> how do we verify these figures?

You think BigGun might be a crook? Now come on... ;-) I don't think the SysInfo screenshot is doctored. What we don't know though is how reliably the chip runs at 120 MHz. It can do a brief benchmarking, but can it run under full load for hours at 120 MHz?

> if they've managed to get those speeds they're not going to want to talk about it.

Yes, the faster a physical 060 chip can go on the NatAmi the smaller the performance advantage of their self-developed N68050/N68070 FPGA implementation is going to be.

This also does not eliminate the possibility that these are re-labeled chips.
Without an MMU or FPU these could be easier to overclock.
It would really hurt to know that the Chinese figured this out and it never occurred to the of us.
But surface mount chips aren't something a hobbyist would normally consider (when there are socketed alternative).

You did notice that the Sysinfo screens did seem to indicate a lack of MMU and FPU functions, didn't you?

The lack of floating point functions is a significant minus for some software.

This whole mystery just just murkier the more its examined.

[ Edited by Jim 06.06.2011 - 15:27 ]

> This also does eliminate the possibility that these are re-labeled chips.

Does it? A full 68060 Rev.6 can be overclocked from 50 or 60 MHz to just over 100 MHz. That's 70+% overclocking. Running a 75 MHz part at 120 MHz only amounts to 60% overclocking. Even running at 133 MHz amounts to no more than 78% overclocking.

> Without an MMU or FPU these could be easier to overclock.

What do you mean "overclock"? Attempting to run the part past 133 MHz? Running a 133 MHz part at 120 MHz ist *under*-clocking, assuming the "possibility that these are re-labeled chips" is eliminated.

> surface mount chips aren't something a hobbyist would normally consider
> (when there are socketed alternative).

Yes, but are there any socketed alternatives to a 133 MHz 060FE? Anything like an 68(EC)060RC133?

> You did notice that the Sysinfo screens did seem to indicate a
> lack of MMU and FPU functions, didn't you?

Of course I did. Or how do you think I concluded it's the "MC68060FE133" board he ran SysInfo on?

No, No, No. That's an error/typo.

It should be
This also does not eliminate the possibility that these are re-labeled chips.

>What do you mean "overclock"? Attempting to run the part past 133 >MHz? Running a 133 MHz part at 120 MHz ist *under*-clocking, >assuming the "possibility that these are re-labeled chips" is >eliminated.

No, I'm still pretty sure these are re-labeled chips. So over, not under.

>Yes, but are there any socketed alternatives to a 133 MHz 060FE? >Anything like an 68(EC)060RC133?

That 133 designation is really getting to you, isn't it?
Its fiction. And no, there are no EC parts with 133 speed ratings (and, for that matter,there are no FE chips rated to 133).

If the Natami team supports this fiction, then they may offend Freescale.


[ Edited by Jim 06.06.2011 - 15:33 ]

> No, No, No. That's an error/typo.

Ah okay, thanks for clarification.

> I'm still pretty sure these are re-labeled chips. So over, not under.

Okay, I fell for your error/typo here obviously.

> That 133 designation is really getting to you, isn't it?

I was arguing from a "does eliminate the possibility that these are re-labeled chips" standpoint here.

> there are no EC parts with 133 speed ratings (and, for that matter,
> there are no FE chips rated to 133).

I'm well aware of what Motorola/Freescale chips are rated to.

> If the Natami team supports this fiction, then they may offend Freescale.

I think the NatAmi is and will forever stay way below Freescale's radar level. Besides, the consumer version isn't supposed to have a physical 060 chip anyway.

Update:

> What we don't know though is how reliably the chip runs at 120 MHz.

From a NatAmi Team member in the previously linked thread:

"It's currently running stable at 120mhz."

Time to figure out an adapter for CSPPCs and other accelerators that have socketed 68060s.

But then again, the lack of an fpu is somewhat crippling.

> Time to figure out an adapter for CSPPCs and other accelerators
> that have socketed 68060s.

If the "MC68060FE133" is a rebadged MC68EC060FE75 that is able to run overclocked at 120 MHz then an MC68EC060RC75 (which wouldn't need an adapter) should do as well, shouldn't it?

That would be worth exploring. I wonder if a custom math library could be created to make up for the loss of the fpu and how much of a performance hit that that tactic would create?

Jim,
Quote:

For that matter, there are no 68060 chips in FE packaging, at least according to Motorolla's (albeit Q4 2001) product selection guide.

@ boot_wb

There are FE packaged 68060s, but this 133 designation is oh so dubious.

Oh and get this, while I'm used to a few inaccuracies in Freescale listings, "Additional Features Floating Point Unit (FPU) Memory Management Unit (MMU)", their MC68EC060RC75 listing credits the processor with two features it doesn't have.

> there are no 68060 chips in FE packaging

There are:

http://www.google.com/search?q=site:freescale.com+mc68060+cqfp

Put out feelers for EC chips from distributors within the US (still a little scared of Chinese suppliers).
I should have quotes by tomorrow.

> Put out feelers for EC chips from distributors within the US

Freescale recommends http://www.rocelec.com for the MC68EC060RC75 and says they have inventory of "> 1000" of this chip.

Wow! I couldn't get that link to work.
Thanks Andreas.
That quantity leaves some room for price negotiation.


[ Edited by Jim 06.06.2011 - 18:11 ]

> That quantity leaves some room for price negotiation.

Direct link to all their 060RC chips:

http://www.rocelec.com/about/searchresults/060rc/

There're 1480 pcs. in stock currently. Seems they're differentiating manufacturer between Freescale and Motorola Semiconductor.

Thanks Andreas,
I asked them for a price quote too.

edit - A company called US Bid just contacted me. They should have a price by tomorrow (although they also gave me a quote on HD6309EPs I'd been looking for).

[ Edited by Jim 06.06.2011 - 23:28 ]

Well, even after a call from US Bid yesterday, no one has sent me any prices yet.
I really expected that at least Rochester Electronics (Freescale's recommended vendor)would get back to me.
Guess I'll give it a few days.
It's not vital anyway as I don't have anything to test it in (the only 68K based system I've still own has an '030 in it).

Only one of the suppliers I wrote to has responded so far, first with only a question of how many I needed, and then the second reply said that they did not have any available.

I wonder if they became "not available", because I was also asking for where they came from, since Freescale claims they are not valid Motorola, or Freescale parts, and also asked if they had an MMU & FPU?

Could be they decided they were not able to supply them because of the MMU & FPU requirement I imposed on the quote.

I've received similar responses from Chinese suppliers. (asking how many were needed)
I don't think they can meet the one requirement you made, as the Natami team's Sysinfo posts show that the "FE133" they have doesn't have an MMU or FPU.

Thomas does claim to have received a stock pile of these though, so they must intend to build some '060 boards with them.

Still no luck on the EC components. I'll re-issue inquiries later today.

> the Natami team's Sysinfo posts show that the "FE133" they have doesn't
> have an MMU or FPU.

This was already shown almost two years ago with another software (see the last link therein).

Over the past couple of years I've bought chips direct from Chinese Manufactures... If you sent your quote request to enough companies you will definitely get multiple quotes... I've found that sometimes only one company has the chip and 20 companies will send me quotes based on what they think they can get out of me after paying the "one" actual company that has stock. always worked for me to be patient and go thru all the quotes... especially for older chips that may not be currently manufactured.

I know that the test results from the Natami team showed that their MC68060FE133 chips do not have MMU or FPU, but Thomas probably built only a few of these surface mount 68060 CPU boards with these chips.

More than one of the sites advertising these chips show them as having both the MMU and FPU. The advertising is probably incorrect, and all of these chips are without MMU and FPU.

Still, it is possible that Thomas and the Natami team could have bought a batch of faulty MC68060FE133's at a discounted price that were known to have non-operational MMU & FPU units, but I doubt this is the case. Surface mount parts are probably very difficult to test, until after they have been soldered to a board.

You can get swappable sockets for them. Was looking at the possibility of this for repairing a cyberstorm in the past - allow swapping the cplds to find the faulty one.

Needless to say the price put me off. (and the footprint was too large anyway). Useful for programming/testing qfp chips though...

>You can get swappable sockets for them. Was looking at the possibility of this for repairing a cyberstorm in the past - allow swapping the cplds to find the faulty one.

Needless to say the price put me off. (and the footprint was too large anyway). Useful for programming/testing qfp chips though...

I've seen similar sockets for BGAs, but they too were far too costly to justify (unless used for testing purposes).

Edit - First price quote in on 68EC060RC75 and I am dumbfounded.
US BID will supply me 3Mhz Hitachi HD63C09EPs for less than $13, but they quoted me $400 for the '060.

Seems a little steep.
I'll have to look elsewhere.

[ Edited by Jim 11.06.2011 - 21:52 ]