Update:

> the annual Freescale Technology Forum. The titles of this year's
> technical sessions have already been published:
> [...]
> "Using Freescale's AltiVec SIMD Engine for Graphics, Image Processing,
> Radar, Video Surveillance and Other Applications"
>
> They didn't have sessions on AltiVec for some years. I guess this is going
> to be a foretaste of AltiVec-enabled QorIQ T.

Detailed descriptions of the sessions are online:

http://www.getregisterednow.com/FSL/CEX/Session.aspx?li=1

From there:

"During this session, attendees will gain insight into the next-generation QorIQ products that provide heterogeneous cores, AltiVec-enablement and assured computing with significant overall system performance within embedded power budgets." (FTF-NET-F0273)

"AltiVec technology, the SIMD accelerator on current e600 core microprocessors, has found a home on the QorIQ processor roadmap and has proven useful in graphics, image processing, printing, radar, sonar, video surveillance and other markets. This presentation will look at some of the performance enhancements achieved with original AltiVec technology and the improvements incorporated in the updated version." (FTF-NET-F0531)

I'm really eager to see what those "improvements incorporated in the updated version" of AltiVec are about. That goes in line with what Glenn Beck, marketing manager for aerospace, defense and single board computing within Freescale's Networking Processor Division, said when he divulged six months ago that with AltiVec they "have taken a proven technology, enhanced it, and moved it to the multicore QorIQ family of processors". I hope it will be better than IBM's VMX128.

Edit: PDF presentation with mention of "Enhanced AltiVec" (page 8):

http://www.advantech.eu/adf/file/munich/agenda/Advantech_Design-In_Forum_QorIQ.pdf

[ Edited by Andreas_Wolf 14.04.2011 - 14:44 ]

Man you're willing to dig to find info.

Three days and 1 one hour session.
FTF-NET-F0531 Thursday 09:00 1 hour Salon F

I dug through a lot of those sessions and the one you quoted is one of the few I'd have any interest in.

Why don't we just contact Bill Mercer after the event and see if he can offer a summary?

> Three days

It's four days, Monday through Thursday.

> the one you quoted is one of the few I'd have any interest in.

Let me guess the others: FTF-NET-F0394 and FTF-NET-F0399. Correct? :-)

> Why don't we just contact Bill Mercer after the event and see if
> he can offer a summary?

We won't have to do that as Freescale will offer all presentation slides for public download, as usual.

>We won't have to do that as Freescale will offer all presentation slides for public download, as usual.

Now you have my attention, sir!
Yeah! Hard info on the T5!

Update:

> 1.5 weeks ago A-Eon changed the X1000's official specs from 1.8 GHz to 2.0 GHz.

3.5 weeks ago they changed it to "nominal 2.0GHz (1.8GHz standard)".

http://www.a-eon.com/x1000.html

It could mean that it is supposed to be delivered underclocked at 1.8 GHz with heatsink but without fan and that the user who wants to run it at nominal 2.0 GHz speed would have to add a fan himself. Assuming this suspicion is true would leave one question though: If the usage of the extra 200 MHz of *one* core should require the addition of a fan what's with the usage of *both* cores at 1.8 GHz, e.g. under Linux?

Quote:

actually the right translation is "PA Semi only took an architectural licence from IBM"

some twisted french syntax

>> Google translation:
>> "PA Semi did not take an architectural license from IBM"

> actually the right translation is "PA Semi only took an architectural licence from IBM"

Thanks for clarification. I put way too much faith into Google it seems ;-)

Addendum:

> I guess there're still some information to be revealed to get the whole "quite a story".

A-Eon didn't reveal it so far (no idea why), but as there're still discussions on this and questions occuring at various places I'm going to tell how I perceive this "quite a story" came along in chronological order:

1. A-Eon decides for the PA6T, approaches P.A.Semi for recommendation of a board design company with PA6T experience and subsequently get turned to Varisys.
2. A-Eon plans on buying PA6T chips from P.A.Semi, but before it's able to place an order P.A.Semi is purchased by Apple (April 2008).
3. Apple decides to supply the PA6T to P.A.Semi's existing customers only, which A-Eon is not but Varisys is.
4. Apple calls P.A.Semi's existing customers for final PA6T orders as the chip is going to be discontinued. Varisys informs A-Eon about this call.
5. A-Eon asks Varisys to place an order for an (unknown to us) amount of PA6T chips. Varisys agrees.
6. Varisys orders the chips from Apple. The chips arrive at Varisys.
7. A-Eon in turn purchases those chips from Varisys, effectively rendering Varisys A-Eon's CPU supplier (the chips remain at Varisys, though, for obvious reasons). As part of this supply agreement between Varisys and A-Eon, Varisys demands an NDA* to be signed.

* The reason for Varisys' NDA requirement is even more speculative than my timeline. Maybe the supply agreement between Apple and its PA6T customers contained the condition that the customer must not resell the chips to 3rd parties, which is what Varisys effectively did though, so they feared penalty from Apple. It's odd at least that Varisys was officially revealed as A-Eon's design partner in June 2010 while the exact type of CPU was only officially revealed half a year later (which must be when the NDA ended).

Addendum:

> > would it make more sense to go for a P5010 based system or
> > one based on the P5020?

> That would depend on the price difference between the two.

No prices on the P5 yet, but seeing the price difference the number of cores makes with the P4 going for the P5010 could be a sensible decision:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=6196&start=160

With the P4 halving the amount of CPU cores cuts off chip price by 25%. Don't know if Freescale will apply this policy also to P5 though.

>A-Eon didn't reveal it so far (no idea why), but as there're still discussions on this and questions occuring at various places I'm going to tell how I perceive this "quite a story" came along in chronological order:

Everyone should keep in mind that Andreas is probably right about this sequence of events (in fact, his explanation is the only on that fits the facts), but this is still speculation.

Obviously, for the reasons stated in Andreas' summary, Trevor and the rest of A-eon haven't been able to reveal some of this information.

> Obviously, for the reasons stated in Andreas' summary,
> Trevor and the rest of A-eon haven't been able to reveal
> some of this information.

Really? Could you name those reasons? I ask because as I stated above I've "no idea why" some of those information should still be confidential when the NDA between Varisys and A-Eon had ended in (or before) December 2010 obviously.

I can think of three reasons for the info not to be released.

1. The NDA is still in effect.

2. The information to be released could be embarrassing to the person/company responsible for releasing, or not releasing it.

3. The person/company responsible for making the decision to release, or not release the information does not think it is important one way or the other and feels there is no reason to respond to any requests for such information.

> 1. The NDA is still in effect.

Unlikely, else A-Eon wouldn't have revealed the type of CPU in December 2010. The information on the CPU was part of the NDA according to A-Eon.

> 2. The information to be released could be embarrassing to the
> person/company responsible for releasing, or not releasing it.

Could be.

> 3. The person/company responsible for making the decision to release, or
> not release the information does not think it is important one way or the other
> and feels there is no reason to respond to any requests for such information.

Unlikely, else A-Eon wouldn't have teased us with its "quite a story" comment.

http://www.google.com/search?q=%22a-eon%22+%22quite+a+story%22

>> We won't have to do that as Freescale will offer all presentation slides
>> for public download, as usual.

> Now you have my attention, sir! Yeah! Hard info on the T5!

I seems that this year Freescale for whatever reason requires to login in order to be able to download the FTF presentation slides. Pity, I'd really like to know what the improvements of the updated AltiVec are.

http://www.freescale.com/webapp/sps/site/overview.jsp?code=FTF_2011_AMERICA_OLL_CAT

@Andreas

The registration process is simple enough, and open to the public.

> The registration process is simple enough, and open to the public.

I know that. But there's no need to as Jim is kind enough to relay the slides I'm interested in. I'm just wondering why Freescale changed the policy after years of making the FTF slides available without having to login. Besides, you just can't easily link to files that are not publically available without registering ;-)

Umm its called data mining.

But you wont be able to register anyway as you are a bot.

> you are a bot.

Seems the other half of your split personality disagrees with you:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=5724&start=40

Yes, having to log in to view the PDFs is a pain since you can't link to them when creating posts. I'm having trouble locating one of the files Andreas asked about, but I've sent him the others.
And I've put in a request for the file I couldn't locate.
they should respond by tomorrow.
I've been really busy at work, so I haven't had a chance to review them.
Can't wait to see what Andreas picks out that's note worthy.
These files along with the announcement of the new processor really have me stoked.

> I'm having trouble locating one of the files Andreas asked about, but I've sent
> him the others. And I've put in a request for the file I couldn't locate. they
> should respond by tomorrow.

Thanks for your efforts. It's really appreciated.

> Can't wait to see what Andreas picks out that's note worthy.

;-) I've yet to have a look at them but will do so ASAP.*

> These files along with the announcement of the new
> processor really have me stoked.

They have despite the fact that Freescale underdelivered compared to what you had hoped for? ;-)


* Addendum:
https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7001&start=665

[ Edited by Andreas_Wolf 24.06.2011 - 05:41 ]

Update:

> I'd really like to know what the improvements of the
> updated AltiVec are.

Page 12 of FTF11_NET_F1176 document mentions "Improved load /store to ease bit alignment" and "New instructions for video analytics".

Edit: Now also there:
http://www.freescale.com/files/32bit/doc/brochure/PWRARBYNDBITSSKD.pdf (page 3)
http://www.freescale.com/files/32bit/doc/brochure/PWRARBYNDBITS.pdf (page 59)

[ Edited by Andreas_Wolf 27.03.2012 - 19:39 ]

>> These files along with the announcement of the new
>> processor really have me stoked.

>They have despite the fact that Freescale underdelivered compared to what you had hoped for?


Oh YES! Between this and the new developments at IBM (and Nintendo) I really excited to see PPC processor development advancing. For all those naysayers that said this was a dead ISA, things sure seem to be moving along nicely.

APM announced a new ppc NAS processor, too.
The APM86491 - Key features: 1.2 GHz 465 core, 256 kb L2 cache, 2x usb 3.0, 2x SATA, 2x Gb Ethernet, 2 x PCI Express, 2.5 W energy uptake, 20 US$ in bulk volume.

http://phx.corporate-ir.net/preview/phoenix.zhtml?c=78121&p=irol-newsArticle&ID=1576895&highlight=

While designed for NAS boxes, this may also be used for a small low cost low end, low cost computer if you attach a video and audio codec to it.

[ Editiert durch Zylesea 25.06.2011 - 01:51 ]

> APM announced a new ppc NAS processor, too. The APM86491

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7001&start=671

;-)

Care to elaborate what you wanna tell with your link? I see the Sam460 specs only...