[rescue] SMP on intel wasteful?

Pablo Ruiz Garcia pruiz at ip6seguridad.com
Tue Jun 25 01:54:49 CDT 2002


----- Original Message -----
From: <ssandau at bath.tmac.com>
To: <rescue at sunhelp.org>
Sent: Monday, June 24, 2002 11:40 PM
Subject: Re: [rescue] SMP on intel wasteful?


> My analogy (and I'm sticking to it) is that x86 hardware is like a 2-stroke
> engine and real servers are like diesels. Each has its place, but you
certainly
> would NEVER attempt to power a bulldozer with a 2-stroke. It may be fast,
but
> loads KILL it.

What impressed my of x86 was that thereis lot's of improvements for code
on x86 platform, while i usually does not see that code improvements developed
on others platforms.

Think for example on MMX, there are lots of "geeks" making optimizations of
software (especially free software) making optimizations with MMX and so..
while, i have not seen that for Sparc.. I know it's not a big example, but I
got a
SUN Ultra60 with 512MB of ram, and I love it, but my PentiumII 400 runs
John the ripper more acuratelly and fast than the ultra :?

>
> In addition, I have never had an IRQ or plug&puke problem with a SPARC. I've
> also never been able to boot an x86 box with a serial terminal. Netbooting
is
> MUCH easier with a SPARC or a HP9000 (only ones I own...) and so on.
>

Problems with Plug&play or HW incompatibility is always caused by the very big
market that have been created around the PeeCee computer, and so, there are so
much
companies seelling so many diferents kinds of products PCI boards, USB
adaptors, etcetc

With Sun/HP Hardware every board or adaptor has been tested to work correctly
with theyr
own hardware, cause this, you sould not have any kind of problem of
incompatibility. (note: you *should*) ;)

> Each does have its place, but x86 was NEVER meant to be a server, and it
shows.
> Man, they don't even have 16 REAL IRQs. The last bunch are just patched onto
> the first bunch!
>

I know that, but you can think on SMP APIC boards for x86 ;)

PD: Some time ago i have seen two diferente BIOS which talks directly to
serial
port, (the first was a Lab desing) and the second if cant remember correctly
was
the one on watchguard Firebox products.
PeeCee not talking to serial directlyis not a PeeCee limitation, but a bios
one.

> > > > I've always found those nebulous arguments against PeeCee hardware
> > > > interesting.  Usually unfounded in recent fact, but interesting
> > > > nonetheless.  When put to the test though I've found that dollar for
> > > > dollar a PeeCee running UNIX will spank a RISC box running UNIX.
> > >
> > >   *bzzzt*  Troll alert.
> > >   *bzzzt*  Don't-know-what-CPAI-is alert.
> > >   *bzzzt*  Possible remote Intel Marketing Clone alert.
> >
> > Sticking your fingers in your ears and saying "la la la" won't make x86 go
> > away.  Believe me, I've tried.  Sorry if the following is grammatically
> > poor, I'm just rambling.
> >
> > Bottom line, x86 is fast.  It is as ugly as sin and has some nasty
> > compromises in it, but it is fast.  The reason for this is dead simple,
> > it's a commodity architecture that discourages stratification within it.
> > Let me relate a little story that I feel has some relevance to this
> > discussion.
> >
> <sorry, snipped>
>
>
> -------------------------------------------------
> This mail sent through IMP: http://horde.org/imp/
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue
>
--
Pablo Ruiz Garcia (Pci)
Consultor de Seguridad - Tiger Team
IP6 Seguridad (pruiz at ip6seguridad.com)
http://www.ip6seguridad.com/pgp/pruiz.asc



More information about the rescue mailing list