>>> it [...] is confirming suspicions that Nema is just a revision of
>>> another P6T design

>> Varisys' pre-Nemo PA6T design is a known product and Nemo surely can't
>> be regarded as a revision of this

> Nemo is the realization of specs created by Ack Systems and Amiga Inc.

As you know, a mere feature list is tremendously far from a working board design. Furthermore, a direct comparison between real Nemo features and ACK's published feature list shows these differences:

- CPU GHz: 1.8 vs. 2.0
- PEG: 2 (1x16 or 2x8) vs. 1 (x16)
- PCIe: 2x1 vs. 1x2+1x1
- PCI: 2 vs. 1
- GbE: 1 vs. 2
- USB2: 10 vs. 6
- PATA: 1 vs. none
- CF: 1 vs. none
- I/O: SB600 southbridge vs. dedicated chips for SATA2 and USB2
- custom CPU+I/O: Xena+Xorro vs. none
- serial/debug: 2xRS232+1xJTAG vs. 2xUART
- Firmware: CFE vs. U-Boot

The only feature that is exactly the same is the RAM slots.

Quote:

And the type of processor (except for its precise speed).
And that was part of what lead me to contact Paul.

So...I give full credit to Trevor for digging up exactly the right company for the job.

And the SB600 came as no surprise to either of us as we had already been discussing its potential use with the MPC8640/8641 (by that I mean US, not the entire forum).

How many years has this been, btw?

> I don't know that it is enough to afford to pay salaries to several employees.

It's not necessarily been paid for full-time working hours.

> As to sourcing the PA6T cpus, as a previous customer, Paul might have been
> able to secure these on his own

Yes, that's what's strange about Trevor's (or Paul's, if correctly reported by Trevor) story. Varisys had purchased PA6T chips before, so should have been eligible for Apple's "final call".

> Trevor HAS paid Varisys for the development of three motherboard.

It's even four: Nemo, mATX Cyrus, ATX Cyrus+, Tabor.

> I [...] would even lay the decision to use the P1022 at Paul's feet - it
> was the type of cpu he stated his preference for over the PA6T

Availability and compatibility aside, 1.2 GHz e500v2 doesn't seem like proper replacement for 1.8 GHz PA6T.

> that preference probably helped cinch the decision to use the
> P5020 and P5040 on X5000 motherboards.

That wasn't a question of preference. In 2012/2013, there simply was no other sensible CPU choice to make for a new, high-performance (in PPC terms, that is) desktop PPC board than QorIQ P5. (I've never seen any PPC476FP-based SoC for sale on the respective manufacturer's website or at semiconductor distributors, so minimum purchase quantity would likely have been too high for A-Eon anyway.)

> I give full credit to Trevor for digging up exactly the right company for the job.

They were recommended to A-Eon by P.A.Semi, so half credit should suffice ;-)

> the SB600 came as no surprise to either of us as we had already been discussing
> its potential use with the MPC8640/8641 (by that I mean US, not the entire forum).

Others were earlier:

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=4306&forum=11&start=7
https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=6124&forum=32&start=12

:-)

> How many years has this been, btw?

About 6, I guess.

>> I don't know that it is enough to afford to pay salaries to several employees.

>It's not necessarily been paid for full-time working hours.

True enough. After all, condider the Friedens arrangement with Hyperion.

>> Trevor HAS paid Varisys for the development of three motherboard.

>It's even four: Nemo, mATX Cyrus, ATX Cyrus+, Tabor

Also the X3500 board, and any design explorations that were not pursued.

>>I [...] would even lay the decision to use the P1022 at Paul's feet - it
>> was the type of cpu he stated his preference for over the PA6T

>Availability and compatibility aside, 1.2 GHz e500v2 doesn't seem like proper replacement for 1.8 GHz PA6T.

I got the impression that Paul's preference was one related to pricing, availability, and the end cost of the product.
And he didn't specifically mention the e500v2 core, just the Qorlq line.



[ Edited by Jim 20.05.2016 - 10:12 ]

>>that preference probably helped cinch the decision to use the
>>P5020 and P5040 on X5000 motherboards.

>That wasn't a question of preference. In 2012/2013, there simply was no other sensible CPU choice to make for a new, high-performance (in PPC terms, that is) desktop PPC board than QorIQ P5. (I've never seen any PPC476FP-based SoC for sale on the respective manufacturer's website or at semiconductor distributors, so minimum purchase quantity would likely have been too high for A-Eon anyway.)

Yes IBM products are unobtainable useless you are a "qualified" buyer.
IBM's marketing methods must be modeled after the Stanley brothers.

There were no Freescale cpus at the time that could match the PA6T (P5s were not available at the time).

And Applied Micro's announced developments often never saw the light of day, or the products capabilities were scaled down.


As to the SB600, that would be my hubris showing.
I should not have underestimated the knowledgeably of some my fellow posters.
At the time, I knew the SB600 was a functional equivalent of the ULi Southbridge used in Freescale evaluation designs.
It had less pins, so there was more multiplexing of pin inputs and outputs to produce the same connectivity.
As such, it wasn't a direct swap, requiring some re-engineering.
But it was the only real logical choice (as the SB450 was not really competitive with ULi components or the later SB600).

So as a natural choice, my default to it was not that original.



[ Edited by Jim 20.05.2016 - 11:10 ]

> Also the X3500 board

That was supposed to be the same board as the one for the X5000.

> he didn't specifically mention the e500v2 core, just the Qorlq line.

Okay, I mistook what you wrote for meaning he recommended the QorIQ P1 line (which only comes with e500v2) as PA6T replacement. Back then, QorIQ line consisted of P1, P2 and P4, i.e. e500v2 and e500mc. I guess he had the e500mc-based P4 in mind in terms of PA6T replacement.

>> I've never seen any PPC476FP-based SoC for sale on the respective manufacturer's
>> website or at semiconductor distributors

> Yes IBM products are unobtainable useless you are a "qualified" buyer.

I was more referring to LSI than IBM.

>> In 2012/2013, there simply was no other sensible CPU choice to make for a new,
>> high-performance (in PPC terms, that is) desktop PPC board than QorIQ P5.

> There were no Freescale cpus at the time that could match the PA6T
> (P5s were not available at the time).

Not in 2010 when Paul Gentle recommended QorIQ over PA6T, yes, but they were in 2012/2013 when the Cyrus specs were conceived. I won't back away from my opinion that even without Paul's recommendation, A-Eon would have decided for the P5 in 2012/2013, simply because there was no feasible alternative for high-performance (in PPC terms, that is) desktop PPC.

>A-Eon would have decided for the P5 in 2012/2013

What else would you have for a high end PPC system then?
Of course you are right.

And it was the core everyone was interested in at that point.

I still find Tabor confusing.
Edit - Sorry anout that last comment.
I let myself drift into general commentary on A-eon boards.

[ Edited by Jim 21.05.2016 - 12:16 ]

Quote:

I think you meant to write Paul Gentle recommended PA6T over QorIQ, not the other way around, in 2010.

It is not often that you make any posting mistakes, so I will take this rare opportunity to point out this slip up, in return for the many times you have corrected my mistakes in the past. (I have made very many)

[ Edited by amigadave 24.05.2016 - 20:21 ]

>> in 2010 when Paul Gentle recommended QorIQ over PA6T

> I think you meant to write Paul Gentle recommended PA6T over QorIQ,
> not the other way around, in 2010.

Doesn't "recommend A over B" mean "recommend A instead of B", similar to how "prefer" would be used ("prefer A over B" or "prefer A to B")?

Jim in April 2010:
"Just yesterday I contact a UK company that builds PPC boards (mainly because the head of the company had designed an XMOS based PCIe sound card). He basically discouraged me from looking into the PA Semi based board he had designed because he was having trouble obtaining the processors."

Jim in July 2010:
"Recently when I discussed PPC based systems with the owner of Varisys (the designers of the X1000 motherboard) I got this response: "In terms of PPC we are thinking QorIQ here for higher end designs.""

amigadave in March 2013:
"I wonder why the people responsible for making the design decisions for the final X1000 motherboard design ignored Paul's advice to stay away from the PA6T and look at using one of the QorIQ chips instead?"

The majority of my posts are laced with opinion and conjecture.
But these comments our fabulous indexer has repeated are not some of those.
Paul specifically recommended the Qorlq line several years ago and as things have developed it still is a good recommendation.

edit - I do not normally quote email messages, but to clarify I want to post the second line from that old message.

"Hi Jim,

In terms of PPC we are thinking QorIQ here for higher end designs.

http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=P4040..."

So, the P4040 with its e500mc cores is what was originally referenced.
Yes, it is not as powerful as a PA6T.
But look how the Qorlq line has expanded.

And right now NXP continues to license technology that would work well with future PPC Qorlq cpus (but will probably be used on their ARM cpus).

Postscript - I was surprised to find that I still had those.

So...it is my assumption that Trevor had a specific cpu (PA6T) in mind and went looking for a design firm with experience with that cpu.
I don't think it was Paul's recommendation.

[ Edited by Jim 25.05.2016 - 20:11 ]

>> I guess he had the e500mc-based P4 in mind in terms of PA6T replacement.

> the P4040 with its e500mc cores is what was originally referenced.

Thanks for confirmation.

> it is my assumption that Trevor had a specific cpu (PA6T) in mind and went
> looking for a design firm with experience with that cpu.

No need to assume when that's what Trevor has been telling all along. As said, Varisys were recommended to A-Eon by P.A.Semi according to Trevor.

> I don't think it was Paul's recommendation.

Yes, according to Trevor, the PA6T was what the OS4 developers wanted. Paul's recommendation was the addition of the xCORE stuff.

>...Paul's recommendation was the addition of the xCORE stuff.

Yes, as I have mentioned it was an Xcore design that originally set me off looking for Paul.
And once I saw what his company was marketing I contacted him for advice.

So...when Nemo was announced I had my suspicions about a link between A-eon and Varisys that was later confirmed.

A fun period.
Just when you thought that PPCs didn't have much of a future in desktop computing A-eon manages to produce a run of boards with a rare PPC that Apple helped kill.
Then Freescale develops not one but two 64bit PPC cores.
And here we are, still discussing the immanent death of the PPC while IBM promotes Power8.

As usual, my mistake.

I thought that the QorIQ CPU's were not yet available for Paul Gentle to recommend in 2010, so thought that you had gotten the two CPU's reversed in your sentence.

Also, great work finding that quote of mine (which I of course can't remember from a thread I was only mildly interested in long ago, and since I have trouble with many aspects of my memory, specially relatively short term memory, these days. Getting older is so much fun!)

Quote:

Yes the memory thing IS so much fun.
I have classes starting in July and while I did great in those I took in the past (the last almost 20 years ago) my recall isn't what it used to be.
It will be challenging.

As to the cpu choice on Nemo, you are lucky that Trevor pressed for his preference.
You have a neat system.

> when Nemo was announced I had my suspicions about a link between A-eon and Varisys

You were in contact with Varisys already in January 2010?

Quote:

Yes, from late the year before.
By February of 2010 Paul had handed my inquiries off to Adam Barnes his Technical Director.

The quote I posted is from April 2010.

I wasn't looking for A-eon's design team.
Rather, I was just interested in Paul's designs.
And I thought I could use his advice with the MPC8640/8641D project I was working on.
Although I must admit to being skeptical when Nemo was announced as so many announcements in the Amiga community are pure vapor.

[ Edited by Jim 27.05.2016 - 01:07 ]

Addendum:

> drivers for [...] onboard Ethernet are still not completed.

The long story as told by Hyperion director Costel 'Cyborg' Mincea five days ago:

http://translate.google.com/translate?sl=de&tl=en&u=http://www.os4welt.de/viewtopic.php?p=32421%23p32421 (relevant part starts at 5th paragraph)

Quote:

How or why anyone would work with any of these companies involved is beyond me. That thread is validating ever single negative yet obvious statement that has been said over the last 10+ years. 2018 most certainly appears to be the year anything and everything OS4 related dies off painfully.

Quote:

Kudos to Cyborg to bringing it to nearly functional state (and dont know why inclusion on FE disc).

Find it astonishing he was never sent another Nemo board to complete it. I would send mine to Germany
to get thing done.

Sound driver was done, net driver, warp3D for all Radeon HD, Timberwolf done and Libre beta (CFE bugfixing ...) and it could be way more pleasent experience then it was.

I dont like OS4 will die out, but its easy to unerstand why and how.


[ Edited by vox 22.01.2018 - 18:30 ]

> Cyborg [...] was never sent another Nemo board to complete it.

If you take the time to read his posting to the end you'll find that he writes he was sent a replacement by A-Eon in 2015.

read the costel post make me really happy with my choice of No More A-eon in My home...
now i have a micro A1 the true real amigaone machine, the quad g5 the best powerpc desktop machine with the hope of morphos on it... and a threadripper 1950x that will make me happy emulating x5000

>micro A1

Ugh, flawed Northbridge.
A Pegasos II maybe, any early AmigaOne, no thanks.

> micro A1 the true real amigaone machine

...that wasn't allowed to be officially called an "AmigaOne" ;-)