BG> Thought you might be interested in the benchmarks of my AMD DX5/133 CPU,
BG> now that I have a motherboard which suits the chip.  Pity it's an AMI
BG> BIOS, but at least it seems to be extremely stable at 40MHz (CPU runs
BG> {at} 160MHz), especially as I'm running 2 x VLB cards in this machine.

RS> Yeah, certainly one of the better buys around at the moment.

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.

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

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

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

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

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

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

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

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

Yeah, interesting to watch, and interesting how cheap say the 486DX2/66s are to
@EOT:

---