FASTEST AMD K5?!

[edwardauskis]

Hi all!

What do you think about this?

[Neon]

Ya, quite confusing.
On a socket 7 motherboard with 66 MHz bus, it would need to set the multiplier at 3.75 to give PR250. I am not aware of any socket 7 motherboard with 3.75x multiplier.
It would perhaps run internally at 166 MHz. This seems plausible, but K5 architecture struggled to reach 133 MHz. I have not heard of any that can run at 166 MHz.
From a marketing standpoint, PR250 would compete against Pentium 200? Pentium II 233? K6-233?
Several K5 ES that I have seen have different print font than the production models. Look at http://cdn.cpu-world.com/CPUs/K5/S_AMD-AMD-K5-PR166ABQ%20(ES).jpg and note the font is stretched higher vertically. Also "ES" is placed in the 4th line of text, not the 3rd line.

[edwardauskis]

only what ca I say, this is not from a China resource.you here more biggest specialist when Im

[frag_]

Except multiplier (it's not a problem at all, it's entirely an inner CPU thing) I totally agree with Neon.
Speed room is the main problem. K5 had practical limit at ~133MHz.
There was almost no room to raise frequncy without extreme cooling and higher voltages.
Note than 3.52V is already the max for almost all s7 motherboards.
Without die shrink or at least new revision almost none of them can do e.g. 150.
It was too late and didn't make any sense to make new K5 revision by the summer of 1997. There were K6 already.
Yes, PR rating is also strange. Should be 233 or 266.
Is there any K5 ES with normal, unstretched font? I can't find any.
Also known samples have no "MALAYSIA" word.
And why 1997 copyright and not 1996?
Imho it's fake.
Proving it is very easy - buy it and theck the real multiplier and frequency +)

What is the source of this photo?

[edwardauskis]

this photos sent me seller who sale scrap lot. this lot have some nice example,so I will not divulge it

[frag_]

I you win it you must make hi-res scans and test it

[Neon]

With the K5, the PR value indicates how to set the motherboard multiplier, even though there was internal multiplier that would change that value. So:
PR100 motherboard 1.5x =100 internal 1.0x = 100
PR133 motherboard 2.0x =133 internal 0.75x =100
PR166 motherboard 2.5x =166 internal 0.70x = 117
PR200 motherboard 3.0x =200 internal 0.67x = 133
PR250 motherboard 3.75x =250 ?

[xsecret]

looks fake to me. Typo is wrong for an ES (that's the retail typo), ES misplaced and looks like photoshoped on front (background is a little brighter) but more important : package code is missing (should be 25676).

[debs3759]

Looks fake to me as well. I'm sure if it was real that someone would have seen it before.

[frag_]

[frag_]

But still, this logic is broken for PR150 as it has 66 bus.
150/66=2.25

[edwardauskis]

I dont think so if photoshoped, this seller does not do that, but if its fake can be.

[black_monolith]

Probably fake, and if so a recent fake. Adding the "ES" would imply it is targetted at CPU collectors.

It's still worth checking out. Have the seller fire it up and see if it works at PR250.

[Neon]

PR120 and PR150 are supposed to run on 60 MHz bus

PR120 motherboard 2.0x =120 internal 0.75x = 90
PR150 motherboard 2.5x =150 internal 0.70x = 105

I think that in practice, some of the PR150 could run at PR166 speed.

[frag_]

Yes, both are physically run with 60MHz bus.
I mean, why PR150 have "66MHz" printed on the lid?
If consumer supposed to set fsb and multi jumpers on the motherboard based on the information printed on cpu,
then it would give right setings for any K5 except PR150.
PR150 would be overclocked to PR166 frequency and likely not stable.