IMG confirm that Intel is using SGX

[tangey]

The Poulsbo chipset , now known post-IDF as the System Controller Hub (SCH), is confirmed as having SGX in it, as well as video related IMG IP.

http://www.imgtec.com/News/Release/index.asp?NewsID=712


Technical documents at intel show that Intel are releasing 3 SCH chips, UL11L has graphics clocked at 100Mhz and is limited to 800x480 internal screen only, UL15L and UL15W are clocking the graphics at 200Mhz and support internal & external screens and full HD video.


AS UMPC and MID are handheld devices, I've put this post in here, please move if inappropriate.

[darkblu]

that's a very good news. thanks for the heads-up.

[tangey]

The following video pretty much confirms that that SGX will also be in Canmore Soc.


Heres a video of a presentation on Socs at the recent Intel IDF. Forward to around 17:30 and you'll hear the Intel guy (in relation to canmore) "...contains 3rd Party IP blocks providing audio/video encoding and graphics and display engines".

http://www.intel.com/pressroom/kits/events/idfspr_2008/video.htm#
Anyone know of a 3rd Party Graphics IP supplier to Intel other than IMG ?


Given that Intel are heavily promoting the idea of re-usable IP in Socs as a way to quickly produce variations of chips for specific markets, this would also mean that it is likely that IMG IP will be used in the Moorestown platform, and even gives an outside possibilty that it'll make an appearance in the graphics enabled variant of Nehelam that will be coming out next year.

[Scali]

I suppose this is the item to which the original post was linking aswell:
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=370

"Imagination Technologies confirms new Intel® Centrino® Atom processor technology utilises POWERVR graphics and video technologies"

[tangey]

Intel at this weeks IDF confirmed that their IA Soc, the CE3100 (formerly known as Canmore) contains the same SGX graphics core as found in Poulsbo. Its actually clocked a little higher @ 216Mhz.
I'm not sure at this time if the CE3100 also contains IMG's video IP, VXD370.

I know that the canmore chip isn't for handheld, butI didnt want to open up another IMG thread.
http://download.intel.com/design/celect/downloads/ce3100-datasheet.pdf

look at page 1.2 for the SGX mention.

[tangey]

Intel has started to talk about the graphics and video tech in the moorestown platform.

http://macnewsupdate.com/intel-mid-market-was-established-by-the-iphone-continues-to-grow/

Highlights are:-
1)The GPU is going to be about the same size as the CPU core
2)code and decode high-definition TV (HDTV) imagery.
3)The volume of data being handled has increased significantly, and we had to expand the memory bandwidth
4)it will deliver improved graphics draw and image processing performance

Interesting that one of the mentioned requirements is increased data bandwidth, in that the latest SGX531 core (remembering that SGX530 is in the menlow platform) announced in Oct was specifically aimed at:-

"emerging trends in higher performance SoCs (Systems on Chip) that incorporate 128-bit or other higher data throughput bus fabrics."

So it looks very likely to me that SGX531 is for moorestown

[Mike11]

Intel has started to talk about the graphics and video tech in the moorestown platform.

http://macnewsupdate.com/intel-mid-market-was-established-by-the-iphone-continues-to-grow/

Highlights are:-
1)The GPU is going to be about the same size as the CPU core
2)code and decode high-definition TV (HDTV) imagery.
3)The volume of data being handled has increased significantly, and we had to expand the memory bandwidth
4)it will deliver improved graphics draw and image processing performance

Interesting that one of the mentioned requirements is increased data bandwidth, in that the latest SGX531 core (remembering that SGX530 is in the menlow platform) announced in Oct was specifically aimed at:-

"emerging trends in higher performance SoCs (Systems on Chip) that incorporate 128-bit or other higher data throughput bus fabrics."

So it looks very likely to me that SGX531 is for moorestown
Well, but just for the MID/UMPC Moorestown Atom!? The Netbook and Nettop variant will get something else (Intel's own IP?)?

Interesting interview. So, MIDs will dominate the world and the iPhone is an MID according to Intel. In that case IMHO he might be right with the domination part but I really don't see the need for Intel (and Windows) in that 4" MID world. 8 inch or bigger ok. But in the all-in-one, the one device that I always-carry-around-with-me-world? No.

[tangey]

Well, but just for the MID/UMPC Moorestown Atom!? The Netbook and Nettop variant will get something else (Intel's own IP?)?

The article was addressing the MID market,and Intel's chipset for that market. The Intel guy does say that he doesn't think its necessary to distinguish between MIDs and netbooks. But does that mean the subsequent mention of Moorestown implies it could be seen in netbooks ?...dunno.

[Lazy8s]

Intel's SCH used the SGX535 specifically, so I don't think they'd use what looks like a variant of the embedded-focused SGX530, the 531 variant, for Lincroft.

[roninja]

Seems as though the Z520 is used in the forthcoming Sony Vaio P11.

http://www.umpcportal.com/2009/01/new-sony-umpc-info-includes-specs-and-a-sub-1000-price

So it seems the Menlow platform is creeping even further into the Netbook space alongside its core MID domain...

[Ailuros]

Seems as though the Z520 is used in the forthcoming Sony Vaio P11.

http://www.umpcportal.com/2009/01/new-sony-umpc-info-includes-specs-and-a-sub-1000-price

So it seems the Menlow platform is creeping even further into the Netbook space alongside its core MID domain...

See also: http://www.fudzilla.com/index.php?option=com_content&task=view&id=11177&Itemid=1

[Mike11]

See also: http://www.fudzilla.com/index.php?option=com_content&task=view&id=11177&Itemid=1
Why would an iPod Touch HD use an Intel Atom? Higher clocked ARM11 CPU or even dual-core ARM11 makes more sense (for fall 09). Things would look different for a MacTablet, but at 7" a "real" Mac doesn't make much sense imho.

[tangey]

Speculation is mounting that Intel's Pineview Soc (part of the Pinetrail platform for notebooks), will integrate a GMA500 style (i.e. SGX) graphics, and not GMA950 graphics (Which was previously using for netbooks along with the N atom series)

http://www.netbookchoice.com/2009/06/10/first-details-on-intel-%E2%80%98pineview%E2%80%99-atom-n400-d400-d500-processors/

That info is not official Intel stuff, but a summation of info received by hkepc.com

The official line from intel can be seen in this .pdf:-
http://www.intel.com/pressroom/archive/reference/Pineview_Moblin_disclosure.pdf

All that says on page 6 is that there will be "increased graphics performance". Yet it is common knowledge that GMA500 on the Zseries Atoms has consistently performed worse than the 950 graphics using in the G945 chipset (for the N series Atoms)

Now the info from the first link above,and also various other sources, are stating that the graphics will run at 200MHz. Thats the same clock speed at the GMA500 currently runs on most versions of Intels SCH chip. So if thats the case, where is the performance increase coming from ? Other than the improvement in having the processor and the graphics on the same die, there isn't one ? Is Intel merely stating that in thier opinion GMA500 is better than GMA950, and that the actual poor performance to date of GMA500 is down to driver issues ?

Another site today stirred the story more.....
http://www.tmcnet.com/usubmit/2009/06/14/4224804.htm
THis site states "the graphics chip Intel is said to be planning to pair with the upcoming Atom CPUs will only be slightly faster than its existing one".

Now I'm not sure what they are comparing too, but I'm assuming its the existing GMA950 being using in existing netbook platform. It goes on to say

"The upcoming Pineview Atom N450 will reportedly be paired with Intel's GMA 500 graphics chip. The GMA 500 will run at a faster clock speed and double the number of graphics "pipelines" to four. But it will still run at half the speed of Nvidia's 9400M chip and boast only a quarter of its graphic pipelines.", and quotes Jon Peddie as saying,The GMA 500 is "indeed better than the old 945," .

Again, I feel that the "faster clock" is relative to the 133Mhz that the existing netbook chipset graphics runs out. Can anyone confirm the number of "pipes" that SGX535 has ?

I wonder if we might see SGX540 in pineview, at 200MHZ it should provide a good performance increase over SGX545, and given that SGX540 taped out in Oct '08, the timing would be about right ?

To confuse the situation even more, the last link says that "For consumers demanding HD video, Intel reportedly plans to offer Broadcom Corp.'s Crystal HD video decoder chip as an optional add-on."

And yet we know that one of the main strengths of poulsbos GMA500 over the existing N-series graphics is its video decode ability, GMA500 can decode 1080p in hardware. So if its GMA500 that is being put into pineview, why the need for an optional separate HD decode chip ? Here's another separate article which would seem to confirm the optional separate HD decode chip:-

http://apcmag.com/intels-next-gen-atom-platform-gets-hd-decoder-chip.htm

Its all very confusing

[Lazy8s]

Two "pipelines" for SGX53x; four for SGX54x.

[iwod]

If it is really SGX, then even if it perform worst then GMA950 i will still be happy. Intel Gfx is the worst i have seen possibly in the whole Graphics History.

I am still puzzled by why Intel need YET ANOTHER chip for decoding when the newest Atom was all about Integrating MORE Chips into its Core.

The have PowerVR video decoder and they are not using it.

[Ailuros]

Again, I feel that the "faster clock" is relative to the 133Mhz that the existing netbook chipset graphics runs out. Can anyone confirm the number of "pipes" that SGX535 has ?




SGX520: 1 Pipe, 1 TMU
SGX530: 2 Pipes, 1 TMU
SGX531: 2 Pipes, 1 TMU, 128 Bit Bus
SGX535: 2 Pipes, 2 TMU, DX9
SGX540: 4 Pipes, 2 TMU
SGX543: 4 Pipes, 2 TMU, USSE2 40% höhere Leistung zu 540, Multi-Core fähig
SGX545: 4 Pipes, 2 TMU, DX10.1

http://www.mitrax.de/?cont=artikel&aid=35&page=4

If someone could find out what the die area for the graphics core would be it would help defining which SGX variant it might be. If it's in the 8mm2@45nm league then the scale weighs more in 545's direction.

If Intel truly integrates SGX also in that one, here's to hope that the finally deliver decent drivers this time. That alone would deliver a huge performance increase even for Poulsbo heh...

[tangey]

Intel just bought a further 25M shares in IMG, bringing their stake from under 4% to just about 14%.

Intel Corporation - No intention to make an offer for Imagination Technologies
Group plc

"Intel Corporation, through its wholly owned subsidiary Intel Capital
Corporation, has acquired a further 25,000,000 shares in Imagination
Technologies Group plc resulting in an aggregate holding of 31,970,422 shares in
Imagination Technologies Group plc (including its existing holdings of 6,970,422
shares held by its wholly owned subsidiaries Intel Capital (Cayman) Corporation
and Intel Capital Corporation) representing 13.99% of Imagination Technologies
Group plc's total issued voting rights.

Intel Corporation has already licensed several generations of Imagination
Technologies Group plc's graphics and video IP cores for deployment in Intel's
PC, mobile computing and consumer architectures. Intel Corporation values its
continued relationship with Imagination Technologies Group plc and the use of
Imagination's technologies by an industry-wide customer base.

Intel Corporation has acquired these additional shares for strategic reasons.
Whilst Intel Corporation will continue to consider its options in relation to
its holding in Imagination Technologies Group plc, for the purposes of Rule 2.8
of the City Code on Takeovers and Mergers (the "City Code"), it confirms that it
has no current intention to make an offer for Imagination Technologies Group
plc."



This is a unique situation, IMGs biggest shareholder Saad are "distressed" and need to offload. They had been holding just under 42M shares.

I fully expect Apple to now take some of the remaining Saad shares.

[Ailuros]

It could very well be that Intel just gives (like themselves in the past and Apple with their own share investment) additional resources to IMG to be able to keep up with increasing R&D resources. It could thus mean that Intel plans a wider than current employment of IMG's IP in their SoCs.

[tangey]

It could very well be that Intel just gives (like themselves in the past and Apple with their own share investment) additional resources to IMG to be able to keep up with increasing R&D resources. It could thus mean that Intel plans a wider than current employment of IMG's IP in their SoCs.

Not quite sure what you mean, the shares were not bought from IMG, or issued by IMG as part of an agreement, they were bought from an Saudi investment house thats in big trouble and needs to offload its entire portfolio. Intel got the shares at 80p, which is 10-15% discount to the market price at the time. IMG got nothing.

[iwod]

Not quite sure what you mean, the shares were not bought from IMG, or issued by IMG as part of an agreement, they were bought from an Saudi investment house thats in big trouble and needs to offload its entire portfolio. Intel got the shares at 80p, which is 10-15% discount to the market price at the time. IMG got nothing.

They are only 350M + right now, Intel could buy them in a snap.
I am not sure why Apple didn't acquire those stocks though. It seems to be Apple is very conservative with its pile of cash.

[Ailuros]

http://www.fudzilla.com/content/view/14347/1/

There's no NOKIA thread mind you guys ;)

[tangey]

http://www.fudzilla.com/content/view/14347/1/



Saw that news earlier on bloomberg, most likely for a mid/tablet, on a "suck-it-and-see" basis. Be awfully brave to commit to phones at this stage. But I'm guessing medfield is looking like it'll be competitive in the phone segment. Given who is hosting the conference call (Intel's Anand Chradraskher), it will almost certainly imply IMG involvment.

[tangey]

Intel and Nokia Announce Strategic Relationship
Full press release here:-
http://www.intel.com/pressroom/archive/releases/20090623corp_b.htm?iid=pr1_releasepri_20090623rb

Vague and aspirational, but if any end product comes from it, it'll probably be good for IMG. The losers are likely to be ARM and possibly Symbian.

[Ailuros]

Intel and Nokia Announce Strategic Relationship
Full press release here:-
http://www.intel.com/pressroom/archive/releases/20090623corp_b.htm?iid=pr1_releasepri_20090623rb

Vague and aspirational, but if any end product comes from it, it'll probably be good for IMG. The losers are likely to be ARM and possibly Symbian.

NOKIA has always been trying to not depend on only one supplier. If NOKIA is truly looking into entering markets above smartphones it could very well be that Intel is a lot closer in having SoCs ready with SGX54x integrated than TI's OMAP4.

[tangey]

By-the-By, Intel just bought ANOTHER 4.85M shares in IMG, they now own just over 16% of IMG.

[Mike11]

By-the-By, Intel just bought ANOTHER 4.85M shares in IMG, they now own just over 16% of IMG.
Why the hell is Intel doing this? Right now they really don't take advantage of IMG hardware and for the future they are working on their own GPUs (plus SIMD Accelerator etc.).

[argor]

the IMG Financial Highlights are out http://www.imgtec.com/corporate/newsdetail.asp?NewsID=472

Pre-tax profit* jumps 56% to £5m and chip unit volume up 83% to 86m

* 12 major agreements involving over 25 silicon IP core licences
o Strategically and financially important deals with Apple, Intel, MediaTek, NXP, Samsung, Sigma Designs, SiRF, and Toumaz as well as a significant agreement with a second major international OEM

[Laurent06]

NOKIA has always been trying to not depend on only one supplier. If NOKIA is truly looking into entering markets above smartphones it could very well be that Intel is a lot closer in having SoCs ready with SGX54x integrated than TI's OMAP4.
Indeed, Nokia always try to get different suppliers. The problem with going with Intel is that only Intel controls Atom core. With ARM at least they can put ST vs TI vs Samsung vs etc.

[Arun]

The number of suppliers, platforms, and OSes at Nokia is really spiraling out of control. Even with their massive scale, I don't buy that they can justify this insane lack of focus. Sorry for being such a party pooper, but it reminds me of one thing and one thing only: Motorola, exactly three years ago. It is noteworthy that according to Qualcomm, Nokia is now behind Samsung for 3G phones. So it would seem much of their current power comes from their amazing position (well earned, IMO) in emerging markets and 2G phones in general.

Regarding IMG: Very impressive, good to see it's all finally paying off :) (not saying it wasn't before, but it'd be hard to argue that these results and especially these forecasts and statements aren't the future aren't extremely positive)

[tangey]

Why the hell is Intel doing this? Right now they really don't take advantage of IMG hardware and for the future they are working on their own GPUs (plus SIMD Accelerator etc.).


Havn't seen anything to suggest that Larabee can address the power/performance/die space in the way that IMG can. Have seen anything to the contrary either, but perhaps thats the issue. Also

Finanicials from IMG today suggest a move into Laptop and PC space with intel (previously only netbook and mids were referred to).

This would encourage me in the belief that pinetrail does indeed have SGX graphics, and that a nehalem based variant will have on chip/die SGX too.

[Ailuros]

Why the hell is Intel doing this? Right now they really don't take advantage of IMG hardware and for the future they are working on their own GPUs (plus SIMD Accelerator etc.).

In what regard isn't Intel taking advantage of IMG's IP?

http://en.wikipedia.org/wiki/PowerVR

see the list of devices for the GMA500/Poulsbo alone.

And that's not even taking into account any upcoming SoCs with >SGX535 as found in Poulsbo.

Yes they're working on "high end" GPUs with LRB but Intel would be extremely dumb if they'd try to embed even one LRB core with its current power consumption and die area into an SoC and that SIMD accelerator has been defined to be a graphics core per se or just an advanced multimedia accelerator?


Just for the record's sake from IMG's final results:

Mobile Computing (MID, UMPC, Netbook devices) - Imagination’s partnership with Intel in the personal computing/UMPC and MID segment has progressed to plan with shipment of the Intel® Atom™ Z range of products that deploy Imagination’s graphics and video technologies. This solution
has already secured many OEM design wins with over 70 products shipping or announced. The partnership with Intel continues strong and with a wider co-operation scope with several significant projects underway.

[Ailuros]

Indeed, Nokia always try to get different suppliers. The problem with going with Intel is that only Intel controls Atom core. With ARM at least they can put ST vs TI vs Samsung vs etc.

Yes but as I said assuming that NOKIA truly wants an SoC for a bigger device than a smartphone and Intel probably being faster done with something more powerful than OMAP3/SGX530, Intel would be the more obvious choice since it'll still take some more time for OMAP4 from TI as an example.

[Mike11]

In what regard isn't Intel taking advantage of IMG's IP?
What I meant is that they only use SGX in an Atom version that sells very little compared to the other Atoms with Intel GPUs (which Intel is keeping around another year). And the SGX that they are selling are so badly supported that it's not even worth mentioning (drivers that are just good enough to run windows compared to what the SGX could really do).

[tangey]

What I meant is that they only use SGX in an Atom version that sells very little compared to the other Atoms with Intel GPUs (which Intel is keeping around another year). And the SGX that they are selling are so badly supported that it's not even worth mentioning (drivers that are just good enough to run windows compared to what the SGX could really do).

SGX is also used in Canmore CE3100. It will also feature in Moorestown and Medfield, likely pinetrail, and possibly a Nehalem variant.

MBX has previously featured in the CE2110 and the 2700G (now sold to Marvell)

The Intel/IMG relation is only at the beginning, putting AI into small platforms only started a year or so ago.

The latest Win7/Vista drivers apparantly are being MUCH better received in some quarters.

[Ailuros]

What I meant is that they only use SGX in an Atom version that sells very little compared to the other Atoms with Intel GPUs (which Intel is keeping around another year).

And in which other coming SoCs are they going to deploy further SGX variants?

And the SGX that they are selling are so badly supported that it's not even worth mentioning (drivers that are just good enough to run windows compared to what the SGX could really do).

That's something to blame Intel for; drivers are being supplied for Intel from Tungsten graphics while PowerVR has reference drivers for their IP that show far higher performance. Word has it that Tungsten isn't using the onboard firmware of the chip and it naturally falls back to software rendering.

How well do you know actually Intel and their driver quality for their own IGPs on the other hand? In other words it isn't a problem that's exclusive to SGX or any other graphics processor up to now sold from Intel. It's a problem that's there for years and it has absolutely nothing to do with what Intel will continue to use in future SoCs or if Intel choses to invest in Imagination Technologies.

One step further when Intel LRB's long chain of delays finally ends and the GPU finally hits shelves I honestly hope they will have changed their philosophy.

[TEXAN*]

the IMG Financial Highlights are out http://www.imgtec.com/corporate/newsdetail.asp?NewsID=472

Pre-tax profit* jumps 56% to £5m and chip unit volume up 83% to 86m

* 12 major agreements involving over 25 silicon IP core licences
o Strategically and financially important deals with Apple, Intel, MediaTek, NXP, Samsung, Sigma Designs, SiRF, and Toumaz as well as a significant agreement with a second major international OEM

SEGA?

There are rumours going around that SEGA may be working on a handheld.

[Ailuros]

SEGA?

There are rumours going around that SEGA may be working on a handheld.

Think of something bigger that starts with an "S" too ;)

[darkblu]

How well do you know actually Intel and their driver quality for their own IGPs on the other hand? In other words it isn't a problem that's exclusive to SGX or any other graphics processor up to now sold from Intel. It's a problem that's there for years and it has absolutely nothing to do with what Intel will continue to use in future SoCs or if Intel choses to invest in Imagination Technologies.
intel have different departments providing support for different intel GPUs. for instance, their current (oss) support for the gma4500 line (g45 & co) is outstanding - it's among the poster drivers in the DRI2 tree. in contrast, last time i checked gma500, it was more-or-less left on auto-pilot (read: was using a premordial tungsten edge for the 3d part).

i'm glad to see intel being so committed to using the SGX architecture, but their left and right hands need to work toward the common goal before you call intel's SGX integrations a viable product.

[tangey]

SEGA?

There are rumours going around that SEGA may be working on a handheld.

Sega could be working on something, but the RNS from IMG stated the licence was from a "major international consumer electronics company". Given that Sega havn't made any hardware in years, I don't think you could describe them as such.

Think Sony.

[Mike11]

By-the-By, Intel just bought ANOTHER 4.85M shares in IMG, they now own just over 16% of IMG.
And there it is:
Apple Inc. is subscribing for 2,200,000 new shares of Imagination Technologies Group plc (LSE: IMG) at £ 1.4275 per share, the mid market close price on 25 June 2009, subject only to listing and admission of the new shares. Apple is a licensee of Imagination's technology. Following the share placement and recent share purchases that Apple has notified IMG that it has made in the open market, Apple will have an ownership interest of 9.5%.
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=473

I thought Apple would buy more, but better than nothing :)

[tangey]

And there it is:

http://www.imgtec.com/corporate/newsdetail.asp?NewsID=473

I thought Apple would buy more, but better than nothing :)

They did buy a lot more, prior to yesterday they held 3%. They bought 2% via a share issue from IMG (in other words IMG got the money), and they bought a further 4.5% in the open market.

So they bought a further 6.5%. They now have 9.5%, Intel has 16%, between them, they now own 25% of IMG.

Thats what I call "Intel Inside"

[Ailuros]

intel have different departments providing support for different intel GPUs. for instance, their current (oss) support for the gma4500 line (g45 & co) is outstanding - it's among the poster drivers in the DRI2 tree. in contrast, last time i checked gma500, it was more-or-less left on auto-pilot (read: was using a premordial tungsten edge for the 3d part).

Current is the keyword here for 4500; how did it look like when the chipset was first introduced?

i'm glad to see intel being so committed to using the SGX architecture, but their left and right hands need to work toward the common goal before you call intel's SGX integrations a viable product.

My problem with Intel is that it takes an unexcusable amount of time for anything graphics up to date to get their drivers on a level they should be since day one. I'm afraid SGX535/Poulsbo might be even obsolete if and when any SGX gets decent driver attention.

[TEXAN*]

You're right, the OEM they mention is probably Sony.

The latest news that's come my direction says that SEGA wants to keep costs down in regards to silicon and thus will probably use the Aurora SoC. The reason for this is because they plan on utilizing an expensive force feedback based touchscreen. Apparently it is so advanced that users shall perceive the sensation of pressing, depressing, releasing of actual physical buttons, keys etc.

They are going to test the technology in the arcades first, on the RINGWIDE hardware. With the first games appearing at the end of this year -

http://www.andriasang.com/e/blog/2009/06/19/sega_dual_screen_arcade/

[roninja]

Would have thought that Aurora "Dreamcast" on a chip is too old, and they'd perhaps look to an SGX variant?

[Ailuros]

You're right, the OEM they mention is probably Sony.



http://www.fudzilla.com/content/view/14430/1/

[Simon F]

Would have thought that Aurora "Dreamcast" on a chip is too old, and they'd perhaps look to an SGX variant?
IF "Aurora" is the SoC I think it is, then it's possibly the most powerful MBX system there is and also includes one Dreamcast feature that no contemporary hardware has - i.e. translucency sorting.

[tangey]

IF "Aurora" is the SoC I think it is, then it's possibly the most powerful MBX system there is and also includes one Dreamcast feature that no contemporary hardware has - i.e. translucency sorting.

From:-
http://www.system16.com/hardware.php?id=732

Hardware : System-on-a-chip 90nm solution with integrated CPU, Graphics and Sound Hardware.
CPU : Hitachi SH-4 32-bit RISC CPU @ 300MHz
GPU : Imagination Technologies Power VR MBX+VGP @ 150MHz
SPU : ADPCM Sound
Triangle Rendering : 10M Polygons/sec
Fill Rate : 150M Pixel/sec
Max Resolution : 1280x1024
Other Features : Ethernet and USB


The oldest game mentioned on that wepage is 2005, which makes the platform 4-5 years old.

[tangey]

You're right, the OEM they mention is probably Sony.

The latest news that's come my direction says that SEGA wants to keep costs down in regards to silicon and thus will probably use the Aurora SoC. The reason for this is because they plan on utilizing an expensive force feedback based touchscreen. Apparently it is so advanced that users shall perceive the sensation of pressing, depressing, releasing of actual physical buttons, keys etc.

They are going to test the technology in the arcades first, on the RINGWIDE hardware. With the first games appearing at the end of this year -

http://www.andriasang.com/e/blog/2009/06/19/sega_dual_screen_arcade/


I wonder if this new product requires some sort of video decode? Reason I ask is that SI electronics took a VXD370 licence "for use in devices targeting the amusement market".

SI electronics was part of the development team that designed the Aurora based system.

[Lazy8s]

And Sega Sammy sold SI Electronics off a little while back.

[TEXAN*]

Yes, they moved all Research & Development in-house. Thus, SI-Electronics wasn't needed anymore.

[tangey]

To get this thread back on topic, I've seen the first confirmation that Moorestown will use SGX graphics.

I've been doing some sniffing around for moorestown drivers, to confirm that it contains SGX. the open community is the best place to look, as its a collaborative environment.

If you look here:-
http://www.opensubscriber.com/message/dri-devel@lists.sourceforge.net/11716475.html
(warning its a BIG page),

You'll find "poulsbo/moorestown DRM driver"

There's a programming flag in there "IS_MRST" which is set when dealing with moorestown variation. In terms of graphics, the moorestown code is pretty much the same as the poulsbo (psb) code, with just some slight modifications and path variants. From that code I can confirm that Moorestown definitely has SGX in it, with the SGX registers in a different area of the memory map than in poulsbo, as this snippet of code reveals:-

if (IS_MRST(dev))
dev_priv->sgx_reg =
ioremap(resource_start + MRST_SGX_OFFSET,
PSB_SGX_SIZE);
else
dev_priv->sgx_reg =
ioremap(resource_start + PSB_SGX_OFFSET, PSB_SGX_SIZE);

later in the code, the offsets are defined as:-
PSB_SGX_OFFSET 0x40000
MRST_SGX_OFFSET 0x80000

The PCI_ID for moorestown is anything from 0x4100-0x4107, meaning they are allowing for the possibility of up to 8 sku variants. This could be simply different packages, clock rate, temp range etc.

#define psb_PCI_IDS \
{0x8086, 0x8108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PSB_8108}, \
{0x8086, 0x8109, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PSB_8109}, \
{0x8086, 0x4100, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4101, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4102, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4103, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4104, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4105, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4106, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \
{0x8086, 0x4107, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_MRST_4100}, \

for comparision, the 8108 and 8109 ID's above for poulsbo are used to differentiate between the MID (L) and UMPC (W) versions of the existing SCH

Note that in general within the code, the moorestown variations seems to be dealt with in pretty much the same way as the poulsbo variation, with the clear suggestion thats its pretty much the same. There is nothing in the code that can be used to determine clock rates, so we dont' know if it operates at the same rate or quicker.

Some more interesting things are in the file.

First, Tungsten graphics are still involved. AS of Nov last year, they have been taken over by VMWARE and there is some VMWARE copyright headers in there.

Second, it appears from looking at the code, that the version of VXD that IMG supplied to Intel for poulsbo (and moorestown) is codenamed "topaz"

The above comments are my own personal interpretation of the code at that publicly accessible webpage. The file is dated March '09, so that must be borne in mind too

[Ailuros]

First, Tungsten graphics are still involved. AS of Nov last year, they have been taken over by VMWARE and there is some VMWARE copyright headers in there.

That doesn't sound either like bad or good news to me :roll:

[tangey]

That doesn't sound either like bad or good news to me :roll:

Quite...I just called it "interesting" :)


Also I've seen some follow up comments regarding those patches, and apparently some of them have been removed as there was some uncertainty regarding GPL licensing issues Vis-a-Vis closed portions of the 3D drivers.

[Tessier]

Intel's GMA 500 drivers do not seem to be mature to say the least...

DXVA works both in XP (1009 driver) and Win7 (1005 driver) using the default renderer (Overlay/EVR). However, in Win7 I had a lot of frame dropping when using MPC or KMPlayer EVR Custom Renderer (which means no subtitles). In XP, KMPlayer's VMR7 Renderless renderer works correctly, but VMR9 Windowed or VMR9 Renderless causes frame droppping.

The Intel/notebook manufacturer drivers offers VS3.0/PS3.0 but no OpenGL at all. Drivers generated with IEGD 10 (Intel Embedded Graphics Driver, XP only) does not support VS and only PS2.0. Still, the performance is almost twice of the normal driver. I did not have any luck using custom video renderers with this driver, although the default Overlay works. Unfortunately I could not generate a driver which keeps the aspect ratio for lower resolutions.

So, for me, it is interesting that PowerVR demonstrated GMA 500 in Vista with concurrent H264 playback and Quake3. It would be interesting to get some performance results measured with their in-house drivers.

[Lazy8s]

Hardware translucency sorting was re-enabled for the Aurora platform (full MBX+VGP @ 150MHz + SH-4 w/FPU @ 300MHz, aka SH3707)!?

Wow! Any possibility of doing that to an SGX platform or two?

[tangey]

The following quote is from IMG's annual report which was published today:-

"We are confident that we have the strategic relationships to be a major player in
mobile computing and to even see some of our technologies in the desktop space."

Thats the first time IMG have openly stated an expectation to be in the "desktop space".

For me that suggests either some pintrail/pineview variants will end up in entry level PCs, or there will be a nehalem variant with on-chip IMG graphics.

[Lazy8s]

Maybe video cores as well?

I think Series 6 will be the evolutionary transition from graphics processor to a main processor best suited to parallelized, floating point intensive workloads.

[Ailuros]

Maybe video cores as well?

I think Series 6 will be the evolutionary transition from graphics processor to a main processor best suited to parallelized, floating point intensive workloads.

Hmmm remember for the last one that they always have to be extremely careful since their largest customers are large semicondunctor manufacturers. You wouldn't want too many colliding interests in such a case ;)

[Tessier]

I did some benchmarks with the end-user (.1009) and IEGD10 drivers (end-user/IEGD).
PowerVR VillageMark: 13/57
PowerVR Fortune (50 cards): 4/42
3DMark fillrate: 190, 330/220, 330
MDolenc - Pure fillrate: 1002/1122
MDolenc - Z pixel rate: 68/1121
MDolenc - Single Texture: 1004/1121
MDolenc - Dual Texture: 1000/1121

PowerVR's Fortune draws several cards on top of each other. For tilers the achieved frame rate should be more or less the same irrespectively of the number of cards and the effective fill rate increases. This is the situation with the IEGD driver, however the end-user driver behaves like an IMR - the frame rate drops as the number of cards increases and the effective fill rate remains the same. VillageMark and MDolenc results also support the theory that there is someting screwed up with the HSR part when using the official drivers. Maybe there is hope for acceptable performance....

[Rootax]

You are on XP ? Because on Vista/Seven, the last end-user drivers are 7.14.10.1005, dated 21/05/2009.

http://downloadmirror.intel.com/16943/eng/relnotes_vista_gfx.htm

[Tessier]

You are on XP ? Because on Vista/Seven, the last end-user drivers are 7.14.10.1005, dated 21/05/2009.

http://downloadmirror.intel.com/16943/eng/relnotes_vista_gfx.htm

Yes, I am on XP. IEGD does not support WDDM. I have the strong feeling that the 1005 driver is quite similar to the 1009 XP (e.g. there is no OpenGL, Shader Model 3.0 is supported).

[Rootax]

Ok. I don't get why the GMA500 drivers by Intel are so crappy, performance wise... Can't they use the Power VR/IMG drivers or something ?

[Lazy8s]

While the IEGD10 drivers are a big improvement, the Villagemark and Fortune scores make me suspect more room for notable improvement is there.

[tangey]

Yes, I am on XP. IEGD does not support WDDM. I have the strong feeling that the 1005 driver is quite similar to the 1009 XP (e.g. there is no OpenGL, Shader Model 3.0 is supported).


Hi There,

Interested to here about improved performance with the IEGD drivers. I'm registered and downloaded the drivers from Intel. From the docs it seems to refer to having to build a Video biso to use with these drivers. Is this true ?, or do you just install them as a normal driver installation. I have a Dell mini 12" with Intel Z atom, so I'm looking to improve the graphics and video performance.

[Tessier]

One more interesting result using Humus's GL_EXT_Reme OpenGL benchmark. For IEGD10 the included OpenGL driver was used (written by Tungsten), for the official Intel driver Scitech's OpenGL->D3D wrapper was used. Results (IEGD --- Scitech):
Overdraw/HSR:
-------------
Overdraw factor 3, back to front: 62.18 fps --- 110.60 fps
Overdraw factor 3, front to back: 62.51 fps --- 110.55 fps
Overdraw factor 3, random order: 62.55 fps --- 110.58 fps

Overdraw factor 8, back to front: 58.53 fps --- 97.39 fps
Overdraw factor 8, front to back: 58.83 fps --- 97.51 fps
Overdraw factor 8, random order: 58.67 fps --- 97.28 fps

Fillrate:
---------
Pixel fillrate: 54.97 MegaPixels / s --- 88.21 MegaPixels / s
Texel fillrate: 232.72 MegaTexels / s --- 131.83 MegaTexels / s


T&L/High polygon count static display list:
-------------------------------------------
Pure transform: 2195968 vertices / s --- 1621888 vertices / s
2 point lights: 690836 vertices / s --- 682604 vertices / s
8 point lights: 225655 vertices / s --- 263798 vertices / s
2 directional lights: 2172489 vertices / s --- 1415564 vertices / s
8 directional lights: 1560491 vertices / s --- 1064740 vertices / s

High memory bandwidth load/texture cache efficiency:
----------------------------------------------------
One 1024x1024x32 texture: 62.05 fps --- 94.30 fps
Four 1024x1024x32 textures: 61.85 fps --- 94.18 fps


So it seems that a wrapper over the not too mature D3D driver provides better results than the "real" OpenGL driver.... (Texel fillrate is lower probably because Scitech only supports 2 textures). The HSR numbers are also interesting, because they look like a TBDR result - what is in contrast with the PowerVR Fortune results posted earlier. I wouldn't say I really understand the reasons...

[darkblu]

thanks for the bechmarking, Tessier.

The HSR numbers are also interesting, because they look like a TBDR result - what is in contrast with the PowerVR Fortune results posted earlier. I wouldn't say I really understand the reasons...
are we sure the fortune test does not alpha-blend?

edit:

tests figures really make little sense. from what i recall, poulsbo's GMA is clocked at at least 100MHz, at 2 pixels/clock top fillrate (SGX535), the quoted 55-88MPix/s seem way too low (no effective HSR assumed). also, intel themselves quote the vertex pipeline as capable of doing a vertex/15 clocks. if we asume that as a maximal rate (ie. no pipelininig, etc), those 2.2MV/s of pure transform would translate to 33MHz shader engine clock - no way in hell that'd be the GMA's clock. smells like SW T&L to me. either that or a broken/inefficient VBO mechanism.

[tangey]

tests figures really make little sense. from what i recall, poulsbo's GMA is clocked at at least 100MHz......

Only the smallest mid format devices use the 100Mhz variant of Poulsbo, the mainstream Poulsbo chipset clocks SGX at 200MHz

[darkblu]

Only the smallest mid format devices use the 100Mhz variant of Poulsbo, the mainstream Poulsbo chipset clocks SGX at 200MHz
well, i went for the worst case scenario. otherwise i seem to remember working on a crown beach board clocked at 200MHz.

[Ailuros]

http://channel.hexus.net/content/item.php?item=20357

[darkblu]

http://channel.hexus.net/content/item.php?item=20357

from the same article:

http://img.hexus.net/v2/channel/news/imaginationimage2-big.jpg

you'd think the article would go with piece of hw that was intel-based, no?

i mean, 'intel have been doing a great job with the miniaturisation of their tech. here's this TI board running our design' does not exactly drive a point home : )

[tangey]

Intel today launched the next SOC in their CE line, the CE4100. It implements SGX535, which the CE3100 also implemented. However its implementation results in two "first"s for SGX.

1) Its the first implementation of SGX at 45nm.

2) The CE4150 variant clocks 535 @ 400Mhz, which is nearly twice as fast as any previous SGX implementation.


http://www.intelconsumerelectronics.com/Download/Sodaville_Product_Brief_.pdf

[Lazy8s]

Sony most likely wouldn't, yet imagining the 400 MHz as a clock speed for a four-core SGX543 PSP2 GPU is exciting. The stencil rate would be monumental!

Sony might as well fold PlayStation4 into PSP2 as a single platform at that point.

[roninja]

http://www.linuxfordevices.com/c/a/News/Pine-Trail-scheduled-for-fourth-quarter/?kc=rss

According to the link above we may well see both
Moorestown and Pinetrail join intel's CE4100 with a PowerVR SGX535 embedded as part of the SOC. Whilst plausible I am rather suprised given the availability of SGX540 which would have been my choice for Moorestown. Meanwhile there's alot of conjecture over the Pineview soc and whether it will be a shrunck 945GSE or a GMA500.

[tangey]

Back in post #56 of this thread, I conjectured that IMG tec might appear in either pineview or in the on-graphics versions of clarkdale/arrandale, given that IMG had stated that they saw their technology appear in the desktop space.

Well it is now apparent that IMG does not feature either in Pineview, nor in Clarksdale/Arrandale :(

[tangey]

Intel have released the first datasheets on the various pineview/trail platforms:-

http://www.intel.com/products/processor/atom/techdocs.htm

[tangey]

5 min video demo of LG prototype smartphone based on Moorestown. (SGX graphics and VXD video IP).

UI is pretty polished given its still 6 months away from production, OS is Moblin.

video shows 3 apps running concurrently on the screen, one of them being a 720p video.

http://ces.cnet.com/8301-31045_1-10430096-269.html?tag=mncol;txt

[tangey]

Intel says it will out-perform all ARM chips

Pankaj Kedia on Moorestown:

"We will surprise everyone," he said. "We will out-perform ARM chips hands-down, and we will have better graphics on Moorestown than anything else."

http://channel.hexus.net/content/item.php?item=22489

Looks like SGX in moorestown might be running faster than the 200Mhz in Poulsbo, (but probably not at the 400MHz as implemented in the CE4100 chip).

In which case given that it launches in Q2, with likely end product likely by end of year, his assesment of it having the best graphics might be on the money. I assume it also has IMG video encode and decode (VXD and VXE)

[rpg.314]

Intel says it will out-perform all ARM chips

Pankaj Kedia on Moorestown:

"We will surprise everyone," he said. "We will out-perform ARM chips hands-down, and we will have better graphics on Moorestown than anything else."

http://channel.hexus.net/content/item.php?item=22489

Looks like SGX in moorestown might be running faster than the 200Mhz in Poulsbo, (but probably not at the 400MHz as implemented in the CE4100 chip).

In which case given that it launches in Q2, with likely end product likely by end of year, his assesment of it having the best graphics might be on the money. I assume it also has IMG video encode and decode (VXD and VXE)

Vendor side PR. doesn't count for anything.

[Ailuros]

tangey,

Do we know what version of VXD Moorestown will have integrated? Download the newest whitepaper and you'll know why I'm asking.

[tangey]

Vendor side PR. doesn't count for anything.

Always worth considering, but given that they currently *do* have the fastest SOC graphics in production with the 400Mhz implementation of SGX535 in the 45nm CE4150, its not just PR speak.

[Entropy]

Vendor side PR. doesn't count for anything.

Not only is it PR, it leaves out the most critical aspect of these products, power draw.
Intel has been capable of outperforming ARM products for a long time. And nobody in these segments have given a damn, of course. Now that it's so bleedingly obvious that small portable computing devices is where the action will be that even the PC companies belatedly wants in, it seems that the PR (and the PC industry observers they talk to) still haven't understood that this ballgame plays by different rules, and that "performance" isn't where it's at. It never was. It was always a second tier consideration, and it still is because for the overwhelming number of applications extended battery life brings much higher value than another 50% in performance.

So rather than any factual content, this little tidbit is interesting because it shows Intel as a PC company trying to sell to PC people.

[rpg.314]

Always worth considering, but given that they currently *do* have the fastest SOC graphics in production with the 400Mhz implementation of SGX535 in the 45nm CE4150, its not just PR speak.

What is the platform-perf/platform-power wrt OMAP4, Tegra2, other competitiors?

[tangey]

What is the platform-perf/platform-power wrt OMAP4, Tegra2, other competitiors?

of the CE4150 ?, I think its around 8watts, but then again thats not meant as a competitor for the platforms you mentioned.

What *I* said was that Intel has the fastest SOC-ed graphics in production, regardless of power/performance. no one has a faster one taking more power...or a faster one taking less power, unless you're in a position to tell me different, in which case I'm all ears.

Now whether their implementation in Moorestown will be the "fastest" for that sector is a different matter, but the Intel PR didn't say fastest, it said "better". It is however liekly that there will be a number of variations of moorestown for different sectors, one of which is possible to have the same 400Mhz SGX as they have implemented in the CE4150, which I would imagine would be "faster" than upcoming Omap4 or Tegra2 which will be in end products in similar timeframes (maybe a bit longer for Omap4).

[Simon F]

Download the newest whitepaper and you'll know why I'm asking.
:?: Whose white paper?

[rpg.314]

of the CE4150 ?, I think its around 8watts, but then again thats not meant as a competitor for the platforms you mentioned.

What *I* said was that Intel has the fastest SOC-ed graphics in production, regardless of power/performance. no one has a faster one taking more power...or a faster one taking less power, unless you're in a position to tell me different, in which case I'm all ears.

Now whether their implementation in Moorestown will be the "fastest" for that sector is a different matter, but the Intel PR didn't say fastest, it said "better". It is however liekly that there will be a number of variations of moorestown for different sectors, one of which is possible to have the same 400Mhz SGX as they have implemented in the CE4150, which I would imagine would be "faster" than upcoming Omap4 or Tegra2 which will be in end products in similar timeframes (maybe a bit longer for Omap4).

In the smartphone market, absolute speed doesn't count, platform-perf/platform-power does.

[Ailuros]

What is the platform-perf/platform-power wrt OMAP4, Tegra2, other competitiors?

We'll see when all of them show up in final devices. If you're having any doubts in terms of Intel claiming to be ahead in process technology it's easy to find out if they're right or exaggerating.

[Entropy]

In the smartphone market, absolute speed doesn't count, platform-perf/platform-power does.

Power draw is an even stronger factor than that. You can´t sell these kinds of devices if battery life isn't up to snuff. The aggravation of having a dead device until you can manage to revive it at a power outlet is MUCH greater than not having a particular power-hog feature. What good is a phone that isn't working? What are the consequences of going off on a business trip, and having forgotten to bring the charger? Or if you're off for the day, forgetting to charge, and then having it die on you in the middle of a conversation?
Bad battery life is unaccepatable, pure and simple. It is a consideration that has higher priority than any other feature, and by a wide margin.

A phone with a flat battery has no feature at all.

(Edit: This is more of a user perspective, my wife and I both buy and use these kinds of devices and have for some time. She has killed two phones by throwing them to the pavement for battery related issues.... And she was right to do so. Battery issues is awful human engineering, and not to be tolerated. And she has refused those brands since.)

[Ailuros]

Beyond doubt; but before you go out on a limb and ask about X or Y platform's power consumption (and then hw characteristics and eventually performance) the more burning question in this case is when each platform will show up in final devices.

The Intel gentleman in that Hexus pic is holding two devices; I haven't seen any OMAP4 powered devices yet.

[tangey]

In the smartphone market, absolute speed doesn't count, platform-perf/platform-power does.

Indeed which is why historically IMG have won graphics seat (including Intel) in places where power alone was the overwhelming criteria, or power/performance was.

[rpg.314]

Power draw is an even stronger factor than that. You can´t sell these kinds of devices if battery life isn't up to snuff. The aggravation of having a dead device until you can manage to revive it at a power outlet is MUCH greater than not having a particular power-hog feature. What good is a phone that isn't working? What are the consequences of going off on a business trip, and having forgotten to bring the charger? Or if you're off for the day, forgetting to charge, and then having it die on you in the middle of a conversation?
Bad battery life is unaccepatable, pure and simple. It is a consideration that has higher priority than any other feature, and by a wide margin.

A phone with a flat battery has no feature at all.

(Edit: This is more of a user perspective, my wife and I both buy and use these kinds of devices and have for some time. She has killed two phones by throwing them to the pavement for battery related issues.... And she was right to do so. Battery issues is awful human engineering, and not to be tolerated. And she has refused those brands since.)

Isn't battery life a function of standby power and the peak power draw when running at full/near-full load? And isn't the latter related to platform perf/W?

[Entropy]

Isn't battery life a function of standby power and the peak power draw when running at full/near-full load? And isn't the latter related to platform perf/W?

Not sure quite what you're asking. For phones, standby power is very important, obviously. A phone that has a habit of being out of battery when you rush out in the morning is aggravating to say the least. That would go for a tablet/pad as well - devices that are discharged when you pick them up for use are simply frustrating.

For more multifunctional devices power draw when active becomes more important than it used to be, because they are simply used more. Not only for classical PDA tasks but, (the big ones) web surfing, gaming and media playing. Battery life is a greater concern now than it used to be, not lesser. The idea is still that these devices can be carried along with you all the time and always be accessible when desired/needed, but we expect them to do more for us without the usage model being compromised.

[roninja]

so whats next for Intel and its PowerVR'd SoC's perhaps....

http://www.netbookchoice.com/2010/02/18/intel-%E2%80%9Coak-trail%E2%80%9D-set-to-replace-atom-z-series/

Intel may be set to launch a replacement of the Atom Z series of processors according to PC Watch. The Atom Z CPUs are found on a number of netbooks including the Sony Vaio P, MSI Wind U115 and Asus Eee PC T91MT. Intel had already announced Moorestown, the successor to Menlow that includes the Atom Z series, as part of its MID/smartphone platform.

Rumours suggest that Intel may be launching a new version of Menlow called “Menlow Plus” or “Menlow Refresh” with clock speeds faster than 2GHz. However, another rumour suggests that Intel is working on a Moorestown version for PCs dubbed Oak Trail. It’s all rumours for now but we’ll keep you posted with any developments

Couple more articles here....

http://nexus404.com/Blog/2010/02/19/intel-atom-oak-trail-coming-soon-report-says-oak-trail-coming-to-replace-atom-z-series-processors/

http://www.pocketables.net/2010/02/zseries-atom-rumored-to-be-replaced-by-intel-oak-trail.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+pocketables/PpUx+%28pocketables%29

http://www.pcworld.com/article/189830/intel_wants_to_put_moorestown_chips_into_tablets.h tml

[MfA]

I doubt Intel has any interest in pushing PowerVR higher up the foodchain, as long as IMG has a good IP portfolio to fill up the gaps for them at the low end they will gladly make use of them ... but I don't see it lasting, they have too many of their own irons in the fire.

[roninja]

This is essentially a Menlow refresh. So I would assume they would continue with the PowerVR involvement in that side of things especially being the (potential) lead partner for SGX 545.

[tangey]

Intel has put up a "white paper" regarding optimising 3D applications for use with SGX". with some demo code.

The abstract refers to SGX being used in "Intel® System Controller Hub US15W, and similar future platforms"

Body of the paper briefly refers to "In Vehicle Infotainment market".

http://edc.intel.com/Download.aspx?id=3422&returnurl=/Software/Downloads/IEGD/default.aspx


At the moment I see SGX appearing in Moorestown Medfield and Tunnel Creek, and possibly "Tioga Peek"

Also Queens Bay/Tunnel Creek will be talked about in another couple of weeks by Intel. My research indicates that this is the successor to Menlow, and will contain 400MHz SGX graphics integrated onto the chip ala sodaville.

I'm also starting to see some reference to "Tioga Peek" which appears to be in the same stream as Sodaville, i.e. for CE products.

So thats Moorestown/Medfield for handheld/phones, Tunnell Creek for embedded, and Sodaville/Tioga Peek for CE such as STB and TVs.

[tangey]

Intel has formally announced the Z600 series ( previously known as Moorestown).

Anandtech have a comprehensive article on it.
Up to 1.5Ghz for phones, up to 1.9Ghz for tablets.

As I have conjectured for quite some time now:-
Graphics are SGX535, from 200Mhz to 400Mhz.
Both Video Decode and Video Encode I/P are included, again supplied by IMG.
(is this the first known implementation of IMGs Video encode IP ?)

400MHz graphics version targetting tablets (that gives it almost twice the graphics engine of the ipad).

Runs Quake III @ 100fps visually confirmed by Anand (resolution unknown)

Power/ times (supplied By intel, not confirmed by Anand)

Sunspider score Z600=sub-2s , ~10 seconds on the iPad:

On the mobile reference plaftorm supplied by AVA (800x480 screen)
Standby time:285 hours
Talk-time:8hrs (2G)
Web_browsing:5 Hrs
audio playback:48 hrs
HD_video playback:5hrs

Platform idle 23mW (snapdragon is apparently 25mW ?)

If those figures are realistic, I'd say this makes this a viable platform. Anand does say that the 5-chip chip-count means there are space issues.

http://www.anandtech.com/show/3696/intel-unveils-moorestown-and-the-atom-z600-series-the-fastest-smartphone-processor/1

[rpg.314]

Intel has formally announced the Z600 series ( previously known as Moorestown).

Anandtech have a comprehensive article on it.
Up to 1.5Ghz for phones, up to 1.9Ghz for tablets.

As I have conjectured for quite some time now:-
Graphics are SGX535, from 200Mhz to 400Mhz.
Both Video Decode and Video Encode I/P are included, again supplied by IMG.
(is this the first known implementation of IMGs Video encode IP ?)

400MHz graphics version targetting tablets (that gives it almost twice the graphics engine of the ipad).

Runs Quake III @ 100fps visually confirmed by Anand (resolution unknown)

Power/ times (supplied By intel, not confirmed by Anand)

Sunspider score Z600=sub-2s , ~10 seconds on the iPad:

On the mobile reference plaftorm supplied by AVA (800x480 screen)
Standby time:285 hours
Talk-time:8hrs (2G)
Web_browsing:5 Hrs
audio playback:48 hrs
HD_video playback:5hrs

Platform idle 23mW (snapdragon is apparently 25mW ?)

If those figures are realistic, I'd say this makes this a viable platform. Anand does say that the 5-chip chip-count means there are space issues.

http://www.anandtech.com/show/3696/intel-unveils-moorestown-and-the-atom-z600-series-the-fastest-smartphone-processor/1

Moorestown is single core, right?

[Rys]

Yeah, single core with HT.

[Rys]

UzWGQaPEF9Y

[MfA]

Over 5 times faster than the demonstration at computex last year.

[wishiknew]

Reading the Anandtech article. Thats a high number of chips to get this working. Even more if the phone has an above 5mp camera.

[Mike11]

Reading the Anandtech article. Thats a high number of chips to get this working. Even more if the phone has an above 5mp camera.
That's my main problem with it too. 3-5 chips instead of 1-2 with ARM SoCs. Hopefully with Medfield this problem will be resolved. H2/2011 could get interesting.

Regarding SGX, I'm a little bit disappointed that Intel didn't choose SGX545 over SGX535. I really thought Intel bould be the SGX545 launch partner.

And finally someone used IMG's VXE core in a SoC (well, compared to ARM SoCs Intel's Z600 doesn't really qualify as a SoC).

[fehu]

I'm curious to see if a dual core A9 would be competitive on perf/W, maybe it's me but isn't easy to imagine all that stuff closed on a phone and lasting enought...

And what about "we don't support WP7"?
WP7 it's built for arm processors, what kind of support is intel talking?

[rpg.314]

And what about "we don't support WP7"?More like MS kicked them out the door (for whatever reasons), and with the benefit of hindsight they figure that grapes are sour.

WP7 it's built for arm processors, what kind of support is intel talking?

I guess MS refused to maintain 2 ISA branches just to focus better on what they have now.

[aaronspink]

Reading the Anandtech article. Thats a high number of chips to get this working. Even more if the phone has an above 5mp camera.

anyone putting a >5 mp camera in a cellphone with no optics or bad plastic optics should be shot. Hell, realistically, I haven't seen a cellphone except the specific cameras with a built in cell phone, that even has optics worthy of 1 MP.

[wishiknew]

Don't know about the optics but Nokia's N8 at least there's suppose to be a big size sensor in there.

[fehu]

More like MS kicked them out the door (for whatever reasons), and with the benefit of hindsight they figure that grapes are sour.

I guess MS refused to maintain 2 ISA branches just to focus better on what they have now.

yes i know, but anand isn't a stupid, would have make joke of this statement if it was completely pr stunt
maybe has something to do with arm emulation over x86? looks fairly impossible btw...

[fehu]

Don't know about the optics but Nokia's N8 at least there's suppose to be a big size sensor in there.

and from the video and photo made with it, its worth it

this first intel generation with limited "phoneness" looks like an advanced research project partially offloaded to the buyers' pockets

[Lazy8s]

Dual-core A9 SoCs launching around Moorestown's release leave it at a competitive disadvantage in many ways, but Medfield takes advantage of Intel's lead in process technology to really shine.

Just because IMG needs to cite a specific clock speed as a point of reference for quoting SGX performance and that point they chose was 200 MHz, it doesn't mean their cores have been running at 200 MHz in the actual phones all this time. Articles always seem to assume this. They also always assume that the performance difference between SGX535 and the 530 is double the triangle rate and not double the texel fill.

[Ailuros]

Don't know about the optics but Nokia's N8 at least there's suppose to be a big size sensor in there.

I'm not sure but I'd guess the N8 contains a Broadcom SoC (at least I can't think of any other SoC with an ARM11). Yes it contains a OGL_ES2.0 GPU in its paperspecs with funky peak triangle rates but I'm afraid that's about it.

That's my main problem with it too. 3-5 chips instead of 1-2 with ARM SoCs. Hopefully with Medfield this problem will be resolved. H2/2011 could get interesting.

Regarding SGX, I'm a little bit disappointed that Intel didn't choose SGX545 over SGX535. I really thought Intel bould be the SGX545 launch partner.

And finally someone used IMG's VXE core in a SoC (well, compared to ARM SoCs Intel's Z600 doesn't really qualify as a SoC).

I'm personally not disappointed at all. SGX545 might have a high feature-set but it also doesn't come for free either. I'd rather prefer Intel to invest for it's next generation of embedded SoCs that die area in performance or else in SGX543 MP. According to IMG under 65LP (@200MHz) 545 equals 12.5mm2 and 543 8mm2 per core. 543 might be limited to SM3.0+ but with a 2MP/16mm2@65LP you get more than twice the performance of a 545.

For the moment and always IMHO fill-rates have a sizable importance in the embedded space. Going to a 545@200MHz vs. 535@400MHz as an example means twice the fill-rate in the latter case, since all cores =/>535 have 2 TMUs. Frequency doesn't scale unfortunately in parallel with die area.

[rpg.314]

yes i know, but anand isn't a stupid, would have make joke of this statement if it was completely pr stunt
maybe has something to do with arm emulation over x86? looks fairly impossible btw...May be he didn't want to offend his contacts @Intel? :???:

[MfA]

Apart from the DRAM I don't see how the extra chips matter, tiny chips with tiny bandwidth interfaces ... their integration wouldn't save much power or area AFAICS.

[tangey]

I'm personally not disappointed at all. SGX545 might have a high feature-set but it also doesn't come for free either. I'd rather prefer Intel to invest for it's next generation of embedded SoCs that die area in performance or else in SGX543 MP. According to IMG under 65LP (@200MHz) 545 equals 12.5mm2 and 543 8mm2 per core. 543 might be limited to SM3.0+ but with a 2MP/16mm2@65LP you get more than twice the performance of a 545.

For the moment and always IMHO fill-rates have a sizable importance in the embedded space. Going to a 545@200MHz vs. 535@400MHz as an example means twice the fill-rate in the latter case, since all cores =/>535 have 2 TMUs. Frequency doesn't scale unfortunately in parallel with die area.

Tunnel Creek is using SGX535 @ 200Mhz-400Mhz.
Sodaville (CE4100) is using SGX545 @200-400Mhz.

Today we get that Moorestown is similarly using SGX535 @ 200-400Mhz.

So the pattern is getting clear, what Intel use on one Soc platform, they use on the others.

Intel said at IDF that the generation after tunnel creek would be x5 the graphics performance of tunnel creek. The said today that Medfield would be x2 the graphics of Moorestown.

Assuming they stick with IMG, clearly looks like will be seeing either 545 or 543MP next year in Medfield, with a higher clocked/increase cores version in Tunnel Creeks successor.

[Mike11]

Apart from the DRAM I don't see how the extra chips matter, tiny chips with tiny bandwidth interfaces ... their integration wouldn't save much power or area AFAICS.
What do you mean with tiny chips? Apart from the extra DRAM chip, there's also Langwell, which is even bigger than Lincroft (14x14x1.3mm vs. 13.8x13.8x1.1mm).
Thank god at least Langwell is gone for sure with Medfield next year (i really hope the extra DRAM chip is gone too).
http://images.anandtech.com/reviews/soc/intel/moorestown/IntelAtomProcessorZ6xx(Lincroft)and_PlatformContro llerHub_package2_sm.jpg

For tablets this isn't that big of a problem, but the main target here is smartphones.
I think 2-3 chips more matter if you want to build compact smartphones with mainboards like this (especially for next-gen devices that launch in Q4/2010):
http://cache-04.gawkerassets.com/assets/images/4/2010/04/500x_open19_01.jpg

[Rys]

Arun, if you're reading, do you have any preliminary thoughts on Z600 and the silicon required for designs, specifically with regards to integration and the area and number of chips (especially the mixed signal chip and what radio ICs you see being paired with the chipset) ?

[rpg.314]

OT: Why are power management ICs always separate from the main SoC?

[MfA]

Mike11 ... okay, but those were already known about ... I just think making a big deal being made about the extra power and RF chips seems overblown.

[tangey]

For tablets this isn't that big of a problem, but the main target here is smartphones.
I think 2-3 chips more matter if you want to build compact smartphones with mainboards like this (especially for next-gen devices that launch in Q4/2010):


The platform doesn't totally seem to exclude iphone sized devices, given that Anandtech stated that the demo phone from AVA was narrower, thinner, but a bit longer than the iphone 3GS, and included a 1500mAh battery

[liolio]

Chipzilla is impressive that should no longer astonish but still :)

By the way the article of techreport (http://techreport.com/articles.x/18866/1)is up.

[Arun]

Arun, if you're reading, do you have any preliminary thoughts on Z600 and the silicon required for designs, specifically with regards to integration and the area and number of chips (especially the mixed signal chip and what radio ICs you see being paired with the chipset) ?Oh hai!

The level of integration is just fine besides for the very real disadvantages of the two-chip approach and the lack of Package-on-Package DRAM for LPDDR1 (a very strange omission since it's widely supported on 14x14 packages - surely Intel didn't *need* to use a 13.8x13.8 package?). One interesting point is that both packages are 0.5mm pitch, versus 0.4mm for high-end ARM SoCs, which could save a little bit of cost on the PCB but besides that it's apparent that Intel's packaging expertise is rather below average.

The I/O partitioning between the two chips is also bizarre: LVDS on Lincroft and HDMI on Langwell, for example - whereas ARM competitors usually integrate HDMI but need a separate chip for LVDS in tablet designs (because LVDS normally needs extra process steps on modern process nodes iirc). Putting the ISP on Langwell is also a bizarre choice and feels more like a late-stage addition (same for HDMI), although in the end it doesn't really matter. On the other hand being limited to 5MP is rather absurd, especially as it's easy to get an on-sensor-package ISP at 5MP or below (see: Apple, they don't integrate an ISP!) but rare above so you'd need a moderately expensive discrete chip.

In the seemingly cancelled LG GW990, the Z600 was combined with a Ericsson 3G module (i.e. pre-certified in several key ways to simplify design/time-to-market but increasing costs) based, afaik, on the ST-Ericsson M340. But it could probably be combined with any of the very popular slim modems in the industry, such as the Infineon X-Gold 616 or Icera ICE8040 for WCDMA and Qualcomm QSC6085 for CDMA. The area penalty of not integrating the baseband is diminishing all the time, and the RF side is probably more important right now anyway (see: Infineon UE2 for a leading-edge example). The decision to work with Nokia on baseband technology for the mid-term highlights Intel's lack of understanding of the field, IMO, and is a negative overall in my mind.

Power management is interesting. The various implementations of Briertown by different analogue companies (not necessarily single-chip in fact but that's not important) are probably the most advanced chips in their field I'm aware of - and also the most expensive. Seriously, NINE DC/DC?! The latest ARM SoCs used 4 at most. Part of it is the extra complexity of the two-chip approach each needing different voltages, but there's also a nice bonus feature: you can power the baseband with no secondary power management chip. It's really a system solution (using partnerships with the likes of ST-Ericsson and Icera) which is nice, even if overall power management is still more expensive and it limits their choices a bit because Infineon for example integrates the PMU (SoC on 65nm, interestingly back to SiP on 40nm - guess they figured the cost/mm2 was too expensive for something that nearly doesn't scale!)

Connectivity-wise, WLCSP and QFN packages are very cheap, and there isn't that much that could be shared between standards anyway. So the lack of integration is irrelevant. The only company that pushed aggressively on that front in the 3G market was Broadcom, and they achieved a grand total of zero design wins for their 'Zeus' baseband (see Linley Group's news on that). Integration is nice, but it's nearly always secondary to other things.

Talking of other things:I could be reading the slides wrong, but it seems to me they're hinting at only 720p support when limited to 32-bit LPDDR1. Being limited to 20Mbps at 1080p is also intriguing, I wonder if this is the same VXD375 as in the A4? The advantage there against TI or NVIDIA or basically nil to be honest (NV is limited to low bitrates at HP on 1080p, but it's enough for YouTube support iirc). At least in terms of 3D graphics, it's truly leading-edge with such a highly clocked SGX535. The idea of putting an audio island on Langwell also seems like a fine strategy even if power consumption there is still a fair bit higher than you'd want it to be.

Overall it's obviously a MASSIVE step-up from Menlow, and Medfield should be truly excellent in many ways. I wonder how Rayfield would react if he realized I was actually (only slightly) too *pessimistic* about Moorestown's power numbers, hah. However, it remains expensive and, most importantly, the game has moved on. When Menlow came out the OS war was still raging and this coupled with Moblin and a 3D UI could have been a very strong contender. But now they're pushing for MeeGo and Android (presumably Chrome OS later) - but Android doesn't make perfect sense. The Java apps will be compatible cross-ISA, but as of 1.5 (or was it 1.6?) there's a native SDK which compiles directly to ARM, and it's what's used for anything interesting 3D-wise for example. So x86 has now become a disadvantage OS-wise (even if not as much as it could have been) which is ironic given Intel's past marketing.

The biggest business opportunity for Intel here is if they can get Nokia completely on board in larger MeeGo form-factors (5-12"), along with a bunch of ODMs and PC-centric OEMs (even though they all have ARM, usually Tegra, R&D programs now). They won't get smartphone volume on this part, no way, but a few flagship design wins could give them credibility with Medfield, and that would spice things up.

Intel's 32nm process seems very strong, and the competitiveness of ARM-based SoCs will depend a bit on TSMC's execution for 28nm - specifically the 28LPT and 28HPM nodes (I don't know about TI, but I know nearly for certain that those are the specific process variants used by both NVIDIA & Qualcomm for Tegra/Snapdragon 3 & 4 respectively). Moorestown feels to be a bit too little, a bit too late to me - but definitely big kudos to the engineering teams either way.

[rpg.314]

Intel's 32nm process seems very strong, and the competitiveness of ARM-based SoCs will depend a bit on TSMC's execution for 28nm So medfield is @Intel 32nm?

What happened to their partnership with TSMC? I thought Langwell @TSMC meant that Medfield will be @TSMC too.

Or are they gonna dump the lamest SoC's @TSMC's doorstep (netbook stuff?), ie where having x86 is actually a benefit. :wink:

[aaronspink]

OT: Why are power management ICs always separate from the main SoC?

Because generally they have very different requirements like being able to support high efficiency power MOSFETs, different pin voltage requires, not really requiring high speed logic, mostly analog instead of digital.

[DavidC]

So medfield is @Intel 32nm?

What happened to their partnership with TSMC? I thought Langwell @TSMC meant that Medfield will be @TSMC too.

Or are they gonna dump the lamest SoC's @TSMC's doorstep (netbook stuff?), ie where having x86 is actually a benefit. :wink:

Didn't that go bonkers because TSMC cancelled their 32nm? Or whatever problem they are having with 32nm-ish generation. I'm pretty sure Intel doesn't want to deal with that. :)

[Lazy8s]

Intel would want to use their process advantage on the critical chip of the platform, so Lincroft and especially the Medfield SoC would be at Intel. Langwell can be a mix of a bunch of 3rd party IP, so TSMC is better suited there.

The Anand article made it pretty clear that politics and not technology is what makes sense as the real, main reason for the lack of a Windows collaboration.

I thought the N8 was using an OMAP3430. Going by the quality of the 5 MP "Carl Zeiss optics" cameras they've been using on previous phones, a similarly-well implemented 12 MP version with relatively large sensor should make for a great camera phone.

[roninja]

Nope N8 is officially listed as a 680mhx Arm11 core which the likes of Arun have pointed to being a Broadcom part on other forums. Nokia seems to continue with a scattergun approach across multiple hardware vendors (TI, Broadcom, Qualcomm, ST-Ericsson, Intel) I am not sure what this approach will achieve in the long run.

[Ailuros]

I thought the N8 was using an OMAP3430. Going by the quality of the 5 MP "Carl Zeiss optics" cameras they've been using on previous phones, a similarly-well implemented 12 MP version with relatively large sensor should make for a great camera phone.

I can't complain about the quality of the quick shots I occasionally shoot on the N95 which has a 5MP (Carl Zeiss lens). However video capture quality is as crappy (dropped frames to oblivion) as on the N900 I've laid my hands on quickly. You'd be surprised how well any iPhone does with video capturing in contrast.

As in all markets funky on paper marketing numbers obviously never tell the entire story.

[Lazy8s]

I'm actually trying to help someone choose a phone and the quality of the still and video photography is probably the biggest factor for them, so your impressions have been helpful. Thanks!

After Samsung's trouble in getting TI's video encoder to let the Omnia i8910 live up to its HD billing, I was a little concerned about recommending the N900 to the person I'm helping. It's otherwise an almost perfect device for them, but TI apparently should swallow their pride for homegrown DSP/custom video processors and license some VXE/VXD.

[Ailuros]

The N900 is a mixed bag and despite a very nice device overall it'll take its time until the OS matures. Web browsing is outstanding but I personally wouldn't use it anyway; too small too impractical for my taste. Tablet sized devices are far better fit for browsing IMO.

There are tons of reviews to find for each device like that one: http://www.techradar.com/reviews/phones/mobile-phones/nokia-n900-655862/review?artc_pg=6

The i8910 is still in a league of it's own: http://www.techradar.com/reviews/phones/mobile-phones/samsung-i8910-hd-531931/review?artc_pg=6

Obviously you can never have it all and it goes without saying that irrelevant of lens, flash types and amount of megapixels in specsheets a mobile phone camera still doesn't equal a good digi-cam.

[tangey]

In a Q&A at the end of their investors meeting today, Intel's Paul Otellini said:-

"in the handheld business, PowerVr is our currently graphics of choice, at least for the next few generations."

He also implied that although they were still working on Larrabee, none of their existing roadmaps were reliant on it.

[tangey]

Intel's expanding line of Atom based Socs with IMG IP in them is starting to look impressive.

TV & STB .....Canmore (CE3100) & Sodaville (CE4100,4130,4150)
In-car/digital signage.............Tunnel Creek, multiple SKUs.
Small form factor.......Moorestown (launching with 7 SKUs)

Q1 2011
Tablet/ netbook..........Oaktrail (aka moorestown -W)
Smartphones.......Medfield.

All of the above have SGX for the graphics and VXD video decode. Most if not all will also have VXE video encode IP.

[tangey]

It looks like Intel has licensed IMG's vector graphics I/P in addition to the 3D graphics and video encode/decode.

The following entry on the Khronos compliance pages was added a couple of days ago.

CPU: x86
OS: Linux2.6
API pipeline:
VG_VENDOR "Imagination Technologies"
VG_RENDERER "PowerVR VGX"
VG_VERSION "1.1"

As far as I am aware, thats the first official sighting for VGX hardware from IMG.

The block diagram that I've previously seen for moorestown doesn't show a separate block for it (could be there of course). So I wonder is this for the TV/STB Socs such as the CE4100,4200 or elk.