Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 30 Aug 2002 08:03:29 -0700
From:      Terry Lambert <tlambert2@mindspring.com>
To:        mark tinguely <tinguely@web.cs.ndsu.nodak.edu>
Cc:        akoskine@cc.helsinki.fi, freebsd-hackers@FreeBSD.ORG
Subject:   Re: More dynamic KVA_SPACE
Message-ID:  <3D6F8941.111A0C39@mindspring.com>
References:  <200208301437.g7UEbxL36566@web.cs.ndsu.nodak.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
Ut-oh.  Time for a rant...

mark tinguely wrote:
> Another interesting processor family is the AMD x86-64 ClawHammer.
> I do not know the progress the FreeBSD/x86-64 project. I would imagine
> the major difficulty will be getting a running compiler.

Actually, the major difficulty is getting a box, given that these
things were originally supposed to been out a while ago, and AMD
keeps not selling the damn things.  Running simulated hardware on
hardware you can already buy off the street really doesn't have a
very big thrill factor, since you might as well run not simulated,
get the same boot messages and programs running, only much, much
faster.

For all we know, AMD has gotten out of the new Silicon business all
together, and will simply be shipping upgraded simulators of things
which would be cool if they were ever taped out, plactic'ed, and
slotted in a ZIF in non-existant motherboards, but which, in reality,
never make it to glass.  Sort of the chip equivalent of the .COM
business model ("Here's my impression of an eyeball").


> I just wish AMD added an 8K page size so the Page Table Maps did not
> eat so much memory.

More proof that hardware people don't ask software people how they
expect to use the hardware after it's built.  Same class of problem
that the Diamond video cards had ...i.e. BIOS written by a hardware
person that didn't have a data table at a findable offset with a version
number, so you could only INT 10 the thing to guarantee that you would
not cook the card and/or your monitor.  They just didn't believe
software people would ever want or need to use an interface that did
not disable hardware interrupts for the whole system in order to avoid
vertical blanking, since the most important component of any computer
is the sparklie-free Diamond video card lacking dual ported RAM.

Any hardware vendor who creates a polled interface for their chips
is doing the same thing ...i.e. the most important thing any computer
can be doing at any point in time is asking our chip if it has more
data available for processing, because it's not like the CPU has any
other work to do, and even if it did, it's nowhere near as important
as talking to our hardware.

Hardware is so annoying.  So unlike software... 8-) 8-) 8-).

-- Terry

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D6F8941.111A0C39>