> > > Thats not a meaningful test btw. The glibc headers take longer to compile
> > > because of all the extensive standards features and optimisations.
> > >
> > > A valid compile test would be to build say the BSD ones with a BSD
> > > target compiler on BSD and Linux.
> > >
> > > Alan
> >
> > Well, this are the facts with the same hardware different OS.
>
> You could equally say "bash compiles faster on Linux than XFree86 on BSD"
>
> Meaningful benchmarking is quite an art. The Hint values themselves Im sure
> are good benchmark info for some things. The 'it compiles at a different speed'
> benchmark is not unless you use the same compilers headers, linkers and final
> target
I think you miss the point, HINT has nothing to do with compiler headers, linkers,
etc.
HINT has to do with hardware and it limits and how the software (kernel) will
limit the maximum performance at the hardware. You can run HINT from a calculator
to a super computer, and in fact if you browse the HINT database you will find
that kind of computers, has nothing to do compilers and linkers, bullshit, i
compile
it with egcs with pentium opts (pgcc), well i get a box and installed stampede
linux
(all the packages, compilers, linkers, kernel, etc are compiled with pgcc) and the
graph the same, a little performance gain (1M net Quip) so has nothing to do with
that.
What i am pointing is that linux have a extremely poor performance compared to
FreeBSD in tha same machine. I could expect a little loss or gain because is the
same hardware but definitively there is a problem with something in the linux
kernel, what is it? i don't know and what i can expect is that you tell me what is
wrong, we are talking about the same hardware, the only diference is the kernel.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/