> >
> > Yes it left gcc 2.95 very close to MSVC 5.0.
> Really? I didn't know that differences to MSVC are so small.
They are very small, in the case of the numeric sort if I use regparm MSVC is a little bit
faster just because the prolog of the function (I mean the push and assigment of the
registers before entering in the body) is generated in a better way (exploiting the
super-scalar features of newer processors).
The string sort have a huge difference, but after using my memove the different was
reverted (now gcc is clearly faster), I'll investigate the rest until I find where the
differences are. But I don't have much time so I'll do it slowly.
> I am getting much
> larger speedups using new backend. Also string functions can be optimized
> a bit. After this we can actually beat MSVC :))
People in the egcs list are confident they can beat MSVC 6.0 with the new IA32 code. I'll
try to check out a copy of this branch and try to compile it.
> Whats about using the string functions and math inlines from glibc.
GLIBC is GPL and DJ's library don't (is GPL but with some modifications to let people link
statically without releasing sources).
> they are much better optimized IMO and especially math inlines brings
> large speedups in apps using them. Are the licences compatible?
I doubt it.
> BTW I am pretty short on time. Are you interested in helping me to finish the
> changes? I really need some help :)
I can't, sorry, I have a lot of things to do in my editor and in the middle time I'm
running these benchmarks...
SET
--
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: set-soft AT usa DOT net set AT computer DOT org
set AT ieee DOT org set-soft AT bigfoot DOT com
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013
- Raw text -