BG> The really strange thing is that once jumpered to take the DX5, the bootup
BG> screen reports the chip as being an "Am5x86/P75", but then
AMI could put
BG> anything they like in their BIOS without it actually being 100% correct.

RS> Turns out its a bit tricky working out what cpu is actually installed,
RS> particularly with the more exotic ones, and the bios authors dont generally
RS> try too hard. Its all rather a can of worms, the CPUID instruction stuff
RS> has turned into the usual abortion you always see in computing, without
RS> a nice clean allocation of that data, particularly with the non Intel cpus.

BG> You'd be horrified to see all the jumpers I had to set
BG> before the board would recognise the chip as a quadrupler,
BG> and this was on a separate sheet, not in the printed manual.

Yeah, thats the downside with those off the mainstream a bit, can be a bit
of a problem with that stuff. Some are twitchy about the cache setting too.

BG> Surely the chipset and/or BIOS manufacturers could
BG> make a board which automatically detected the CPU
BG> type and speed, a bit like "plug and play" for CPUs.

Part of the problem is that the manufacturers of cpus have been
playing silly buggers with some of the pins on 486 pinout cpus.

They've even been doing brain farts on the non 5V voltages too.

The industry is absolutely notorious for this massive brain fart stuff.

BG> This new board already auto-detects the CPU voltage, which I though
BG> was rather elegant.  Pity about the god-awful WinBIOS though.

Dunno, they all have their downsides. Those fucking Awards which have
the same version number, even right down to the trailing subversion
letter letter, some with and some without LBA support are a total
pain in the arse coz you cant be sure of LBA support from the version.

RS> There is a crying need for something
RS> as robust as the ID stuff with IDE drives.

BG> Yeah, couldn't agree more,

Yeah, bit more tricky with a cpu tho coz obviously
it has to be useable to be able to ID anything much.

BG> although if you don't set up the IDE parameters in BIOS, this board
BG> will also auto-detect the drive size at boot-up, but it does tend
BG> to add a rather noticeable delay to the boot time, as you'd expect.

Yeah, we really are seeing some decent progress on
all that stuff now. No doubt there will be plenty
of warts and fangs in the arse getting there tho.

RS> Presumably they are just taking the easy way out, not attempting to keep
RS> up with the newer more exotic cpu variants and just decide its an oddy.

BG> Dunno, the ability to use a DX5 is definitely BIOS-related here.

Sure, I meant more the status display on bootup, rather than the settings.

BG> Some of the fellows in the internet newsgroups have successfully
BG> run the DX5 with a 50MHz clock (200MHz internally) on PCI boards,

RS> I'm gunna report them to the RSPCC. You want to watch out, it is
RS> possible to kill a cpu if you get too carried away. Which can ruin
RS> the day rather more than somewhat with the more expensive cpus.

BG> So people keep telling me, but just like unapproved modems,
BG> nobody can actually name anybody to whom this has happened.

Vernon Frazee in the International Tech echo has
killed some. Its certainly very uncommon to do so tho.

BG> but mine doesn't like that at all with VLB.

RS> Yeah, thats what you would expect. Do you even know that motherboard
RS> is happy with a 50MHz external speed, say with a 486DX50 ?

BG> I'd almost bet my house that it wouldn't work.  Maybe
BG> it would with all ISA, but not with my two VLB cards.

There are some that will, but its not common at all.

BG> Still, it makes you wonder just how much more they can squeeze
BG> from 486 technology.  Seems it isn't totally obsolete yet.

RS> Yeah, interesting to watch, and interesting
RS> how cheap say the 486DX2/66s are too.

BG> I bought my original i486DX2/66 second-hand only ~2 years
BG> ago for $450, and you can now buy one new for just $65.

Yeah, not sure its a good idea to do this sort of thing
tho, one of these days I might cark it in apoplexy. For
some reason the price I paid for a 200MB drive, $1350, sticks
in my mind. Shit, I could get around 5GB for that today.

BG> Even the DX5/133 can be had for less than $120 these
BG> days.  Pity the Pentiums haven't yet followed suit.

Its always like that, whatever Intel has a monopoly on, they
gouge like mad to recover their R&D costs while they can.
@EOT:

---