Update:

Quote:

APM86791 announced ("Keelback"):

http://investor.apm.com/phoenix.zhtml?c=78121&p=irol-newsArticle&ID=1570965
http://www.apm.com/products/embedded/multicore460/apm86791/

It seems that Gopi already included this when talking about their FY2011 SoCs one month ago. But it's apparent that Applied Micro is not really following its own public roadmap as according to this the chip after Diamondback should have been faster than Mamba, yet Keelback is just as slow as Diamondback.

Not bad, but slow.
If APM could get their SoC's to operate at higher frequencies they might have something.

Update:

> As of November 2010 there's a fresh roadmap on QorIQ T series:
> http://www.freescale.com/files/training/doc/dwf/AMF_NET_T0425.pdf (page 7)
> T5: T5040, T5020
> T4: T4160, T4120
> T3: T3080, T3060
> T2: T2040

Seeing as the QorIQ T4240 just got announced with 12 dual-threaded cores we now know the meaning of the nomenclature of the QorIQ AMP series members. Accordingly, the T5040 should have two dual-threaded cores and the T5020 should have only one single dual-threaded core.

Edit: Correction:

According to page 6 of the FTF11_NET_F1176 document the QorIQ T5 will have "1-8 Single threaded Cores". This should mean that the T5040 won't have two dual-threaded cores but four single-threaded cores and the T5020 won't have one dual-threaded core but two single-threaded cores. Furthermore, this should mean that there'll also be a T5010 with one single-threaded core.
(Besides, that page says up to 8 cores for the T5 while the press release says only up to 6 cores. And the page also says "1-8 cores" for T1/T2, which would be 2 to 16 threads, while the press release says only up to 8 threads, which would be 4 cores.)

[ Edited by Andreas_Wolf 27.03.2012 - 19:44 ]

Update:

>> 3+ GHz
>> [...] this gives us some idea of what they're shooting for with the
>> next revision of the e5500 core. [...] 3Ghz or higher operation.

> Keep in mind that this clock rate is what more than two years ago they
> thought they'd be able to deliver the T5 at. This doesn't mean it really
> will come at that, not even that it's still planned to come at that.

e6500 core announced to run at up to 2.5 GHz:

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=7901&forum=11

This is the same maximum clock speed that was announced for its predecessor, the e5500 core.

Update:

> Just for the purpose of overview I compiled a list of the DMIPS/MHz
> per core figures for various PPCs in ascending order (AltiVec capability
> is generally ignored, obviously):
> [...]
> e5500: 3.0

The table on page #2 of the QorIQ AMP fact sheet PDF indicates that the e6500 core didn't improve in that regard (comparing e6500's per-thread performance to e5500's per-core performance). The only thing that makes me wonder is that on said page Freescale makes it sound as if the e5500 never existed:

"The 64-bit e6500 core [...] enables more than twice the performance improvement as well as twice the power efficiency over Freescale's previous generation e500mc core through enhancements to the core itself, the move to 28 nm process technology, as well as dual-threading capability."

I'd rather think it's the e5500 core they enhanced to realize the e6500 core and didn't start with enhancing the e500mc core all over again ;-)

Quote:

It's called marketing, it's not meant to make sense

Quote:

Sounds like they've taken the e5500 and added in a second thread and AltiVec.

Then they stuck a load of them together and bolted on a monster 6MB cache.

Sounds pretty much like where the G5 would have gone next.

> It's called marketing, it's not meant to make sense ;-)

;-) To make the improvements seem even better they could as well have the e6500 compared to its ancestor from a decade ago, the e500v1.
But to be not too unfair towards Freescale I will note that while comparing the e6500's performance to e500mc's instead of e5500's makes the e6500 shine even brighter it's the exact other way round regarding the power efficiency comparison as the e5500 is actually even less power efficient than the e500mc. See:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7001&start=435
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7001&start=439

So substituting the e500mc by the e5500 in Freescale's statement would read something like this:
The e6500 core enables about 60% performance improvement as well as 2.6 times the power efficiency over Freescale's previous generation e5500 core.

> Sounds like they've taken the e5500 and added in a second thread and AltiVec.

Yes, that's exactly what I think they did to come up with the e6500.

> Sounds pretty much like where the G5 would have gone next.

Yes, considering a "PPC980" would surely have featured AltiVec like the PPC970 and probably been derived from POWER5 which was the first POWER CPU to have SMT capability.

Addendum:

>> "AltiVec SIMD has long been recognized for its performance in the
>> PowerQUICC processor line"
>> http://www.freescale.com/webapp/sps/site/homepage.jsp?code=QORIQ_HOME
>>
>> AltiVec in PowerQUICC? Huh? While it's true that they present QorIQ P1 to P3
>> as PowerQUICC replacements and QorIQ P4 and P5 as MPC86xx replacements
>> that surely doesn't make MPC86xx a PowerQUICC, does it? Are they now trying
>> to retroactively subsume MPC86xx (or even MPC74xx) as PowerQUICC?

> Just discovered that Freescale indeed lists the MPC8610 as PowerQUICC III
> (while MPC864x is listed as "Host Processor") there:
> http://www.freescale.com/webapp/sps/site/overview.jsp?code=PRDCT_LONGEVITY_HM
>
> Furthermore, MPC74xx and MPC86xx are listed in Freescale's 'PowerQUICC and
> QorIQ Processor Selector Guide':
> http://www.freescale.com/files/32bit/doc/brochure/PWRARCHQIQSG.pdf
>
> Explanation that adds up, anybody?

Recent example of the confusion Freescale has been causing over this matter:

"The e6500 core integrates an enhanced version of the AltiVec single instruction, multiple data (SIMD) vector processing unit, which has long been available on Freescale's PowerQUICC processors"
http://www.linuxfordevices.com/c/a/News/Freescale-AMP-QorIQ/

It's certainly not going to make more sense when Freescale's own John Dixon, Power and DSP products marketing manager, claims regarding the e6500 core:

"We also have the reintroduction of AltiVec, which is an enhanced version of our AltiVec that we had in our PowerQUICC products in our previous generations."
http://videos.cache.magnify.net/60X0NG37F35BHBT2-FreescaleEmbargo0621_483_336_384x216.mp4 (at 0:57)

Update:

> PacketPro was originally announced to sport USB 3.0 as well as 10GbE.
> So far, Mamba only offers USB 2.0 and 1GbE.

Still no 10GbE, but eventually USB 3.0 announced in APM86491 ("Catalina"):

http://investor.apm.com/phoenix.zhtml?c=78121&p=irol-newsArticle&ID=1576895
http://www.apm.com/products/embedded/multicore460/apm86491/

Update:

>> By the time Mamba is introduced, the 5010 and 5020 will be available.

> According to Applied Micro, production quantities of Mamba are expected in
> Q2/2011. You told me some weeks ago that Freescale anticipated Q2/2011
> for P5010 production as well (which was originally set at H2/2011). So let's
> see who'll be first :-)

Q2 has passed and the winner seems to be Mamba. Avnet lists Mamba at 1.2 GHz for 77 USD single-core and 116 USD dual-core in120 pcs. quantity. No prices for the QorIQ P5 yet.

Addendum:

>> According to IBM* the PPC476FP core even delivers 2.7 DMIPS/MHz.
>> That puts it above e500mc (2.5 DMIPS/MHz).
>> * PPC476FP Core Product Brief (August 2010, page 2)

> Confirmed by other IBM documents:
> Introduction to PPC476FP White Paper (April 2011, pages 3 and 7)
> PPC46x and PPC47x Feature Differences Application Note (May 2011, page 4)
> PPC470S and PPC476FP Technical Overview (February 2011, page 14)

Seconded by LSI for the PPC476FP core based Axxia processor:

"Provides Dhrystone performance of more than 2.7 DMIPS/MHZ"
http://power.org/wp-content/uploads/2011/09/LSI-Powerdotorg_webinar_final.pdf (page 8)

"Dhrystone performance > 2.7 DMIPS/MHZ"
http://power.org/wp-content/uploads/2012/09/10._LSI_Cooky_Shi_upd.pdf (page 14)


Edit: PDF links updated

[ Edited by Andreas_Wolf 18.09.2012 - 23:10 ]

Update:

> So if A-Eon are pleased with Varisys's performance regarding
> development of the Nemo board I don't see a reason why they
> wouldn't want to expand their business relation to an X1000's
> successor (if there really will be any, that is), based on QorIQ.

Sounds like that's exactly the plan:

"Throughout the Nemo development phase Varisys has proved an extremely knowledgeable and highly responsive partner. A-EON Technology director Trevor Dickinson said, “We have been very satisfied with Varisys commitment, dedication and quality. When it comes to leading edge PowerPC hardware their level of technical expertise is possibly unrivalled. As a result Varisys has contracted to work on several new development projects to improve and extend the AmigaOne product line. Adam Barnes, Varisys Technical Director said, There is some exciting cutting edge PowerPC technology on the horizon and we are delighted to have the opportunity of bringing this to the AmigaOne platform. Varisys Managing Director Paul Gentle added, “We enjoyed working on the Nemo development which provided our first experience of the global Amiga community. We look forward to extending our relationship with the new projects and will do our best to 'Keep this party going'."
http://amigaprez.wordpress.com/2011/10/20/amigaone-x1000-to-be-at-amiwest-2011/
http://www.amigakit.com/x1000/newsrelease.pdf (page 2)
http://www.a-eon.com/news.html

Admittedly, QorIQ is not mentioned explicitly, but what else should "exciting cutting edge PowerPC technology on the horizon" mean? :-)

> When the Sam440ep was released in October 2007 it still took Hyperion
> one year for a first OS4.1 version to be released for it:
> http://www.acube-systems.biz/index.php?page=news&id=15
> http://www.acube-systems.biz/index.php?page=news&id=38

Hyperion now claims this:

"22 October 2007
ACube's Sam440ep released for end users with a beta version of AmigaOS 4.0"
http://www.amigaos.net/content/10/history-amigaos

Does anybody remember this? I clearly don't. AFAIR there was never ever OS4.0 for any Sam board available (at least not for end users). What am I missing here?

Moreover, it's claimed under the same link:

"AmigaOS 4.0's inception dates back to 2002 when an agreement was reached to port AmigaOS to the PowerPC platform."

Didn't this agreement between Hyperion, Amiga Inc. and Eyetech being reached happen already in 2001 when Hyperion (re-)started the OS4 project, not in 2002?

Unfortunatly Andreas, Hyperion don't share your fondness for complete accuracy. Especially when it suits them otherwise.

> Hyperion don't share your fondness for complete accuracy.

Now I'm only waiting for someone to reflect this new "truth" in Wikipedia and referencing that "history" page as source ;-)

> Especially when it suits them otherwise.

"2002" instead of "2001" could be a typo (which of course should be corrected nonetheless). But yes, Hyperion's nonsense claim that the Sam440ep was released for end users with OS4.0 available (when in fact it took them one year to release a first OS4 version for it) is really dubious and does look like it was put there on purpose.

> Hardware development is mostly done by Pegatron -with the cooperation of bPlan

I just noticed that back in August bplan put up a web page dedicated to their latest hardware developments done for Genesi:

http://www.bplan-gmbh.de/genesi_en.html

Andreas_Wolf,
Quote:

Well duh!!

The OS4 related Wikipedia pages have been a work of fiction for a long time. If someone tries to correct it, hordes of BHF go on the attack.

Update:

> As of November 2010 there's a fresh roadmap on QorIQ T series:
> http://www.freescale.com/files/training/doc/dwf/AMF_NET_T0425.pdf (page 7)
> T5: T5040, T5020
> T4: T4160, T4120
> T3: T3080, T3060
> T2: T2040

Revised QorIQ T roadmap as of August 2011 (page 4):

http://www.freescale.com.cn/cstory/ftf/2011/pdf/0772.pdf
http://2011ftf.ccidnet.com/pdf/0772.pdf

T5: T5080, T5040
T4: T4320, T4160
T3: T3080, T3060
T2: T2080, T2040, T202x

The most interesting changes to my mind are the removal of the T5020 and the addition of the T5080. What's puzzling is that the T4240 (in June 2011 announced for early 2012) is missing from this roadmap. Or maybe they increased its core count from 12 to 16 and accordingly call it the T4320 now?

At the AmiWest 2011 Show I asked Trevor Dickinson what was next for him. I thought that after all the criticism he received about the X1000 computer he would be discouraged and not wish to spend any more money trying to satisfy the Amiga community with more hardware projects. To my surprise, he instead announced that he is moving forward with more new computers from A-Eon with Varisys as his partner and AmigaKit was also joining him as a partner to produce new AmigaOne computers. He also expressed that although his personal preference is OS4, he hopes to bring the Amiga community back together and one way to help accomplish this is to provide hardware that all Amiga inspired OSes can run on.

Of course he cannot accomplish this without cooperation from the AROS and MorphOS developers, as well as his current cooperation with Hyperion to have AmigaOS4.x ported to any new computers. It will be interesting to see what his next system(s) is/are going to use for a CPU, as OS4 and MorphOS are clearly not moving away from PPC in the near future and AROS is already available for PPC, though I have not read up on which PPC platforms AROS has been ported to, or if work continues on the PPC port, or not. The QorIQ CPU's are the most likely choice, I would guess.

Trevor also said that he will not be talking about what he is doing, or releasing any specs on his next project(s), until they are ready to release. It was an obvious statement on his displeasure on the highly critical nature of many users in the Amiga community that love to tear apart and criticize plans and timeframes that are released early, or who moan and complain endlessly about the direction some people are taking to advance the Amiga experience, instead of doing something themselves. We have no shortage of people that love to sit on the sidelines and complain, instead of actually contributing anything useful to the Amiga community as a whole. So, with this choice to hold back information about new projects, we will have to wait until a product is close to being ready for release, before we will know what CPU any new computer from A-Eon will be using.

I for one, will be interested in seeing what Trevor will come up with next.

2012 should be another interesting year for new Amiga hardware announcements. Now if we could just make some progress on bringing back more developers and speed up the writing of more useful software and drivers for the hardware we already have, or will soon have, the future of Amiga inspired systems would be better than it has been in many years.

[ Edited by amigadave 09.12.2011 - 11:41 ]

> he instead announced that he is moving forward with more new computers
> from A-Eon with Varisys as his partner and AmigaKit was also joining him
> as a partner to produce new AmigaOne computers.

True, he specifically said that they contracted Varisys to develop three new products, one of which will be an XCore processor card. I guess the remaining two will be standalone PPC boards for running OS4. One of them could be a successor to Nemo, and the other maybe a mobile solution (i.e. quasi-successor to Hyperion's OS4 netbook), but that's just wild speculation from my part.

> he hopes to bring the Amiga community back together and one way to help
> accomplish this is to provide hardware that all Amiga inspired OSes can run on.
> Of course he cannot accomplish this without cooperation from the AROS and
> MorphOS developers

It's not that MorphOS can't be ported to existing Sam4x0 boards or to the X1000 but just that the MorphOS Team doesn't want to do this for several comprehensible reasons. So it will be interesting to see how he's going to create appeal for the MorphOS Team to port MorphOS to any future A-Eon hardware.

> It will be interesting to see what his next system(s) is/are going to use for
> a CPU [...] The QorIQ CPU's are the most likely choice, I would guess.

Absolutely. A-Eon already hinted at Freescale's QorIQ P4 and P5 chips for future hardware. In my opinion it would be better for them to wait for the availability of the AltiVec-enabled QorIQ AMP (= T series) chips, at least for the X1000 successor, as even the top end P series chips are only a little better performing than the PA6T, if at all, and they lack AltiVec.

> I have not read up on which PPC platforms AROS has been ported to

So far, AROS for PPC runs natively only on Efika 5200B and Sam440.

> Trevor also said that he will not be talking about what he is doing, or
> releasing any specs on his next project(s), until they are ready to release.
> It was an obvious statement on his displeasure on the highly critical
> nature of many users in the Amiga community that love to tear apart and
> criticize plans and timeframes that are released early

I've understood it more like a statement on his displeasure with A-Eon's own rush job at announcing the X1000 for half a year later, two whole years in advance of its release.

> 2012 should be another interesting year for new Amiga hardware announcements.

So given that Trevor Dickinson said he would "not be talking about what he is doing, or releasing any specs on his next project(s), until they are ready to release" you believe there should be new A-Eon hardware released as early as 2012?

Andreas_Wolf,
Quote:

It is more likely that the 4240 is canceled for 2012 (delayed to 2013, 4320).
IMO it's a bad sign.
I am corious if and when the 4160 will see the light of day. The TDP will be very interesting, also!

>> What's puzzling is that the T4240 (in June 2011 announced for early 2012)
>> is missing from this roadmap. Or maybe they increased its core count from
>> 12 to 16 and accordingly call it the T4320 now?

> It is more likely that the 4240 is canceled for 2012 (delayed to 2013, 4320).

That's in essence what my speculation you replied to was. And you're right that this would mean a delay from early 2012 to early 2013 according to the roadmap depiction. In June Freescale said they would "provide more details regarding the T4240 product later this year", which still has 3 weeks left. So I guess when there won't be any details on the 12-core T4240 this year we can conclude that it's probably cancelled and "replaced" by the 16-core T4320 which is supposed to come one year later.

> I am corious if and when the 4160 will see the light of day.

According to the roadmap depiction, the 8-core T4160 should be available as initial sample when the T4240 was originally said to be available by Freescale, i.e. early 2012. We'll have to wait and see I guess.

> The TDP will be very interesting, also!

http://www.freescale.com.cn/cstory/ftf/2011/pdf/1417.pdf
http://2011ftf.ccidnet.com/pdf/1417.pdf

This document has on page 16 mentioned "<50W thermal max at 1.8GHz" and "<40W thermal max at 1.67GHz" as power targets for the 16-core T4320. From these figures we may infer the targeted TDP for the T4160 with half the amount of cores.

A little bit more than 20W for an 8 core CPU sounds sweet.
It should be easy to double that figures from - King of the hill - IBM

David's conversion with Treavor brings up topics that I've discussed with him via e-mail myself.
He has stated that he would like to develop systems that support OS4 AND MorphOS.
That would be interesting considering his partner's involvement with Hyperion.

> He has stated that he would like to develop systems that support OS4 AND MorphOS.

First off, it's the OS that must support the hardware, not the other way round. As said before, the MorphOS Team is free to port MorphOS to hardware that is already supported by OS4 (Sam4x0, X1000). I remember that in March you and me discussed the possibility that Trevor would attempt to have new systems created outside of A-Eon (i.e. without Ben Hermans' involvement) in order to reduce the MorphOS Team's reluctance. So far it seems he won't do that. Another show stopper is the compulsive bundling of OS4 to the hardware. For the MorphOS Team to consider a port the hardware would have to be available without OS4 optionally.

> That would be interesting considering his partner's involvement with Hyperion.

As we know, Trevor says A-Eon would very much appreciate a port of MorphOS to the X1000. So it seems Ben Hermans wouldn't object to that either.