Mail Archives: djgpp/1998/12/09/07:16:50
Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> wrote:
> On Mon, 7 Dec 1998, Salvador Eduardo Tropea (SET) wrote:
>
> > In some days I'll upload a new table using geometric average, that's a little
> > bit better for comparisson in this case (can be much better in other cases).
>
> What is especially significant is that, while the average performance is
> almost identical, there's some very large (40% or more!)
Yes that's clear in the graphic that contains all.
> differences in
> several individual tests. It would be interesting to know whether these
> tests have any practical implications (e.g., it might be that those tests
> use code which will rarely happen in real-world programs).
All the algorithms are real-world ones. They are just very well known
algorithms. Some of them are floating point intensive, others are pure
integer. All are relative complex and messure the CPU and compiler.
> If the tests are of practical importance, I think it would be interesting
> to know what are the cases where GCC-generated code is much slower than
> Watcom's, and vice versa. From time to time, people ask here questions
> about relative performance, and all we usually tell them is ``almost
> identical''.
Lamentably I don't have a Watcom compiler to see the assembler generated (I
don't want to disassemble a huge EXE ;-). But if somebody can make it we can
learn a lot looking at the routines that Watcom optimizes better.
But perhaps we can't get the best of both because the
differences are in different trade-offs in the design of the optimizer.
SET
------------------------------------ 0 --------------------------------
Visit my home page: http://set-soft.home.ml.org/
or
http://www.geocities.com/SiliconValley/Vista/6552/
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Alternative e-mail: set-soft AT usa DOT net set AT computer DOT org
ICQ: 2951574
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA
TE: +(541) 759 0013
- Raw text -