View Full Version : Apple going with IMG for years to come ?
IMG (Imagination Technologies) today announced a multi-year, multi-use, multi-technology "extension" to the relationship with the "electronics systems company" that they first signed up and announced in July '07.
http://www.imgtec.com/News/Release/index.asp?NewsID=392
Although it has never been confirmed by IMG, there is a broad consensus that the "electronics systems company" mentioned in July last year is Apple.
IMG state in the PR that a number of new Socs will be developed as a result.
Given that the Jul-07 announcement signalled the licensing of SGX (graphics) and VXD (video), one assumes that todays announcement implies other members of the SGX & VXD families, or indeed IP that has not yet fully been developed. It would therefore be possible that this might include a move into the Apple Laptop/notebook realm. IMG have previously stated that a move into Laptop segment would require a lead partner.
This announcements ties in well with the recent ARM Archeitectural licence announcement from the end of July with a major "handset OEM". The similarity is that ARM has not disclosed the licencee by name, and the tie-in is that Apple has states that the recent purchased of PA semi was to faciliate the design of chips for the ipod and iphone areas.
The above would suggest that Apple has gone with IMG graphics technology even though ARM's Mali technology was probably available at little extra cost given the wide scope of the ARM license.
roninja
04-Sep-2008, 18:23
http://www.eetimes.com/news/design/showArticle.jhtml?articleID=210300503
Looks like the most significant deal IMG have achieved to date, and congratulation all involved including those that frequent this board, and like Tangey states above the opportunities are almost endless. I think IMG are looking for a lead partner for certain SGX cores, and I feel Apple might be looking at a whole collection of them.
Probably the best announcement after announcing intel as their licensees. Now I wonder if Apple is going to come up with a game console. It's not too unlikely.
RudeCurve
05-Sep-2008, 09:15
It would be petty sweet if Apple entered the portable gaming sector.
roninja
05-Sep-2008, 10:44
Thought they already had with iPhone / iPod touch etc, its the fact that future generations will incorporate SGX and beyond that implies gaming will become even more compelling on the platform.
RudeCurve
05-Sep-2008, 10:53
They've only touched the surface with iPhone/iPod, more experimental than full all out portable gaming console. To seriously compete with Sony and Nintendo, Apple needs to have a portable that's competitive price-wise.
Edit: Actually the new iPhone 3G is already competitively priced at just $200.
Ailuros
06-Sep-2008, 07:04
It might sound like a dumb question, but why has it to be Apple? I mean most of the interested parties know by know that Apple has used IMG IP for its iPhone, which doesn't justify the re-newed secrecy. Anyone want to take any bets on something like SONY? ;)
roninja
06-Sep-2008, 21:18
Well IMG have never publically announced they are in the iPhone for starters, and secondly the PR refers to the first licence deal in July 2007 and we know of the subsequent Samsung manufacturing licence for SGX/VXD. We know of the relationship with Samsung providing the SoC for the 1st and 2nd iPhones. We know about PA Semi. We know about the ARM architectural license again to unnamed company back in July. There is too many coincidences here to think anyone other than Apple being the mystery licencee.
Skrying
06-Sep-2008, 21:43
They've only touched the surface with iPhone/iPod, more experimental than full all out portable gaming console. To seriously compete with Sony and Nintendo, Apple needs to have a portable that's competitive price-wise.
Edit: Actually the new iPhone 3G is already competitively priced at just $200.
Except a Sony or Nintendo handheld doesn't come with a extremely expensive two year contract.
RudeCurve
07-Sep-2008, 05:59
Except a Sony or Nintendo handheld doesn't come with a extremely expensive two year contract.
True but you can't make phone calls with the SONY or Nintendo portables either. I could see the iPod touch being in direct competition though since the price could be reduced and there's not contract, Apple just needs more games support.
Skrying
07-Sep-2008, 10:26
True but you can't make phone calls with the SONY or Nintendo portables either. I could see the iPod touch being in direct competition though since the price could be reduced and there's not contract, Apple just needs more games support.
Eh? Apple would then have to impose a licensing cost per game on developers, it'd have to actual create tons of business ties, etc, etc. The iPhone cost a significant amount of money to manufacturer and the iPod Touch is still not cheap. It just doesn't make sense to me at all for Apple to entire the gaming business in any way. To long to build up and even then they might never succeed.
roninja
07-Sep-2008, 21:33
Well by the end of 2008 Apple would have sold 20m units of iPhone and iPod touch since introduction which as a user base is pretty impressive in my mind, and a great opportunity for the likes of EA etc to exploit.
Skrying
07-Sep-2008, 23:05
Well by the end of 2008 Apple would have sold 20m units of iPhone and iPod touch since introduction which as a user base is pretty impressive in my mind, and a great opportunity for the likes of EA etc to exploit.
With value marketed games, yes. But good luck selling a $30 game to that market with any type of success. It's very different.
Entropy
07-Sep-2008, 23:35
With value marketed games, yes. But good luck selling a $30 game to that market with any type of success. It's very different.
In a way, it is similar to selling a game to the home PC market. Most home PC owners aren't interested in games at all, many however have some interest, and a few are very interested. The iPhone will probably have a much larger percentage of users interested in some kind of gaming but will otherwise have a similar distribution.
When the market is big enough, you can choose which demographic to target. While the iPhone/iPod represents a different market than that of the DS and PSP, there is little doubt that it will eventually be larger. There is room for many winners in that pond, occupying different places in the ecosystem. Also, it doesn't necessarily overlap the pure handheld gaming devices completely, so they could lead healthy lives even in the event of a successful gaming market for the iPhone. However, if Sony or (much less likely) Nintendo try to respond by trying to outdo the iPhone on its own turf, I think they will fail to do so. Spectacularly.
What I find particularly interesting in the press release is the referal to "future" IP: "a new multi-year, multi-use license agreement which gives it access to Imagination’s wide range of current and future POWERVR® graphics and video IP cores." The SGX is pretty impressive in its own right, but having more advanced developments already covered in the deal... Nice. Can't help wondering about what it might be, and what the future might bring.
Skrying
08-Sep-2008, 00:38
Huh? The iPhone/iPod Touch market will never reach the size of the Nintendo DS or PSP individually, let alone the last two combined. Not only that but those who care about gaming on their iPhone/Touch and willing to pay beyond $10 is going to be extremely small. It's a device that could have only the extreme end of casual and the games would be less than shallow for the most part. So as a "real" gaming device I don't really see the point of the discussion, it's not going to happen in any meaningful way.
Entropy
08-Sep-2008, 06:20
Huh? The iPhone/iPod Touch market will never reach the size of the Nintendo DS or PSP individually, let alone the last two combined. Not only that but those who care about gaming on their iPhone/Touch and willing to pay beyond $10 is going to be extremely small. It's a device that could have only the extreme end of casual and the games would be less than shallow for the most part.
Well, while I respect that this is your opinion, you have to accept that others have no real reason to accept it as fact. Why would people who have the solvency to pay for the phone and the associated subscription plans balk at paying for a game, if the product warrants it? I think that if you look at a typical iPhone user, they are very likely to make more money per year than the average PSP user.....
The last numbers available for the iPhone production rate (800000/week) exceed the combined for the DS and the PSP (roughly 600000/week). And that is without counting the iPod touch. It is also for the iPhone3G alone, a single model where there will likely be a family. Still not available in all countries. And always carried by a single carrier, which in many territories is a disaster in terms of sales volume, where no carrier has a strong dominance, and the consumers rightfully expect to be able to choose between carriers regardless of handset used.
The mobile phone market as a whole is roughly 20 million handsets/week. How it will grow, and how much of that is destined to be iPhone compatible, nobody knows. How attractive would a device that was software compatible with the iPhone but couldn't connect to the traditional carriers (but only via WiFi) be? Again, nobody really knows.
So as a "real" gaming device I don't really see the point of the discussion, it's not going to happen in any meaningful way.
"Real" gaming?
Skrying
08-Sep-2008, 06:36
They didn't buy the device for gaming in most cases. When someone purchases a Nintendo DS or Sony PSP the main goal (well, at least for the DS) is for it to be a gaming device. They are prepared to purchase games on a semi-regular basis at an established cost. With an iPhone user you're starting with the established market, then reducing it further for those who will purchase from the store, then further for those who will purchase games, then further for those who will purchase games within a higher cost bracket. This greatly reduces the market and reduces who the games are going to be targeted at.
I think there will games out for the iPhone but I don't think its game software shipments of games costing greater than $20 will ever be worthy of truly caring about.
"Real" gaming is when you have a regular stream of a games that have the scope of say Zelda or Final Fantasy.
roninja
08-Sep-2008, 09:44
http://www.phonesreview.co.uk/2008/09/07/apple-iphone-and-ipod-touch-to-gain-9-games-from-ea/
Enjoying your Apple iPhone or Apple iPod Touch? Like playing mobile games? Then you will be pleased to hear that EA has announced they are to bring nine new games to the Apple iPhone and iPod Touch in the near future.
Apart for the much awaited Spore Origins mobile game, Electronic Arts is also bring a juicy line up of games such as, MONOPOLY: Here & Now, The World Edition, SimCity, The Sims 3, Need for Speed Undercover, Lemonade Tycoon.
And if that’s not enough, how about, Tiger Woods PGA TOUR 09, Mahjong, EA Mini Golf, and YAHTZEE Adventures; with this line up it is obvious EA Games is serious about mobile gaming for the Apple iPhone and iPod Touch.
Not bad for starters....
Ailuros
08-Sep-2008, 11:43
They didn't buy the device for gaming in most cases. When someone purchases a Nintendo DS or Sony PSP the main goal (well, at least for the DS) is for it to be a gaming device. They are prepared to purchase games on a semi-regular basis at an established cost. With an iPhone user you're starting with the established market, then reducing it further for those who will purchase from the store, then further for those who will purchase games, then further for those who will purchase games within a higher cost bracket. This greatly reduces the market and reduces who the games are going to be targeted at.
I think there will games out for the iPhone but I don't think its game software shipments of games costing greater than $20 will ever be worthy of truly caring about.
"Real" gaming is when you have a regular stream of a games that have the scope of say Zelda or Final Fantasy.
While I tend to agree to most of it, there are still too many unknown factors to make any predictions for the future, since any license agreement rarely means SoC shipments within the next year. First and most important question mark would be here, which IHV is going to provide IP for SONY's next generation handheld. I've no idea who will but I'm as certain as I can be that NVIDIA for one doesn't have that particular deal in its pocket. It could of course also mean that they might have developed something themselves, but in the end it would be a lot easier for them to have one graphics platform both for their high end mobile phones as well as future handhelds.
I wish I knew who's the lucky one (out of the handful remaining graphics providers for handhelds/PDA/mobile devices), yet if it should be IMG after all then the entire picture might also influence other devices with probably more humble graphics performance like iPod and the likes.
Frankly I haven't tried yet to play on an iPod anything, meaning my judgement on handling and/or ergonomics for gaming is non-existant on that one. Pure handheld gaming devices have the advantage that they're a lot easier to handle for games and that's where the point for "true gaming" on small form factor devices actually starts.
However that should not mean that future Apple devices could not see an increase in interest considering games.
RudeCurve
08-Sep-2008, 13:10
IMO the primary disadvantage Apple has with the iPhone/iPod Touch as a potentially serious gaming machine is cost and user control. Without tactile controls it's gaming potential is severely limited to Nintendo touch type games and motion control. If they could add a low profile nub/thumpad and a few hard buttons it would be guaranteed to take off because the install base is alreayd there and the programming model is very simple. I could see developers/publishers making games for it just because they could sell quite a few to people who like games and who already own a iPhone/iPhod.
Personally I'm planning on getting an iPhone after my ATT/Cingular contract expires and I'm willing to pay $30 for a high quality game as I would on a DS portable depending on the game. For example if I could get a 3rd party Touch type game on the iPhone that is graphically superior to the Nintendo DS equivalent, I would get the iPhone version rather than the DS version. Of course if I wanted to play some Zelda or Mario I would need to buy a DS too but that doesn't mean I would not continue to buy graphically superior games for an iPhone.
The big loser is going to be SONY. Their software support for PSP is severely lacking and the PSP itself is just a PMP with no phone or GPS abilites out of the box. Even when PSP2 arrives (with a touchscreen) it would still be trailing the iPhone's capabilites. I think SONY is in a difficult position because A) they have to decide whether or not to keep the UMD drive for PSP2 and B) they have to decide whether or not the PSP2 should be an iPhone copycat. As a PMP and game machine the PSP is looking more unattractive as we move forward because the iPhone and iPod Touch will be doing the same things and more. The UMD movie thing has died a long time ago.
It might sound like a dumb question, but why has it to be Apple? I mean most of the interested parties know by know that Apple has used IMG IP for its iPhone, which doesn't justify the re-newed secrecy. Anyone want to take any bets on something like SONY? ;)
Well of course there is nothing concrete that points inextricably to Apple, but looking at various circumstantial evidence, I'd be extremely confident it is Apple.
Yes Apple iphone/itouch has IMG IP in it, but that came about as a result of Samsung licensing IMG IP, the royalities are paid by Samsung to IMG, there is no licensing agreement between IMG & Apple in that respect.
IMG have never publicily or even informally admitted to any relationship whatsoever with Apple, even though various sources have confirmed that MBX is referred to many times in the iphone SDK. Apple is notoriously secretive regarding its partners and technology. IMG are keen to refer to the N95, and various other Nokia and SE end products that have their IP in them via a licence agreement with their respective chip suppliers, but never utter a word about the iphone, except to cite the general "iphone effect". One would think they'd be keen to point to the iphone given its ultra high profile.
The agreement recently announced referred back to an initial licence agreement in July '07 with the same "electronic systems company". That announcement was for, at the time, IMG's "future video and graphics IP. The agreement now announced basically says that the licencee has access to pretty much eveything current and future that IMG designs. This means that IMG probably have no more legal requirement to ever make an announcement in relation to this licencee; if you have just said they are taking the rights to everything you are going to do for years to come, you never need to make another announcement again. This fits perfectly with Apples "minimal news" requirment.
finally, Apple has a known long range plan for things in mobile and "small format" that require top end graphics and video in an efficient manner. Steve Jobs is already on record in stating that their purchase of PA semi was for future iphone/ipod, and we can be sure that he was not being candid in giving us an exhaustive list of its uses. That necessitates the purchase of graphics and video IP to fit into the vaious Socs that will be designed by PAsemi.
Ailuros
08-Sep-2008, 19:13
Tangey,
I asked an honest question above; I don't doubt that it is for Apple after all. I'm merely saying that at the same time the SONY PSP successor is a huge question mark. If now SONY has somewhere a deal under wraps with IMG (mark a huge "IF" since I don't know), then all I'm saying is that a large handheld deal would indirectly help i-whatever in the foreseeable future too.
Tangey,
I asked an honest question above;
I took it as an honest question, and also thought it reasonable to generally state my reasons for the conjecture, for those who aren't completely tuned into the whole scenario.
I'd take a Sony deal too :)
iPhone gaming is an exciting market and very likely to be big business. It's caught the eye of big name VC money like Kleiner Perkins and several game publishers. Hell, I bought a Mac Book Pro and a Touch just to get my feet wet with the SDK.
I have several friends who are already developing games for it, and the exciting thing is that many of these games will be out in only a few months! Digital distribution combined with small scale casual game development -- it's a match made in heaven!
I see appleinsider is running a story that a senior Apple engineer let slip on an online CV that he was working within PAsemi on Arm based processors.
If true this would add further credence to the speculation that PAsemi is integrating ARM processors and IMG graphics IP for next gen iphone platforms.
The CV has since been removed.
http://www.appleinsider.com/articles/08/09/15/apples_pa_semi_working_on_arm_chip_for_next_gen_ip hone.html
As the originator of this thread, its nice to be able to add to it that IMG have finally been able to confirm that Apple is a licencee, via a news announcement stating that Apple have taking a stake in IMG.
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=420
This is the first time IMG have been able to formally identify Apple as a licencee. Previously, Apple was referred to as an "international electronic systems company"
Ailuros
18-Dec-2008, 12:16
Frankly at the ridiculous rate the IMG stock value has fallen to, it's barely a wonder that Apple takes the opportunity to invest in IMG.
Something else: http://www.fudzilla.com/index.php?option=com_content&task=view&id=10997&Itemid=1
Entropy
18-Dec-2008, 18:02
Frankly at the ridiculous rate the IMG stock value has fallen to, it's barely a wonder that Apple takes the opportunity to invest in IMG.
They could take the opportunity to invest in a lot of things. No, this is about future products.
Ailuros
19-Dec-2008, 10:19
They could take the opportunity to invest in a lot of things. No, this is about future products.
Of course is it also in relative terms a form of an aid to IMG for future development. In any case it isn't a coincidence either that Apple has bought shares when they were close to the lowest rate of the entire year:
http://finance.yahoo.com/echarts?s=IMG.L#chart1:symbol=img.l;range=1y;indic ator=volume;charttype=line;crosshair=on;ohlcvalues =0;logscale=on;source=undefined
Would they had bought the shares when they closed the long term licensing deal, they might have ended up paying nearly twice as much per share.
roninja
19-Dec-2008, 13:57
Astute timing these Apple folks, clearly these guys know a thing or two about making money as well as some good gear! Given the Apple folks are now shareholders along with Intel, I guess they may have even more say about IMG's future roadmap.
FWIW, I'll repeat what I said in another thread. This is what I speculate Apple's PowerVR-based roadmap looks like:
0) 90nm/ARM11/MBXLite/In-House or Samsung Audio&Video (2007/2008 iPhones)
1) 65nm/ARM11/SGX520/VXD330/In-House Audio&ISP (2009 iPhone & iPhone Nano)
2) 45nm/Cortex-A9/SGX540/VXD380/VXE280/In-House Audio&ISP (2010 iPhone)
However there is still the mystery of the higher ARM clock speed for the iPod Touch 2G (which I love BTW, it does feel smoother compared to my 1G :)) - I wonder what's up with that.
announcement stating that Apple have taking a stake in IMG
hum :arrow: http://www.appleinsider.com/articles/08/12/19/intel_ups_stake_in_imagination_following_apples_bu y_in.html
Apple is advertising games this Holiday season for the Touch and iPhone. But is it possible that they'd integrate faster GPUs for purposes other than gaming? Or for general UI effects or OS performance while getting better gaming performance in the bargain?
Haven't they said something about how in the Mac OS would use available cycles from the GPU?
I don't think Apple is going to displace Nintendo any time soon, if ever. Best they could hope is that they get the spillover from the handheld and console spaces. That is, developers could leverage titles developed for other platforms and port it for a simple-to-license, simple-to-distribute platform which has an attractive target demographic.
Ailuros
20-Dec-2008, 05:38
FWIW, I'll repeat what I said in another thread. This is what I speculate Apple's PowerVR-based roadmap looks like:
0) 90nm/ARM11/MBXLite/In-House or Samsung Audio&Video (2007/2008 iPhones)
1) 65nm/ARM11/SGX520/VXD330/In-House Audio&ISP (2009 iPhone & iPhone Nano)
2) 45nm/Cortex-A9/SGX540/VXD380/VXE280/In-House Audio&ISP (2010 iPhone)
However there is still the mystery of the higher ARM clock speed for the iPod Touch 2G (which I love BTW, it does feel smoother compared to my 1G :)) - I wonder what's up with that.
I've no idea what the iPhones of the next two years will look like, but the ballpark between MBXLite and 520 sounds too small to me and the differenc between 520 and 540 too big. Are you sure the 2009 thingy won't contain something like a 530 for instance? Merely because the 530 is "older" than the 520 and its integration has better chances.
silent_guy
20-Dec-2008, 06:34
Best they could hope is that they get the spillover from the handheld and console spaces.
They're going after a much larger group: those who'd never consider buying handheld gaming devices or console or a PC game, but don't mind spending a few dollars on an impulse buy that may or may not work out, as long as it just require a few taps on a screen.
IOW, they aiming straight at me and, boy, it is working. :wink:
Entropy
20-Dec-2008, 22:52
I find this (http://www.appleinsider.com/articles/08/12/20/future_iphones_to_wield_opencl_acceleration.html) article in Appleinsider quite interesting as well. It deals with IMG hiring OpenCL engineers, but also goes into the role of OpenGL in the future.
It introduced a viewpoint that I hadn't really considered much, but that seems quite valid. Definitely another perspective than what is typical on these forums as a whole.
Ailuros
21-Dec-2008, 07:07
I find this (http://www.appleinsider.com/articles/08/12/20/future_iphones_to_wield_opencl_acceleration.html) article in Appleinsider quite interesting as well. It deals with IMG hiring OpenCL engineers, but also goes into the role of OpenGL in the future.
It introduced a viewpoint that I hadn't really considered much, but that seems quite valid. Definitely another perspective than what is typical on these forums as a whole.
Nice finding. Apple's longterm goals were clear when they introduced OpenCL and I suspected that IMG will jump on the OpenCL bandwagon since afaik SGX was designed around GPGPU capabilities amongst others.
Eventually it aids both companies: Apple can broaden its software and hardware presence from top to bottom and in relative terms prevent any possible future domination of Microsoft's OSs in the small form factor markets and IMG can try to scale itself up step by step instead of being stuck only in mobile phones/PDAs/handhelds.
The perspective though is there for quite some time as some parts of it can be seen f.e. here:
http://www.imgtec.com/corporate/presentations/InterimDec08/index.asp?DisplayPage=26&#ViewTop
http://www.imgtec.com/corporate/presentations/InterimDec08/index.asp?DisplayPage=37&#ViewTop
Nice finding. Apple's longterm goals were clear when they introduced OpenCL and I suspected that IMG will jump on the OpenCL bandwagon since afaik SGX was designed around GPGPU capabilities amongst others.
I think you'll find IMG has been "in" on OpenCL from the start as they are one of the co-authors of the OpenCl standard...
http://www.khronos.org/news/press/releases/the_khronos_group_releases_opencl_1.0_specificatio n/
in which Tony King-Smith from IMG said
"Imagination is delighted to have been involved in the authoring of OpenCL, which we see as a significant development for the future of GP-GPU based computing for multimedia.”
I've no idea what the iPhones of the next two years will look like, but the ballpark between MBXLite and 520 sounds too small to me and the differenc between 520 and 540 too big. Are you sure the 2009 thingy won't contain something like a 530 for instance? Merely because the 530 is "older" than the 520 and its integration has better chances.Remember the MBX Lite in the current iPhone is clocked at 50MHz it seems, while a SGX 520 on 65nm would be clocked at 150-250MHz. That seems like a very reasonable performance improvement to me if Apple doesn't increase the screen solution, which I wouldn't expect them to do before 2010.
I find this article in Appleinsider quite interesting as well. It deals with IMG hiring OpenCL engineers, but also goes into the role of OpenGL in the future.It's indeed not a bad article but I can't help but roll my eyes at things like this:
Apple's strength in the iPod segment and its strong start with the iPhone are both helping build critical mass around OpenGL in the mobile development space.
[...]
Additionally, OpenCL's similarities to the OpenGL APIs will help entrench both open standards in mobile development before Microsoft's DirectX has a chance to monopolize the market. That in turn will create a mobile bulwark which will likely help marginalize the dominance of DirectX in the broader computing landscape, just as the iPod pulled the wind from the sails of Microsoft's Windows Media DRM strategy. In the game console market, DirectX on the Xbox 360 faces Nintendo and Sony, both in the OpenGL camp. Seriously guys, what the hell? I know your love Apple, that's the entire point, but that doesn't force you to love every damn thing loosely related to Apple no matter how absurd the reasons! Rest of the article isn't bad at all though, but I still think the boost from GPGPU on handhelds will remain much smaller than on desktops.
Entropy
22-Dec-2008, 13:22
One of the aspects I thought was interesting was exactly that the Microsoft virtual monopoly for PCs is still mostly undisputed (even though it leaks in the seams). It is simply gradually becoming less important in the greater scheme of things as overall interest shifts to other areas.
How important OpenCL on mobile platforms becomes will depend on a lot of things, but I think it can be fairly argued that it is likely to be more important there than on Wintel PCs equipped with increasingly capable floating point hardware as standard. For just about every floating point code I come across, the limitation already lies in the memory subsystem rather than the FP hardware. And Intel is set to provide additional capabilities in their next generation CPUs. In mobile space, performance is still an issue, and having underutilized transistors is not only a cost in terms of money, but the cost in power draw is a very serious issue. If having OpenCL has the potential to save transistors, or even entire chips elsewhere in the system, this is quite significant. Is this the case though? I honestly don't know.
It is simply gradually becoming less important in the greater scheme of things as overall interest shifts to other areas.Absolutely agreed, but AppleInsider goes quite a bit beyond that, which I find rather absurd.
How important OpenCL on mobile platforms becomes will depend on a lot of things, but I think it can be fairly argued that it is likely to be more important there than on Wintel PCs equipped with increasingly capable floating point hardware as standard.Have you look at ARM's Neon? On the A8 it's 64-bit, but on the A9 it's 128-bit FMAC single-cycle. It's not as flexible in terms of issuing instructions as the x86 CPUs, but in terms of raw GFlops throughput per MHz it's the same as Barcelona or Nehalem :!:
For just about every floating point code I come across, the limitation already lies in the memory subsystem rather than the FP hardware. And Intel is set to provide additional capabilities in their next generation CPUs.If what you said was 100% true in all cases, HPC GPGPU would never deliver speed-ups significantly above memory bandwidth differences, which it definitely sometimes does. But yeah, I agree it's a substantial factor.
and having underutilized transistors is not only a cost in terms of money, but the cost in power draw is a very serious issue.Power islands are pretty damn omnipresent nowadays, I'd argue that really isn't the biggest problem... :) What is much more problematic is having unused transistors *inside a given subsystem*, and presumably OpenCL could minimize that a bit by not forcing everything to be so damn flexible (rare workloads could just be offloaded to the GPU) but I don't think it should be overestimated. However it is interesting to ponder the possibility that Apple would do video encoding on the SGX 520 for their next-gen 65nm SoC, not on the ARM11.
If having OpenCL has the potential to save transistors, or even entire chips elsewhere in the system, this is quite significant. Is this the case though? I honestly don't know.I don't see any killer app for it right now, but I'm sure eventually somebody will find something more useful to do with it than the proposed "analyze image to tell what monument I'm looking at because I'm too lazy to look at the goddamn plate in front of it or ask the locals" usage! ;)
Entropy
22-Dec-2008, 15:49
Have you look at ARM's Neon? On the A8 it's 64-bit, but on the A9 it's 128-bit FMAC single-cycle. It's not as flexible in terms of issuing instructions as the x86 CPUs, but in terms of raw GFlops throughput per MHz it's the same as Barcelona or Nehalem :!:
Yes. I can see a lot of places where the A9 in sympathetic surroundings could be a very viable alternative to x86, when binary x86 compatibility isn't a major issue.
Off topic:
If what you said was 100% true in all cases, HPC GPGPU would never deliver speed-ups significantly above memory bandwidth differences, which it definitely sometimes does. But yeah, I agree it's a substantial factor.
Even though computational science is my branch, you'll note my almost complete absence from the GPGPU forum. It just doesn't do anything for the problems at hand. This is an area where I have reason to tread lightly, but as far as I can see high performance computing have been addressing a continously shrinking set of problems with time. It is easy to understand why, since we have moved from faster FPUs and memory, to vector, to parallell, to vector-parallell, to massively parallell to vector-massively... The number of problems that effectively map to these increasingly demanding structures is naturally shrinking.
There are important problems that DO map well enough, and of course (and in my opinion this is what drives the community these days), HPC architectures and programming is an interesting technological field in its own right. What I'm typically seeing now though is HPC folks looking for problems that map to their systems, rather than the other way around.
But most scientific/technical problems do not benefit greatly from the Big Iron, and in many cases the underlying approximations are coarse enough, or the computational load light enough, that a PC can deal with the problems so well that going further just isn't justified over the utility of doing your work on a laptop.
And I say that as part of a community that burn supercomputing cycles like there is no tomorrow.
I don't see any killer app for it right now, but I'm sure eventually somebody will find something more useful to do with it than the proposed "analyze image to tell what monument I'm looking at because I'm too lazy to look at the goddamn plate in front of it or ask the locals" usage! ;)
Indeed. In mobile space at least efficiency is still a big issue, so even modest overall efficiency gains has some value there. Whether sufficient to provide tangible benefits to the end-user remains to be seen. It seems reasonable to assume that Apple has some particular applications in mind though, and once those are revealed along with the underlying platform to support them, it will be easier to judge the (initial) value of the technology.
FWIW, I'll repeat what I said in another thread. This is what I speculate Apple's PowerVR-based roadmap looks like:
0) 90nm/ARM11/MBXLite/In-House or Samsung Audio&Video (2007/2008 iPhones)
1) 65nm/ARM11/SGX520/VXD330/In-House Audio&ISP (2009 iPhone & iPhone Nano)
2) 45nm/Cortex-A9/SGX540/VXD380/VXE280/In-House Audio&ISP (2010 iPhone)
However there is still the mystery of the higher ARM clock speed for the iPod Touch 2G (which I love BTW, it does feel smoother compared to my 1G :)) - I wonder what's up with that.
I'm not so sure that Apple could come up with a custom Cortex-A9/SGX54x for the iPhone in mid 2010. Cortex-A8 powered smartphones enter the market in H1/2009, so mid 2010 seems awfully early for Cortex-A9. And Apple acquired PA Semi in 2008, so that would only be 2 years for a custom design with new technology from start to device. It took TI from 2003 til 2009 to get the Cortex-A8 OMAP3 into smartphones with 45 engineers working on it. OMAP4 (Cortex-A9/SGX540) is not expected before the end of 2010 at the earliest for smartphones. And TI has a lot of experience and expertise in the custom ARM chip business...
I'm not so sure that Apple could come up with a custom Cortex-A9/SGX54x for the iPhone in mid 2010. Cortex-A8 powered smartphones enter the market in H1/2009, so mid 2010 seems awfully early for Cortex-A9. And Apple acquired PA Semi in 2008, so that would only be 2 years for a custom design with new technology from start to device. It took TI from 2003 til 2009 to get the Cortex-A8 OMAP3 into smartphones with 45 engineers working on it. OMAP4 (Cortex-A9/SGX540) is not expected before the end of 2010 at the earliest for smartphones. And TI has a lot of experience and expertise in the custom ARM chip business...
I forgot to add: Qualcomm's Snapdragon has been a four year, $350 million project.
Over a month ago, OMAP3430 product such as NEC's N-01A became available.
The theoretical specs of Cortex-A8 may seem less efficient than ARM11, but benchmarks of Pandora indicate an impressive improvement in real-world performance. I'd expect the further improvement Cortex-A9 brings will end up owing as much to the benefit of time and fab process as architectural design.
Ailuros
23-Dec-2008, 06:52
I think you'll find IMG has been "in" on OpenCL from the start as they are one of the co-authors of the OpenCl standard...
http://www.khronos.org/news/press/releases/the_khronos_group_releases_opencl_1.0_specificatio n/
in which Tony King-Smith from IMG said
"Imagination is delighted to have been involved in the authoring of OpenCL, which we see as a significant development for the future of GP-GPU based computing for multimedia.”
That is true, however OpenCL was Apple's baby all along. Apple had even a patent (which I've no idea if it was filed or not) under a different name before OpenCL was announced yet unfortunately my link to it is dead by now.
Arun,
Remember the MBX Lite in the current iPhone is clocked at 50MHz it seems, while a SGX 520 on 65nm would be clocked at 150-250MHz. That seems like a very reasonable performance improvement to me if Apple doesn't increase the screen solution, which I wouldn't expect them to do before 2010.
That might open the gap between MBX Lite and 520 (and I doubt the latter would go higher than 200MHz@65nm), but the gap between a 520 and 540 would still be quite large.
Anyway if Apple is as conservative with power consumption as in the initial iPhone (which makes perfect sense if you consider that if I play a demanding game on the N95 I need to search for a power plug asap heh) you might have a point, albeit I'd easier bet on a SGX530@45nm for the third attempt then.
That might open the gap between MBX Lite and 520 (and I doubt the latter would go higher than 200MHz@65nm), but the gap between a 520 and 540 would still be quite large.
Anyway if Apple is as conservative with power consumption as in the initial iPhone (which makes perfect sense if you consider that if I play a demanding game on the N95 I need to search for a power plug asap heh) you might have a point, albeit I'd easier bet on a SGX530@45nm for the third attempt then.
SGX530@45nm for the third iPhone attempt (2009)? I'm not so sure that 45nm is practical by Q3/2009. All the other 2009 smartphone parts seam to be in 65nm (OMAP3, Snapdragon etc.).
Simon F
23-Dec-2008, 10:35
That is true, however OpenCL was Apple's baby all along. Apple had even a patent (which I've no idea if it was filed or not) under a different name before OpenCL was announced yet unfortunately my link to it is dead by now.
Did you mean granted rather than filed? A patent is automatically filed the instant you present it to the patent office.
I'm not so sure that Apple could come up with a custom Cortex-A9/SGX54x for the iPhone in mid 2010. Cortex-A8 powered smartphones enter the market in H1/2009, so mid 2010 seems awfully early for Cortex-A9.Let me answer point by point... :)
And Apple acquired PA Semi in 2008, so that would only be 2 years for a custom design with new technology from start to device.Apple already had a semiconductor team in-house before PA Semi. The original iPhone SoC is not a vanilla one they bought from Samsung, but rather one they co-designed with Samsung using Samsung IP. The last page from this PDF should give you a hint of what kind of ready-made blocks are available: http://www.samsung.com/global/business/semiconductor/products/strategicfoundry/downloads/foundry_brochure_200611.pdf
The goal with PA Semi is likely to reduce their dependence on Samsung and to ARM/PowerVR IP themselves, rather than through Samsung. They're also obviously moving away from Samsung's in-house multimedia IP, although probably not from their I/O IP. So Apple isn't new to the semiconductor game unlike what some sites are claiming; they are only moving incrementally more of the process in-house.
It took TI from 2003 til 2009 to get the Cortex-A8 OMAP3 into smartphones with 45 engineers working on it.45 engineers working only on the CPU part, right? Obviously those engineers weren't active from 2003 to 2009, more like 2003 to 2006 and the numbers at the start were likely smaller... Anyway there are two big differences there:
- TI did a fair bit of (semi-)custom circuit work on the Cortex-A8 and they were the lead customer. Neither will likely be the case for Apple on 40nm; the A9 was made so that it would be viable to use normal synthesis tools to implement it.
- The time from tape-out to final product availability can be much lower for Apple because they can co-design hardware and software at the same time (something nobody else can do) and the chip they do in-house doesn't require baseband qualification etc. like for Nokia.
I forgot to add: Qualcomm's Snapdragon has been a four year, $350 million project.They simply licensed the ISA and did *everything* in-house. I don't think this is comparable at all... ;)
Yes. I can see a lot of places where the A9 in sympathetic surroundings could be a very viable alternative to x86, when binary x86 compatibility isn't a major issue.Yup, what's most interesting with NEON is that it's very much aimed at phones too though. The FPU in Qualcomm's Snapdragon is also a 128-bit FMAC unit...
Even though computational science is my branch, you'll note my almost complete absence from the GPGPU forum. It just doesn't do anything for the problems at hand.
[...]
But most scientific/technical problems do not benefit greatly from the Big Iron, and in many cases the underlying approximations are coarse enough, or the computational load light enough, that a PC can deal with the problems so well that going further just isn't justified over the utility of doing your work on a laptop.
And I say that as part of a community that burn supercomputing cycles like there is no tomorrow.Oh, I fully understand where you're coming from. Many problems are indeed simple enough that high-end HPC is kind of a fool's errand. But I still think the high-end HPC market will remain large enough that GPGPU will matter; and really, the problem you're describing also applies to handhelds, where most of the cool seemingly compute-intensive stuff really isn't as intensive as it might appear on first hand. There even more, GPGPU depends on finding a cool app to make it mainstream or at least to demonstrate a clear power or cost advantage as you said.
That might open the gap between MBX Lite and 520 (and I doubt the latter would go higher than 200MHz@65nm), but the gap between a 520 and 540 would still be quite large.Why? SGX 540 is ~4x faster per MHz, isn't it? If it's also clocked 50% faster, that means 6x faster. If you go from 320x480 to 480x720 for example, the performance boost per pixel goes down to ~2.5x... That already seems much less crazy to me! :) As for power consumption numbers, don't confuse the core 3D numbers with what's required for the rest of the SoC including the ARM11 and other systems which might have to be online due to bad system/power gating design. SGX540@45/40nm is still a very small core, and I'd assume power to be very reasonable.
SGX530@45nm for the third iPhone attempt (2009)? I'm not so sure that 45nm is practical by Q3/2009. All the other 2009 smartphone parts seam to be in 65nm (OMAP3, Snapdragon etc.).He meant third SoC not second so 2010+, however I only partially agree with you. TI SoCs coming out for the likes of Nokia in 2009 are likely to be in 45nm; except for those however, and maybe a few phones with Qualcomm's MSM7850 (CDMA Rev.A/B with ATI's OpenGL ES 2.0 3D core), everything in 2009 will indeed be 65nm.
As for TSMC's 40nm process, I don't think it's completely impossible that we would see a few personal media players or things like that in 4Q09 with that, given the shorter design cycles, but who knows and that depends more on execution that anything else...
Let me answer point by point... :)
Apple already had a semiconductor team in-house before PA Semi. The original iPhone SoC is not a vanilla one they bought from Samsung, but rather one they co-designed with Samsung using Samsung IP. The last page from this PDF should give you a hint of what kind of ready-made blocks are available: http://www.samsung.com/global/business/semiconductor/products/strategicfoundry/downloads/foundry_brochure_200611.pdf
The goal with PA Semi is likely to reduce their dependence on Samsung and to ARM/PowerVR IP themselves, rather than through Samsung. They're also obviously moving away from Samsung's in-house multimedia IP, although probably not from their I/O IP. So Apple isn't new to the semiconductor game unlike what some sites are claiming; they are only moving incrementally more of the process in-house.
Thanks for the clarification. Great post!
45 engineers working only on the CPU part, right? Obviously those engineers weren't active from 2003 to 2009, more like 2003 to 2006 and the numbers at the start were likely smaller... Anyway there are two big differences there:
- TI did a fair bit of (semi-)custom circuit work on the Cortex-A8 and they were the lead customer. Neither will likely be the case for Apple on 40nm; the A9 was made so that it would be viable to use normal synthesis tools to implement it.
- The time from tape-out to final product availability can be much lower for Apple because they can co-design hardware and software at the same time (something nobody else can do) and the chip they do in-house doesn't require baseband qualification etc. like for Nokia.
Good points. I didn't know the A9 required less hand-crafting then the A8 (OMAP3) for decent results.
They simply licensed the ISA and did *everything* in-house. I don't think this is comparable at all... ;)
Doesn't P.A. Semi have a license to do the same thing? But I guess I can read between the lines that a custom ARMv7 Apple core would not be worth the effort?
He meant third SoC not second so 2010+, however I only partially agree with you. TI SoCs coming out for the likes of Nokia in 2009 are likely to be in 45nm; except for those however, and maybe a few phones with Qualcomm's MSM7850 (CDMA Rev.A/B with ATI's OpenGL ES 2.0 3D core), everything in 2009 will indeed be 65nm.
As for TSMC's 40nm process, I don't think it's completely impossible that we would see a few personal media players or things like that in 4Q09 with that, given the shorter design cycles, but who knows and that depends more on execution that anything else...
My bad, I thought he meant third device generation as in 1. iPhone, 2. iPhone 3G, 3. iPhone 09
So after all this information I'm pretty confident that a mid 2010 iPhone with higher resolution screen and Apple designed Cortex-A9 and SGX540 in 40nm is quite possible. If Apple wants to do it :)
Apple hasn't pushed higher resolution displays in the computer business.
If the form factor remains the same, a 3-4 inch screen, what applications would push them to use higher resolution?
Apple hasn't pushed higher resolution displays in the computer business.
If the form factor remains the same, a 3-4 inch screen, what applications would push them to use higher resolution?Because if they don't everyone else will have sharper displays by 2010-2011! Smartphone market leader Nokia has at least 640x360 on a smaller screen size on all their S60 touch smartphones. Blackberrys have at least the same resolution, but a smaller screen size. Windows Mobile and Android will go for 800x480 on most devices. So Apple has to upgrade at some point. Imho most likely in 2010.
Plus everything looks better with a higher resolution. And you can definitely see the difference between 480x320 and 800x480 on a 3.5" screen.
Ailuros
23-Dec-2008, 18:23
Resolutions on mobile devices will IMHLO scale only up to a reasonable size of the entire device. The size of the shoe isn't necessarily ergonomic or practical for a mobile phone LOL.
Just saying, their laptops, which well very well, have lower resolutions than competitors.
I'd like to see higher resolution too. Main benefit is in videos and games, perhaps. Although with higher quality screens, maybe games development becomes a more expensive proposition, unless they can easily port console games. That won't help smaller developers though.
Higher resolution in the UI or for things like Safari may not be appealing to older consumers though.
Doesn't P.A. Semi have a license to do the same thing? But I guess I can read between the lines that a custom ARMv7 Apple core would not be worth the effort?I thought that was for the PPC ISA? Either way ARM seems to be willing to license the ARM ISA to whoever wants to take the plunge; quite different from a certain other ISA, if I may take the chance to point that out...
So after all this information I'm pretty confident that a mid 2010 iPhone with higher resolution screen and Apple designed Cortex-A9 and SGX540 in 40nm is quite possible. If Apple wants to do it :)Exactly, if Apple wants to do it is a very good question :) i.e. how much money they want to invest into it etc. as well as how far along they were with their 65nm SoC when they bought PA Semi (remember they licensed the PowerVR cores as early as mid-2007...) - btw last quick note: Samsung doesn't have a 40nm process per-se it seems, although in terms of marketing figures it's mid-way between TSMC's 45 and 40...
I'd like to see higher resolution too. Main benefit is in videos and games, perhaps. Although with higher quality screens, maybe games development becomes a more expensive proposition, unless they can easily port console games. That won't help smaller developers though.Yeah, resolution has also had the problem it increases the screen/backlight's power consumption. I don't have exact figures but it's definitely not negligible IIRC. In Apple's case, the big question is whether they will eventually increase the screen size to, say, 4"... If so, a larger resolution starts making even more sense.
Higher resolution in the UI or for things like Safari may not be appealing to older consumers though.How is that relevant when you can zoom in as you see fit and the pixel size of icons etc. would likely be adjusted to remain reasonably sized? I honestly don't think it should be too much of a problem, or at least I hope it won't be; it's not quite like on a personal computer.
Higher resolution in the UI or for things like Safari may not be appealing to older consumers though.
I also don't see the problem here. It's not like the fonts etc. get smaller, they just get sharper. For everyone who surfs and/or reads a lot on the iPhone a higher resolution screen would be a blessing.
Exactly, if Apple wants to do it is a very good question :) i.e. how much money they want to invest into it etc. as well as how far along they were with their 65nm SoC when they bought PA Semi (remember they licensed the PowerVR cores as early as mid-2007...) - btw last quick note: Samsung doesn't have a 40nm process per-se it seems, although in terms of marketing figures it's mid-way between TSMC's 45 and 40...
Is there some kind of exclusivity deal between Apple and Samsung till 2010 and beyond like with AT&T?
Is there some kind of exclusivity deal between Apple and Samsung till 2010 and beyond like with AT&T?I don't think there is anything like that, but since Samsung got the manufacturing license for several SGX and VXD/VXE cores, that would simply to imply Apple plans to manufacture more than one new SoC at Samsung. That would indicate they're locked in for 65 & 45nm...
Apple is going to try to keep the form factor as small and thin as possible.
These design goals would seem to clash with bigger screens, more power hungry chips and the higher battery density needed to support these niceties.
Apple is going to try to keep the form factor as small and thin as possible.
These design goals would seem to clash with bigger screens, more power hungry chips and the higher battery density needed to support these niceties.
I think the combined power consumption of the new chips and displays we discussed here would be in the same ballpark as todays chips/displays or even lower for comparable tasks. And all the other components become more power efficient/smaller too (broadband/storage/GPS etc.)
Entropy
25-Dec-2008, 12:27
Apple is going to try to keep the form factor as small and thin as possible.
These design goals would seem to clash with bigger screens, more power hungry chips and the higher battery density needed to support these niceties.
OLED.
Or how about the rollable OLEDs
But what are the costs of a 3-4 inch OLED screen compared to LCD?
prabindh
25-Dec-2008, 16:32
Let me answer point by point... :)
- The time from tape-out to final product availability can be much lower for Apple because they can co-design hardware and software at the same time (something nobody else can do) and the chip they do in-house doesn't require baseband qualification etc. like for Nokia.
Baseband qualification is for the modem, and the full system itself has to go through the wireless certification anyway, so I do not think it is a great advantage over Nokia. I think Apple has a big advantage in its browser (webkit) core and UI than anything else compared to competition.
On the SGX540, what new applications do you foresee that will drive the adoption of this 3D engine in the iphone in 2010 ?
Baseband qualification is for the modemYou do realize Nokia's custom OMAP SoCs integrate both the app processor and the modem, right? This is how TI still has such a large percentage of the baseband market despite having no 3G IP of their own. Therefore, to switch to a new application processor, they have usually been required to go through baseband qualification...
This is something I think many of these companies are getting tired of as it also requires them to make the design choices even more years in advance, which is not usually a good thing in this business to say the least. So I'm not expecting things to work that way anymore beyond 2009/2010, but I could be wrong.
On the SGX540, what new applications do you foresee that will drive the adoption of this 3D engine in the iphone in 2010 ?Have you seen the success of the games on the application store? Seems like more than enough of a justification to me :)
On the SGX540, what new applications do you foresee that will drive the adoption of this 3D engine in the iphone in 2010 ?
GAMES, a higher resolution screen, OpenCL, display games via video-out, etc. etc.
:razz:Have you seen the success of the games on the application store? Seems like more than enough of a justification to me :)
The big games cost $10 right now. $7 go directly(!) to the game developers. An overall number 10 sells ca. 3000 units a days (12/18/08), number 1 ca. 13.000 (12/22/08). That's between $630.000 and $2.730.000 a month (for something like SimCity). Or between $315.000 and $1.375.000 for a $5 game (for something like Tetris, Fieldrunners, Crash Bandicoot).
And that's just right now. The App Store is just 6 months old. And the installed user base just 1.5 years.
http://www.joelcomm.com/updated_iphone_app_sales_the_f.html
prabindh
26-Dec-2008, 05:19
GAMES, a higher resolution screen, OpenCL, display games via video-out, etc. etc.
Display a touch based game on a video out to TV..Thanks, I would stick to my console :)
I would rather like to think bigger screens would drive more productivity applications. (If only MS would make their UI better).
Also, I would think making a case for having a virtual world application makes a lot of sense, than just games, at higher resolutions and screens. That would make me go wow.
prabindh
26-Dec-2008, 05:23
So I'm not expecting things to work that way anymore beyond 2009/2010, but I could be wrong.
Have you seen the success of the games on the application store? Seems like more than enough of a justification to me :)
Precisely - if things do not work that way by 09/10, then Apple is not alone in having that advantage.
I am not so sure of hardcore games that would necessitate higher performance, coming to small screens. Maybe its just me. :)
Apple is going to try to keep the form factor as small and thin as possible.
These design goals would seem to clash with bigger screens, more power hungry chips and the higher battery density needed to support these niceties.
Does anybody know how big the MBXlite (@90nm) in the iPhone is? Just to compare it to the speculated SGX54x (@40/45nm) for the 2010 iPhone (IMHO SGX545@65nm = 12.5mm2). Thx!
Ailuros
27-Dec-2008, 06:13
I am not so sure of hardcore games that would necessitate higher performance, coming to small screens. Maybe its just me. :)
Games for mobile devices and relevant game ports won't for sure become "lighter" as time goes by, that's a given. Besides even if the hardware is supposed to be "overkill" ISVs can always invest in additional IQ improving features. It's a shame IMG didn't make a full game out of PowerVR Racer. It runs with 4xFSAA on MBX.
Does anybody know how big the MBXlite (@90nm) in the iPhone is? Just to compare it to the speculated SGX54x (@40/45nm) for the 2010 iPhone (IMHO SGX545@65nm = 12.5mm2). Thx!
No idea frankly about Lite@90nm; I'd figure though that a theoretical SGX540@40nm to be somewhere in the ~6mm2 region.
No idea frankly about Lite@90nm; I'd figure though that a theoretical SGX540@40nm to be somewhere in the ~6mm2 region.
If Apple goes with the proposed SGX520@65nm (2.6mm2) for the iPhone 09 than that would mean double the GPU size for 2010. The SGX520 is really really small.
Entropy
27-Dec-2008, 10:28
But what are the costs of a 3-4 inch OLED screen compared to LCD?
I don't think it's possible to say. Not only are Apples volumes for their iPhone and iPod products such that they would get very good deals, but in this case production facilities would be created for the purpose. Apple would effectively create the market. There has recently been production facilities set up for 3.5" AMOLED production (unfortunately digitimes reserves its article database for subscribers), but it's impossible to say if the volumes targeted could cover iPhone volumes. The wording doesn't suggest it.
OLED would be a great technology choice for phones though. Thin, good energy/legibility properties, extremely fast response time for games, films and camera viewfinder function. Life time is good enough to definitely be a non-issue for phones, so cost is the major question mark. But even going by existing products using OLED, it shouldn't be much of an issue for an iPhone. We aren't talking about $15 retail gadgets here, and the added value to the device would definitely be greater than the added cost. Plus, it would be a competitive advantage for Apple vs. its smaller volume rivals.
Ailuros
27-Dec-2008, 11:33
If Apple goes with the proposed SGX520@65nm (2.6mm2) for the iPhone 09 than that would mean double the GPU size for 2010. The SGX520 is really really small.
Exactly the reason why I have my doubts for Arun's proposed (speculative) roadmap. One aspect though I haven't thought of so far would be Apple targeting also some sort of competition to future handhelds, which I wouldn't expect earlier than say at least 2010 anyway.
I agree with Entropy, also both Samsung and TMDisplay seem to be doing a very good job of preparing to ramp AMOLED in the 2009-2010 timeframe; TMDisplay's 2.2" screen with only 100mW of power consumption total would be incredibly attractive for a RAZR-like phone. In fact, I have a small hunch we'll see that in 2H09...
As for SGX 540: you're massively overestimating its size. PowerVR publicly claimed it's substantially less than 2xSGX530, while that is only about 5.5mm² for <=150MHz on TI's not-very-dense 65nm process. SGX520 would likely be 3.5mm² on such a process IMO (remember 2.6 is also pre-layout iirc). So my guess is SGX540 would be maybe 9mm² on that same process, which might translate into 4mm² on Samsung's 45nm process unless you aimed it at a substantially higher clock speed, which is possible.
As for MBX Lite, it's really small; since MBX Non-Lite was 6.5mm² on TI's 90nm process, maybe MBX Lite would be ~3mm² on Samsung's 90nm process? I'm really just guessing there though.
I'm hoping for breakthroughs with OLED in big screens.
I know OLED is used in mobile/portable devices.
What about touch-screen capabilities?
Going by the 'less than 2 to 8 mm^2' range @90nm Imgtec used to quote for the MBX family, ~3 mm^2 does sound right for the iPhone's MBX Lite + VGP Lite.
Ailuros
27-Dec-2008, 21:12
As for SGX 540: you're massively overestimating its size. PowerVR publicly claimed it's substantially less than 2xSGX530, while that is only about 5.5mm² for <=150MHz on TI's not-very-dense 65nm process. SGX520 would likely be 3.5mm² on such a process IMO (remember 2.6 is also pre-layout iirc). So my guess is SGX540 would be maybe 9mm² on that same process, which might translate into 4mm² on Samsung's 45nm process unless you aimed it at a substantially higher clock speed, which is possible.
Their own SGX whitepaper states 12.5mm^2@65nm for the SGX545.
3.5 square millimeters for just 2 TMUs is a tad too much for your 540 theory don't you think? I was thinking originally in the ~11mm2 region for 540@65nm and made a rough estimate for a non perfect shrink to be within any worst case scenario.
I agree with Entropy, also both Samsung and TMDisplay seem to be doing a very good job of preparing to ramp AMOLED in the 2009-2010 timeframe; TMDisplay's 2.2" screen with only 100mW of power consumption total would be incredibly attractive for a RAZR-like phone. In fact, I have a small hunch we'll see that in 2H09...
I don't know about OLEDs in mobile phones. Indoors they are great but outdoors in the sunlight they are much worse than good transflective TFTs. I guess the popular Nokia N85 with it's 2.6" OLED screen is the mass market leader for OLEDs in mobile phones. I had the chance to compare one with my Nokia E-Series transflective Display in a bright autumn sunlight and you almost couldn't see anything on the N85, while the E-Series was totally usable. And I guess the sunlight wasn't very bright compared to a Spain, Mexico, California etc. summer.
I know that OLEDs have many advantages, but for people in sunny countries I wouldn't be so thrilled about having one in my mobile phone.
Similar experience, even in the dark and rainy UK:
http://www.allaboutsymbian.com/features/item/The_Nokia_N85_and_OLED_screens_Not_perhaps_the_Hol y_Grail.php
As for SGX 540: you're massively overestimating its size. PowerVR publicly claimed it's substantially less than 2xSGX530, while that is only about 5.5mm² for <=150MHz on TI's not-very-dense 65nm process. SGX520 would likely be 3.5mm² on such a process IMO (remember 2.6 is also pre-layout iirc). So my guess is SGX540 would be maybe 9mm² on that same process, which might translate into 4mm² on Samsung's 45nm process unless you aimed it at a substantially higher clock speed, which is possible.
4mm2@45nm (Samsung) vs. 6mm2@40nm (TSMC). Quite a difference you got there. Only time will tell :smile:
mboeller
29-Dec-2008, 13:19
At the moment I think the best new feature for a small gaming system and/or phone could be a combination of the unipixel screen with a "nano-touch" system.
http://www.unipixel.com/
http://www.unipixel.com/assets/unipixel_whitepaper_20070717.pdf
http://www.patrickbaudisch.com/ (the picture says it all)
http://research.microsoft.com/en-us/um/people/baudisch/projects/nanotouch/
I don't know about OLEDs in mobile phones. Indoors they are great but outdoors in the sunlight they are much worse than good transflective TFTs.I think that's partially model-dependent. Look at Page 6 of this presentation: http://www.picommissioncoree2008.net/files/multimedia_card_player_samsung_sdi.pdf
SIA is a digital algorithm that increases the effective contrast and does result in a "washing out" effect with fewer shades, but it's better than nothing; and a good AMOLED is definitely much better than a transmissive LCD at least, although at 100K+ Lux I doubt even that and SIA can save you...
At the moment I think the best new feature for a small gaming system and/or phone could be a combination of the unipixel screen with a "nano-touch" system.
http://www.unipixel.com/
http://www.unipixel.com/assets/unipixel_whitepaper_20070717.pdf
http://www.patrickbaudisch.com/ (the picture says it all)
http://research.microsoft.com/en-us/um/people/baudisch/projects/nanotouch/
Long term, maybe. But apart from OLEDs, I can see other display technologies that are imho more likely to enter the touchscreen(smartphone) mass market in the next 2 years :
Better haptic feedback for touchscreens:
1. Real localized haptic feedback like Nokia's Haptikos (real life prototype was already shown over a year ago).
"Don’t be fooled by simple vibrational imitations folks, this is the real McCoy – you press a key on the screen, and it clicks under your finger with exactly the same sort of fingertip feedback as if you’d pressed a conventional keyboard key."
http://www.redferret.net/?p=9533
2. Real static tactile feedback (actually feeling the keys/deformed surface). No actual hardware was shown, yet (imho). Apple patent. Most likely further into the future than Haptikos.
http://www.unwiredview.com/2007/10/25/apples-touch-surface-keyboard-with-tactile-feedback/
Touchscreens with better multi-touch and more:
Sharp's new LCD screen with embedded optical sensors in each pixel. Supports multi-touch input and scanning. Only half as thick as the iPhone's touchscreen, prototype had the same 480x320 resolution as the iPhone. Mass production was announced for mid 2008. AU Optronics (AUO) and Toshiba Matsushita Displays (TMD) have already shown similar products, too.
It would provide more flexible multi-touch (more than two inputs, works with gloves etc.) and scanning (documents, business cards, could eventually be used for fingerprint recognition etc.)
http://www.electronista.com/articles/07/08/31/sharp.multi.touch.proto/
http://www.vnunet.com/vnunet/news/2198082/sharp-multitouch-scanner-lcd
http://www.engadget.com/2007/10/02/sharp-shows-off-multi-touch-optical-scanning-portable-lcd/
I think that's partially model-dependent. Look at Page 6 of this presentation: http://www.picommissioncoree2008.net/files/multimedia_card_player_samsung_sdi.pdf
SIA is a digital algorithm that increases the effective contrast and does result in a "washing out" effect with fewer shades, but it's better than nothing; and a good AMOLED is definitely much better than a transmissive LCD at least, although at 100K+ Lux I doubt even that and SIA can save you...
Since OLEDs can't "reflect" and use the sunlight, they can only try to become much brighter on their own or use "cheap" tricks like these. But transmissive screens were once totally unusable outdoors (my Casio Cassiopeia PDA was unreadable), now they are just bad. So there's hope :)
Edit (because of the OLED):
Was just leaked: Samsung S8300 with 2.8" AMOLED WQVGA touchscreen, 8 MP camera, slider, just 12.8mm thick, made of duralumin, release march 09
http://dailymobile.se/2008/12/29/extra-samsung-s8300-amoled-touch-screen-slider-with-8-megapixel-camera-to-tit-stores/
So we're definitely getting there. From Nokia's N85 with 2.6" QVGA non-touchscreen to 2.8" WQVGA touchscreen. 2009, the year OLED took over (the smart/featurephones). Also, OLED touchscreen, front facing camera, 8 MP camera, dual-LED flash, all neatly packed in just ~11mm (display part of the slider). So no excuses for the iPhone 09 in terms of camera and flash ;)
Is the BlackBerry Storm using haptic feedback for their click screen? Seems the reaction to it is mixed.
We'll see if any games try to use it.
Still think for most phone makers, gaming seems a sidelight so gaming won't drive design. Rather, game developers will work around whichever design sells well.
Is the BlackBerry Storm using haptic feedback for their click screen? Seems the reaction to it is mixed.
We'll see if any games try to use it.
Still think for most phone makers, gaming seems a sidelight so gaming won't drive design. Rather, game developers will work around whichever design sells well.
It depends, do you consider a button depressing to be haptic feedback. That's what the Storm's screen is, a button. You feel that you've clicked something, but it's not like a real keyboard where you can feel the button before it's pressed.
I like it, but there's still a lot of room for innovation in this area.
Is the BlackBerry Storm using haptic feedback for their click screen? Seems the reaction to it is mixed.
Definitely not the same thing.
More info to Haptikos:
http://www.unwiredview.com/2008/07/08/nokia-haptikos-tactile-touchscreen-details-emerge/
Patent:
http://www.unwiredview.com/wp-content/uploads/2008/07/nokia-haptics.pdf
I just realized I made a huge mistake and it makes no sense that SGX 520 would be the first SGX core used by Apple. Oops! :D This is an Imagination Technologies PR I stumbled upon from October 2008, it's by far the most revealing of all PRs from IMG about this:
Imagination's partners for POWERVR SGX include Intel, NEC, Renesas, Samsung, Sigma Designs and Texas Instruments as well as a major OEM. POWERVR SGX535 and SGX530 are sampling or in volume production. SGX540 is licensed to more than five partners, and has taped out this quarter in multiple SoCs, while SGX531 and SGX520 are in design with several licensees.This data has bunch of consequences, including:
- No tape-outs for SGX520/531 yet.
- There likely aren't any licensees for SGX 545.
- There are 5 partners for SGX 540, which is a huge number.
- [2;5[ SGX 540 tape-outs in Q4, in time for MWC/CTIA announcements.
It'd certainly be a very interesting development if Apple was one of partners, which also leaves open the question of what video core they licensed (VXD 380? please? be still my beating heart...) - of course, this doesn't exclude the possibility they just didn't tape-out anything yet and are lagging behind badly. Who knows.
Ailuros
03-Jan-2009, 11:44
I just realized I made a huge mistake and it makes no sense that SGX 520 would be the first SGX core used by Apple. Oops! :D This is an Imagination Technologies PR I stumbled upon from October 2008, it's by far the most revealing of all PRs from IMG about this:
This data has bunch of consequences, including:
- No tape-outs for SGX520/531 yet.
- There likely aren't any licensees for SGX 545.
- There are 5 partners for SGX 540, which is a huge number.
- [2;5[ SGX 540 tape-outs in Q4, in time for MWC/CTIA announcements.
It'd certainly be a very interesting development if Apple was one of partners, which also leaves open the question of what video core they licensed (VXD 380? please? be still my beating heart...) - of course, this doesn't exclude the possibility they just didn't tape-out anything yet and are lagging behind badly. Who knows.
Hate me but I said that 53x has a huge leadtime ahead of 520. In any case I'd guess that any partner would want to use anything 5x0 for something that's rather in the mobile/PDA direction and anything 5x5 rather something that is in the =/>netbook direction with XP or Vista accordingly.
SGX530 sounds quite reasonable to me as a 2nd generation iPhone chip and since I figure that Apple doesn't intend to not touch the =/>netbook markets too later on, there's always headroom for 545 or something even more powerful in the less foreseeable future.
Here again the relevant link: http://www.fudzilla.com/index.php?option=com_content&task=view&id=11177&Itemid=1
Hate me but I said that 53x has a huge leadtime ahead of 520.Yeah, you're right. SGX 530 is now the most likely possibility, SGX 540 second, and SGX 520 third.
In any case I'd guess that any partner would want to use anything 5x0 for something that's rather in the mobile/PDA direction and anything 5x5 rather something that is in the =/>netbook direction with XP or Vista accordingly. Moorestown is very much aimed at MIDs, so I wonder if they could be using SGX 540 instead of SGX 535. I guess it depends on whether the former can support DX9, and my guess is that it nearly certainly can. Hmmm.
Here again the relevant link: http://www.fudzilla.com/index.php?option=com_content&task=view&id=11177&Itemid=1We'll see what happens, I still personally believe 7-9" is in No Man's Land (i.e. neither pocketable nor very large) except for the ultra-cheap netbook market which I doubt Apple cares about.
Ailuros
03-Jan-2009, 19:04
We'll see what happens, I still personally believe 7-9" is in No Man's Land (i.e. neither pocketable nor very large) except for the ultra-cheap netbook market which I doubt Apple cares about.
Some sort of modest "trial run"?
Tablet PCs haven't sold very well, but Apple' cult Messiah Steve Jobs implied Apple was watching small-device categories like tablets and Netbooks to see if they actually take off.
We'll see what happens, I still personally believe 7-9" is in No Man's Land (i.e. neither pocketable nor very large) except for the ultra-cheap netbook market which I doubt Apple cares about.
I think it will work as a more powerful iPod Touch, but not as a netbook. 7" touchscreen, just 10mm thick, upgrade the iPod Touch internal specs 50%-100%, tweak the iPhone OS and allow multitasking but stay backwards compatible. Would be a great living room device and video player/ebook reader/gaming machine/surf tablet. I got an extra iPod Touch for that (and more) right now in my house, it's great but too small.
So how does the TI OMAP chips compare to IMG?
Observers at the Palm Pre demo today seem to think that the Pre scrolls smoother and the interface is snappier than the iPhone.
Of course, by the time the Pre ships, there could be a newer iPhone model.
Assuming its an Omap3 with 3D acceleration, then compares well with IMG, as the 3D is an SGX solution.
So how does the TI OMAP chips compare to IMG?
Observers at the Palm Pre demo today seem to think that the Pre scrolls smoother and the interface is snappier than the iPhone.
Of course, by the time the Pre ships, there could be a newer iPhone model.
TI OMAP3 use SGX530 the iPhone use powervr MBX Lite + VGP Lite + FPU
sgx530 is a newer core than MBX Lite and performs is greater TI grade powervr sgx530 at 11 MPolys/sec
it is like this
http://en.wikipedia.org/wiki/Powervr
With KYRO 3 (2D/3D AIB) products shelved due to STMicro closing its graphics division, PowerVR concentrated on the portable market with its next design, the low power PowerVR MBX. It, and its SGX successors, have become the de facto standards for mobile 3D, having been licensed by seven of the top ten semiconductor manufacturers including Intel, Texas Instruments, Samsung, NEC, NXP Semiconductors, Freescale, Renesas, and Sunplus, and in use in many high-end cellphones including the Apple iPhone, Nokia N95, Sony Ericsson P1, and Motorola RIZR Z8.
There are two variants: MBX and MBX Lite, both have the exact same features, except that MBX is optimised for speed and MBX Lite is optimised for low power consumption
vs
PowerVR SGX & SGXxt (pixel, vertex, and geometry shader hardware)
* next generation fully programmable universal scalable shader architecture
* exceeding requirements of OpenGL 2.0 and up to DirectX 10.1 Shader Model 4.1
* licensed to Apple Inc, Sony, Intel, Renesas, NEC, TI, NXP Semiconductors, Samsung, Sigma Designs, SigmaTel, Sirf and others
* 8 variants announced:
o SGX 510(canned). SGX520 (7 MPolys/sec), and SGX530/1 (14 MPolys/sec) for the handheld mobile market
o SGX535 and SGX540 (28 MPolys/sec) for handheld high end mobile, portable, MID, UMPC, consumer, and automotive devices
o SGX540 (1000Mpix/s, 20-35MPolys/s),SGX543(35M poly/s), SGX545, SGX555
i was i hurry :grin:
I think we can fairly assume that the iPhone09 will have at least a SGX530 like the Palm Pre (H1/2009) too. SGX531 or SGX535 would be nice but i guess 09 is too early for them.
Ailuros
09-Jan-2009, 06:10
I think we can fairly assume that the iPhone09 will have at least a SGX530 like the Palm Pre (H1/2009) too. SGX531 or SGX535 would be nice but i guess 09 is too early for them.
I'm not so sure the next iPhone would really need the added bandwidth or texel fillrate of either 531 or 535, unless Apple is willing to go for a lot bigger screen on that one (sounds more like a 2010 target to me).
As an example if you'd want to spot a difference in real time between f.e. a N95 and a iPhone, you'd have to use something like Quake3 at least and enable antialiasing. It might not be the best example yet software/games lack that much behind for the time being, that you'd really need corner cases to spot differences.
Point being that with a 530 real time performance rises over 2x times over MBX Lite. At 320*240 it sounds good enough to me.
After the Apple conference call and the emphasis on unchanged (iPhone) specs and on advantages of a fixed software platform (non-moving target for the developers; the app store will be just a year old), do you think it's possible that this years iPhone won't have any updated CPU and GPU? I mean no new architecture or big frequency jumps. Let's say iPod Touch 2G CPU frequency and of course 65nm CPU and GPU. But that's it for these chips.
Apart from this we would "only" get e.g. better GPS, Wireless, HSUPA, a lot better camera, front facing camera, more storage, same screen size/resolution, better accelerometers plus MEMS gyroscopes etc. Maybe more RAM for the new OS3.0 multitasking (just wishing), but each app's RAM will be artificially limited so that it can still run on a singletasking iPhone/iPhone 3G.
Wishmaster
27-Jan-2009, 19:14
I think that this year won't be a year of apple. Probably new iPhone will get released in 2010 and as usual they will shock the world with the specs and great marketing.
IMO this year will be the year of WM :D
All those new upcoming hardware bases, wm 6.5 and the commitment from MS about working closer with hw manufacturers should bring us some truly exciting new products at this year MWC and later.
IMO this year will be the year of WM :D
All those new upcoming hardware bases, wm 6.5 and the commitment from MS about working closer with hw manufacturers should bring us some truly exciting new products at this year MWC and later.
Maybe, but if this is the year of WM with WM6.5, then what about WM7 next year? And WM6.5 (afaik Q3/09) faces strong competition this summer. A New Palm, second gen Android phones, S60 Touch (maybe not in the US, but worldwide) and imho a new iPhone. But WM6.5 could be good, Microsoft certainly spend enough R&D money in it's E&D department. That has to be worth something :lol:
Wishmaster
27-Jan-2009, 20:35
Maybe, but if this is the year of WM with WM6.5, then what about WM7 next year? And WM6.5 (afaik Q3/09) faces strong competition this summer. A New Palm, second gen Android phones, S60 Touch (maybe not in the US, but worldwide) and imho a new iPhone. But WM6.5 could be good, Microsoft certainly spend enough R&D money in it's E&D department. That has to be worth something :lol:
Palm is nice and eye candy(even more than iphone) but I'm worried that it won't be as successful as iPhone was.
Android is great and it gets even better with every update. Unfortunately I think that it is still too limited with that linux java combination for programming.
S60 touch is nice but still it looks exactly the same as normal version. Even now after introduction of n97 most people are looking at palm pre and what microsoft will bring at MWC.
Probably we'll see windows mobile 7 introduced along with windows 7 desktop and on devices at 1Q of 2010 at least that's what I think.
But I'm cool with wm6.5 if it improves core efficiency of wm6.1 and drastically improves GUI. If not skymarket should give us place from which we could download GUI replacements and some eye candy programs.
People have been touting WM and HTC ever since the iPhone was unveiled. Mostly geeks who like HW specs. but don't understand that people who aren't geeks comprise a large portion of the market.
That said, there's a lot of room for improvement in the iPhone HW but Apple isn't going to boost specs. for the sake of boosting specs. They'd boost the specs. to enable new applications or features in the OS to help set it apart from the competition, even as others like Palm improve their UI.
But they like the thin form factor, even if they have to sacrifice battery life. So that may limit things like video chats (if there are even mobile networks which can support that affordably) or high-end gaming.
They may also need to find an answer for netbooks, before their MacBooks, which did really well the past quarter, are impacted.
Entropy
28-Jan-2009, 12:56
People have been touting WM and HTC ever since the iPhone was unveiled. Mostly geeks who like HW specs. but don't understand that people who aren't geeks comprise a large portion of the market.
That said, there's a lot of room for improvement in the iPhone HW but Apple isn't going to boost specs. for the sake of boosting specs. They'd boost the specs. to enable new applications or features in the OS to help set it apart from the competition, even as others like Palm improve their UI.
I feel that wco81 has good points here. Let's be reasonable - Apple didn't buy Nat Semi, recruit key employees and strike long term deals with IP suppliers all for nothing. They intend to use their capabilities.
However, they are very unlikely to use those capabilities to please spechunters or megahertz braggers, that's just not a valid target demographic. (Indeed, it's people you may be happier without entirely.)
Applications are the key here, and Apple has indicated that games is one genre of applications that it considers important. What other applications can benefit from more processing power? One that seems pretty obvious is that the iPhones/iPods need to be able to scale HD content, and output to external screens. They already connect to the iTunes store, they already have 16 to 32 GB of storage, so it won't be long before the devices can act as portable media centers, obsoleting stationary hardware for all but the biggest video libraries. And even then it complements such collections with portability. 3D-mapping/drive-through is another application requireing some processing capabilities.
What applications with reasonably wide appeal do you see for portable devices, that require more processing power than the iPhone already offers?
What applications with reasonably wide appeal do you see for portable devices, that require more processing power than the iPhone already offers?
Web browsing? ;) I mean, the iPhone is a great little device but you don't have to go look for new applications to find something that would benefit from faster hardware. Even on WiFi the iPhone/Safari is noticeably slower at rendering web pages than the average PC. Reduced app startup times would also be great – these are probably mostly limited by NAND speed though.
Ailuros
28-Jan-2009, 16:03
I'm personally buffled how can anyone even have the patience to browse on such small devices, especially if you'd want a page that hasn't been optimized for mobiles. In such cases performance is the least of my concerns, primary being that the screen size does hardly help for proper browsing.
I'm personally buffled how can anyone even have the patience to browse on such small devices, especially if you'd want a page that hasn't been optimized for mobiles. In such cases performance is the least of my concerns, primary being that the screen size does hardly help for proper browsing.
I like browsing on the iPhone, I do it all the time (mostly on the go), but I try to avoid non-optimized sites.
But browsing could definitely be improved on the iPhone: faster rendering, faster window/tab switching, higher resolution, full screen mode, more stable and last a bigger screen size (max. 4"). For the first three more powerful hardware would be useful.
I don't think a bigger screen is likely. You don't want some unwieldy form factor.
They could improve the screens, use things like OLED which would enhance battery life.
I'm not altogether sure that display HD content, either on a portable-sized screen or outputting to a bigger screen, is in the cards.
For one thing, their content partners may not want to see iTunes video content eat into content on other forms of media. Could be why they require some DRM chip to hook up iPods to TVs these days.
Would be good to know how many people worldwide commute to work on public transport where they could make use of video. Outside of a couple of big cities in the US, all Americans drive so it wouldn't seem to be a big potential market in the US (even though they've obviously bought iPods with video playback in good numbers, iTunes video downloads aren't exactly setting the world on fire).
According to a recent RBC report the next iPhone will come out in July and it will have an improved graphics processor, 3.2MP camera (video recording), a 32GB model, 3G HSUPA and will be 0.1" thinner.
http://www.appleinsider.com/articles/09/03/30/rbc_says_iphone_3g_pro_wont_boost_iphone_momentum. html
I think this is one of the more believable predictions. The CPU is not mentioned but my money is on a slightly higher clocked 65nm ARM11. The improved graphics processor is most likely a 65nm IMG SGX530/531. Although the next iPhone is supposed to do video recording, I don't think we will see HD recording (like OmniaHD) in this generation, most likely just VGA (maybe 720x480 like LG). And IMHO no IMG VXE, because AFAIK Samsung still has no manufacturing license.
I'm personally buffled how can anyone even have the patience to browse on such small devices, especially if you'd want a page that hasn't been optimized for mobiles. In such cases performance is the least of my concerns, primary being that the screen size does hardly help for proper browsing.
Works surprisingly well on my girlfriends HTC Touch HD. Zoomed out overview of entire webpage at 480x800 and easy navigation for zoom/panning makes it pretty snappy at browsing. The biggest problem for handhelds is still input IMO.
Cheers
I think this is one of the more believable predictions. The CPU is not mentioned but my money is on a slightly higher clocked 65nm ARM11. The improved graphics processor is most likely a 65nm IMG SGX530/531. Although the next iPhone is supposed to do video recording, I don't think we will see HD recording (like OmniaHD) in this generation, most likely just VGA (maybe 720x480 like LG). And IMHO no IMG VXE, because AFAIK Samsung still has no manufacturing license.
I think you are right, I havn't seen any VXE license for Samsung.
With regard to the processor, I see Samsung have rather QUIETLY released a Cortex A8 App Processor , the S5PC100. Doing a google returns a grand total of 10 hits. There has been no press release for the product that I can find, and none of the tech websites have discussed it. It is not listed AT ALL under the app processor section of their website:-
http://www.samsung.com/global/business/semiconductor/products/mobilesoc/Products_ApplicationProcessor.html
However there is a brochure available:-
http://www.samsung.com/global/business/semiconductor/support/brochures/downloads/systemlsi/s5pc100_brochure_200902.pdf
Its clocked at 833Mhz, the brochure states 720P playback and record, and a 2D/3D graphics block is shown. There is no other mention whatsoever about the capabilities of the 2D/3D block.
Arm have started to use this processor as a comparision point for Moorestown:-
http://media.corporate-ir.net/media_files/irol/19/197211/presentations/ARMDevCon2008.pdf
See page 24.
I find that highly interesting, as of course Intel had at one time targetted the next gen iphone as a killer product for moorestown.
I hear also that there is another variant called the S5PC110, with gets a grand total of 1 hits when googled. Note that samsung never published ANY data on the MBX-ed processor used in the iphone. I think that if the processor in the next gen is samsung, this could be the likely candidate, a customised version with SGX graphics core. The one caveat is that there is no hint or information as to how long this processor has been available. The samsung brochure is dated Feb '09, the ARM .PDF however is dated Oct '08 which references chip size and both operational and standby power times.
roninja
01-Apr-2009, 15:10
Whether it is this processor or a customised version it does confirm the fact that Samsung now have an A8 in silicon. I bet we will find a variant in the next iPhone (can't see Apple being underpowered compared with the Pre for example).
Oh yeah, I saw the S5PC100 being demoed at MWC, although Samsung's stand didn't have quite as much glitter as TI's or NVIDIA's. It also didn't have all that many details, although there was a roadmap in there mentioning 1080p next year... I asked about the 3D core, and while the person wasn't incredibly clear (mostly because of language issues, we both missed a few words of what the other was saying), it seemed pretty clear that this was 100% in-house R&D and unrelated to SGX or Mali... In fact I already talked about that in the context of another SoC here: http://forum.beyond3d.com/showthread.php?t=47776
Most impressive part of this chip for me: 720p decode *and* encode apparently supports H.264 High Profile. I do suspect their memory bandwidth and therefore power consumption numbers won't be competitive with the likes of OMAP4 and Tegra, though...
---
Either way this has *NOTHING* to do with Apple. I don't know how many times I'll have to repeat this, but Samsung is just a manufacturing partner which actively helps in the design process. Apple has its own semiconductor team, and their involvement this generation will be higher than last time. And even then the rumours that the same SoC was used by another company were pure lies. For the ARM core, they might use a pre-synthetized version done by Samsung and so the clock speeds here would be interesting but that's about it really. I'm still expecting ARM11 either way TBH.
FWIW, I do expect that RBC report to be incorrect. They are known for being wrong and too pessimistic about Apple products - they clearly are launching rumours to either make money themselves or help their clients doing so. I'm hoping for at least a 65nm SoC with ARM11, SGX 530, and VXD 370. That's assuming there will be a single new model though, and I wouldn't be surprised if there were two... (one probably with the old-gen SoC and a smaller screen?)
roninja
01-Apr-2009, 16:06
Arun, you think they'll stick with an ARM 11 and not an A8?
Arun,
I understand what you are saying, but I my understanding is that processor for the current iphone was totally designed by Samsung. Apple did not have any relationship at all with IMG, no license of any sort. The only people that had the license was samsung. This license was originally used to incorporate MBX into an ARM9 core.
http://www.samsung.com/global/business/semiconductor/productInfo.do?fmly_id=229&partnum=S3C2460
I don't think that core was ever taken up by anyone, it was then re-used by samsung into their ARM11, So samsung had to do the bulk, if not all, of the design work for that App Pro.
I'm not tuned into whether Apple has a previous license agreement with ARM, but as I recall the ARM archeitectural (sp!!) license that was announced for an unnamed company was assumed to be Apple, and that came out towards the end of last year. If that is the first license Apple has with ARM, the timeframe is all wrong for any fruits(!) from that to be forthcoming.
It is not entirely out of the question therefore, that Samsung will still do the lions share of the design work for this years iphone products, even acknowledging the fact that Apple have taken various licenses since lasttime around. This would leave the former PA semi people to work on the next next gen stuff.
I understand what you are saying, but I my understanding is that processor for the current iphone was totally designed by Samsung. Apple did not have any relationship at all with IMG, no license of any sort. The only people that had the license was samsung. This license was originally used to incorporate MBX into an ARM9 core.Correct, Apple did not have a relationship with the IP guys in the timeframe for the original iPhone and most of the design work was done by Samsung, especially for all the third-party IP. The way this was done, AFAIK, is via Samsung's foundry business which had 'ready' versions of various IP blocks: http://www.samsung.com/us/business/semiconductor/news/downloads/FoundryBrochure0507.pdf
This was then combined with a small amount of internal design work inside Apple. My guess is that this generation, Apple is doing 100% of the RTL in-house while the synthesis is shared between Apple and Samsung; for example, Samsung might do the ARM core and the I/O and Apple everything else).
roninja: I don't really know, but I expect Apple to remain fairly cost conscious (A8 is far from free) and I can't see them liking NEON much from a functionality point of view either. From a performance POV, obviously an A8 would be beneficial though so here's hoping even if I don't like that core so much...
Ailuros
02-Apr-2009, 11:26
Allow me to drive the conversation a tad away from the above and nearer to the possibilities of the second half of the thread title:
In addition to the healthy backlog level carried forward into the second half, we have concluded a significant license with a major consumer electronics company in November as well as a number of smaller deals. We are currently progressing a number of significant licensing agreements with existing and new partners which we are looking to close during this second half. We are also seeing an active pipeline of interest in our technologies for future years. Whilst we expect licensing revenue to continue to make progress across the calendar year as a whole, as always it remains difficult to predict the precise timing of the closure of licensing deals.
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=450
Their preliminary results (for the past year) will be presented on the 30th of April.
Ailuros,
I started this thread on the back of a license announcement made by IMG in Sept '08, which I identified as being Apple. Basically that license said that Apple has now a license on a very broad range of current and future IMG IP. given Apples secrecy, the broad nature of this announcement means that it diminishes the likeilhood of any future Apple-related license announcements having to be made.
For completeness here is a link to the original license announcement.
http://www.imgtec.com/News/Release/index.asp?NewsID=392
Whether it is this processor or a customised version it does confirm the fact that Samsung now have an A8 in silicon. I bet we will find a variant in the next iPhone (can't see Apple being underpowered compared with the Pre for example).
I really think they will stick with ARM11 for a while (Tegra is ARM11 too, no one seems to mind that) and SGX53x for this year's iPhone and ignore the A8 completely and go straight to the in-house designed multi-core A9 and SGX54x (or ARMv7 if they have the architecture license) in the following iPhone SoC (2010/11). Apart from possible die shrinks I really don't see them changing their SoC every year. I think they go the Intel route, alternating new model and refresh from year to year, coinciding with the new model then refresh cycle we've seen them do before with the iPhone/3G and countless iPods.
Laurent06
03-Apr-2009, 07:27
and ignore the A8 completely and go straight to the in-house designed multi-core A9 and SGX54x (or ARMv7 if they have the architecture license) in the following iPhone SoC (2010/11).
Assuming they have an architecture license which they would have acquired about a year ago at the same time they bought PAsemi, I would be extremely surprised if they could have an ARMv7 CPU ready for production by 2010/2011.
Assuming they have an architecture license which they would have acquired about a year ago at the same time they bought PAsemi, I would be extremely surprised if they could have an ARMv7 CPU ready for production by 2010/2011.
I guess you're assuming a Snapdragon like redesign. But Apple could just change some small A8/A9 part (for whatever reason, maybe save 0.1mm2) and that would make it a non-A8/A9 and an ARMv7 CPU ;)
Laurent06
03-Apr-2009, 13:52
I guess you're assuming a Snapdragon like redesign. But Apple could just change some small A8/A9 part (for whatever reason, maybe save 0.1mm2) and that would make it a non-A8/A9 and an ARMv7 CPU ;)
I'm not sure an architectural licensee can do that. Unless perhaps it also buys the A8/A9 design license with a specific clause (but I've never heard of such a thing, which doesn't mean it's impossible, just that no one claimed that).
Anyway, all of that is speculation so we can let our imagination wander, can't we? :grin:
Okay, sorry guys - new data made me change my mind. iPhone coming out in June is very likely based on the same SoC as the iPod Touch 2G (different one as the iPhone 3G though, remember! just look at the firmware hacking websites...) - but I also suspect the Touch 2G SoC already supports video encoding, FWIW. The only possible changes would be in terms of connectivity chips, baseband, and the camera. Royalty revenue for IMG should be similar.
DigiTimes is hinting (http://www.digitimes.com/news/a20090403PB200.html) at the existence of a higher-end model later this year with a 5MP camera, which I presume to be the one with SGX/VXD. I suspect the 3.2MP sensor they are talking about is http://www.ovt.com/uploads/parts/OV3640_PB(1.01)_web.pdf and the 5MP one is http://www.ovt.com/uploads/parts/OV5642_PB(1.0)_web.pdf - it's worth noting both of those have integrated ISPs...
One thing that disturbs me is that a phone shipping in July 2009 won't be able to use Infineon's 65nm baseband. So either they're sticking to the exact same baseband as in the iPhone 3G (and still won't support HSUPA because of the RF chip) or they're changing supplier. InterDigital is out of the running, so the only real options I can see for a moderately slim baseband are ST-Ericsson, Qualcomm, and Icera. I can't see them switching to Qualcomm, and I'd be (positively, mind you!) surprised by Icera. So I can only see ST-Ericsson left, which would be intriguing but I'm still slightly skeptical; sticking to the old-gen Infineon solution seems more likely to me.
As for WiFi, CSR's 65nm chip wouldn't be ready either (would be for the high-end model later though) so I can only see the same Broadcom solution as in the iPod Touch 2G again. Alternatively, maybe Marvell's new 150Mbps 802.11n chip for embedded - certainly would explain the "faster internetz!!!" rumours :)
EDIT: I also obviously apologize for how many times I've had to change my predictions - I think everyone will agree Apple isn't the easiest company to track though... ;)
It would be very surprising if they intro'd new iPhone more than once this year, other than a minor SKU change like more storage.
They've been pretty disciplined with iPod refreshes at once a year, just before the Holiday quarter.
Maybe releasing new iPhone models in the middle of the year doesn't line up with SOC releases? It seems RIMM and others release new models several times a year including the late fall. So far, Apple hasn't yielded to suggestions about branching or diversifying their iPhone product lines.
If they stay with one product line differentiated only by storage options, then there is only one opportunity per year to do major architectural changes.
What might accelerate their architectural revisions is if RIMM, Nokia and others gain share by pushing new models out throughout the year. It's one thing to dictate the pace of change when the market share is over 70% as in the case with iPods.
But smart phones are a more hotly-contested market so they may try to go more to a computer-like model (for example, they release new MacBooks, some major revs, some minor speedbumps, usually twice a year). On the other hand, they can't easily count on repeat sales during the year to the installed base since phone customers are locked-in. Plus more than one release per year may antagonize customers or worse, make them delay purchases until the latter half of the year.
Ailuros
04-Apr-2009, 06:31
EDIT: I also obviously apologize for how many times I've had to change my predictions - I think everyone will agree Apple isn't the easiest company to track though... ;)
That's true; Apple keeps its plans so close to its heart that it's close to impossible to know anything about them unless you're a true insider.
Ailuros
04-Apr-2009, 06:39
Ailuros,
I started this thread on the back of a license announcement made by IMG in Sept '08, which I identified as being Apple. Basically that license said that Apple has now a license on a very broad range of current and future IMG IP. given Apples secrecy, the broad nature of this announcement means that it diminishes the likeilhood of any future Apple-related license announcements having to be made.
For completeness here is a link to the original license announcement.
http://www.imgtec.com/News/Release/index.asp?NewsID=392
My own highlights in that quote were targeting "new partners" for one and on the other hand "We are also seeing an active pipeline of interest in our technologies for future years.". The latter doesn't sound to me like just one partner or more specifically in this case "just" Apple.
You are right of course about the unlikeliness of Apple-related license announcements, but the Multi-core Expo must have triggered some added interest.
Okay, sorry guys - new data made me change my mind. iPhone coming out in June is very likely based on the same SoC as the iPod Touch 2G (different one as the iPhone 3G though, remember! just look at the firmware hacking websites...) - but I also suspect the Touch 2G SoC already supports video encoding, FWIW. The only possible changes would be in terms of connectivity chips, baseband, and the camera. Royalty revenue for IMG should be similar.
DigiTimes is hinting (http://www.digitimes.com/news/a20090403PB200.html) at the existence of a higher-end model later this year with a 5MP camera, which I presume to be the one with SGX/VXD. I suspect the 3.2MP sensor they are talking about is http://www.ovt.com/uploads/parts/OV3640_PB(1.01)_web.pdf and the 5MP one is http://www.ovt.com/uploads/parts/OV5642_PB(1.0)_web.pdf - it's worth noting both of those have integrated ISPs...
One thing that disturbs me is that a phone shipping in July 2009 won't be able to use Infineon's 65nm baseband. So either they're sticking to the exact same baseband as in the iPhone 3G (and still won't support HSUPA because of the RF chip) or they're changing supplier. InterDigital is out of the running, so the only real options I can see for a moderately slim baseband are ST-Ericsson, Qualcomm, and Icera. I can't see them switching to Qualcomm, and I'd be (positively, mind you!) surprised by Icera. So I can only see ST-Ericsson left, which would be intriguing but I'm still slightly skeptical; sticking to the old-gen Infineon solution seems more likely to me.
As for WiFi, CSR's 65nm chip wouldn't be ready either (would be for the high-end model later though) so I can only see the same Broadcom solution as in the iPod Touch 2G again. Alternatively, maybe Marvell's new 150Mbps 802.11n chip for embedded - certainly would explain the "faster internetz!!!" rumours :)
EDIT: I also obviously apologize for how many times I've had to change my predictions - I think everyone will agree Apple isn't the easiest company to track though... ;)
Hm, a two model summer/Christmas normal/pro iPhone strategy? Interesting, I didn't think Apple would go down this route. But good for Palm if this summer's iPhone offers only a storage/camera upgrade.
Not sure it's fair to say 'just storage/camera'. First remember that SoC runs at ~600MHz instead of ~400MHz, and as I said I suspect it supports video encoding. As for the camera it's not just more pixels but also much better signal processing. The WiFi chip might (or might not) go to 1x1 802.11n. Finally, if the rumours are true it should support HSUPA (which makes me curious about the baseband) and be thinner. Oh, and there OS 3.0 too...
As for releasing twice a year - remember in terms of unit share, Apple is still fairly small; they have a lot of room for growth if they get more aggressive... Think about it this way: many people have been expecting a Nano. But in reality, there's nothing about the iPhone that makes it so expensive in the long-term; screen and baseband costs go down naturally. So why not reduce its price to $99 for the minimum capacity version and create a new product line starting at $299 for example? Hardly a better way to improve your line-up, IMO.
Not sure it's fair to say 'just storage/camera'. First remember that SoC runs at ~600MHz instead of ~400MHz, and as I said I suspect it supports video encoding. As for the camera it's not just more pixels but also much better signal processing. The WiFi chip might (or might not) go to 1x1 802.11n. Finally, if the rumours are true it should support HSUPA (which makes me curious about the baseband) and be thinner. Oh, and there OS 3.0 too...
Wasn't it 412MHz vs 532MHz and isn't the iPod Touch SoC still at 90nm?
Of course the new iPhone will be better than the 3G, but now it sounds like the difference will be at the same scale as the iPhone-iPhone3G update.
As for releasing twice a year - remember in terms of unit share, Apple is still fairly small; they have a lot of room for growth if they get more aggressive... Think about it this way: many people have been expecting a Nano. But in reality, there's nothing about the iPhone that makes it so expensive in the long-term; screen and baseband costs go down naturally. So why not reduce its price to $99 for the minimum capacity version and create a new product line starting at $299 for example? Hardly a better way to improve your line-up, IMO.
If this summer iPhone becomes the new "low-end" model starting at $99 and Apple introduces a Pro model with big improvements in the winter then I'm looking forward to it. Because a Pro version would give Apple the chance to change/upgrade something fundamental like the screen resolution/ratio, buttons or introduce multitasking etc. This "Pro" version of the iPhone OS could then also be used for the rumored iTablet :)
Yeah, you're right - it's 532MHz. Oops. As for it still being 90nm, I don't really know; my understanding is none of the teardown guys bothered to check, assuming it was the exact same SoC. Yay?
Anyhow it looks like it's the iPod3,1 that will have 802.11n via the BCM4329 (http://www.appleinsider.com/articles/09/04/04/next_ipod_touch_iphone_to_support_low_power_802_11 n_mode.htm) - so it's not clear that the iPhone2,1 announced probably in June will have that change too (but it might), and it looks like the Marvell and CSR SoCs are out of the running.
Does 802.11n make sense?
How does the power consumption compare, especially if running at 5 Ghz?
And most public hotspots are still at 802.11g no?
802.11n single-stream (i.e. without MIMO) makes perfect sense. I had a very good conversation with someone from CSR at MWC, and came out very impressed (both about CSR and 802.11n). What I say here should mostly also apply to Broadcom product, but I can't be certain about that.
Theoretical speed only goes up slightly, but real-world peak goes from ~20 to ~50Mbps and in bad signal conditions you can still expect a reliable ~30Mbps, which is more than good enough for most 1080p video streaming (not for the best non-transcoded Blurays of course). That is something you absolutely cannot expect to do on 802.11g. Also, having 11g devices on your 11n network is going to hurt everybody so you want all devices to migrate rapidly.
As for power consumption, it should be very good. Regarding hotspots, I guess patience is a virtue... :)
I guess.
As it is, any bandwidth-intensive application is best-used with Wifi, not mobile data.
The mobile data roadmap sounds rosy with LTE and such.
But I bet the reality won't be so great and carriers will price faster mobile data in such a way as to discourage intensive use. Probably premium for at least some types of video. Because they would probably need to upgrade not just the base stations but also the backhaul.
Or maybe these companies developing such capable mobile silicon have more faith in carriers than warranted.
Yes, 3G/4G's marketing is largely a scam, and unlimited data plans are not sustainable unless you'd try to add a premium for QoS guarantees. Consider this: Verizon only got 20MHz of spectrum depth at 700MHZ; that means the 100Mbps+ of LTE throughput has to be shared by all cell users - and that's before you consider the huge wastes in bad signal conditions.
Even worse, the spectral efficiency gains from HSUPA to HSPA+/LTE come mostly from MIMO and 64QAM, which are only of large benefit in good signal conditions; in bad ones, Rx Diversity is practically equivalent. If you want better real-world throughput you neef many smaller cells, but that is difficult when they need to be maintained by the operator. Femtocells aren't a solution beyond your own usage at this point.
LTE will be good enough for VGA video streaming and basic 720p, and the lower latencies will be nice for web browsing etc., but don't get your hopes up too much IMO... You won't be giving up on your wireline and WiFi anytime soon.
Hmm, and Verizon was the biggest winner of the 700Mhz auctions.
They were crowing about rolling out LTE early and getting service out to some rural areas which they've never served before.
Maybe if all these carriers built one network together and shared it instead of redundantly cherry-pick the most profitable metros, ending up with several networks layered on top of each other, with no single network having the best coverage.
But that would be too efficient and not sufficiently proprietary.
I don't know why users are so big on streaming or downloading over the air as opposed to sideloading. You're going to have to hook the thing up to charge it (and in the case of the iPhone and products with even more powerful chips, you're going to have to charge often) anyways and the data transfer time is less than the charge time.
A sideloaded video could probably be at much higher bitrate than a streamed one. That is where better video silicon makes sense.
Yeah. Remember that's just the new spectrum though; down the road they can switch 2G/3G spectrum to LTE. But the costs will remain high, and I agree you won't get away from sideloafing for anything above a few Mbps... And as you say charging makes that attractive anywsy if the user realizes what's going on.
Ceetainly online games with low latency/high bandwidth (even 1Mbps+) requirements are going to be a breeze on LTE, which is nice for 3D... :) Surely that should help the GPU attach rates, although I'll be very curious to see what the killer app is like and how the controls work.
Anyway down the road (i.e. for LTE Advanced or the like) I don't see how we'll get away from user-installed super-femtos with a range of up to 1Km perhaps. The only question is how does the wireline bandwidth come in; FTTH is still expensive, but you need at least 100Mbps for this to make sense. And what about spectrum? If a 'carrier' needs to buy it, where does it make the money to amortize it? Fun stuff!
I've seen the first indication that iphone 3.0 firmware supports OpenGL es2.0. If true this would mean that its designed to support a platform that has something other than an MBX 3D graphics accelerator, presumably that would be SGX.
"The final 3.0 firmware is said to support OpenGL ES v2.0"
http://www.boygeniusreport.com/2009/04/07/exclusive-apple-iphone-30-screenshots-leak-out/comment-page-3/#comments
Apple have recently hired Bob Drebin (nothing to Frank !), who's graphics C.V. looks rather impressive.
CTO Graphics product group, AMD
Engineering fellow, ATI
Architecture, ArtX...designed graphics processor for Gamecube.
Chief Engineer, SGI
http://www.linkedin.com/pub/dir/bob/drebin
Heres the news story
http://www.electronista.com/articles/09/04/27/apple.hires.ati.cto/
Another piece in the PA Semi / Arm / SGX Soc jigsaw ?
I see the mainstream media has picked up on the recently hiring the Bob Drebin and imminent arrival of another former ATI/AMD high end guy, and forming an entire chip division around them. Citing sources the WSJ said it was as much about security of their IP as anything else. Imagination gets a mention in the registers take on the story.
http://www.theregister.co.uk/2009/04/30/iphone_soc/
http://online.wsj.com/article/SB124104666426570729.html
...and now today its revealed that a director of the Xbox team is moving to Apple, what is apple planning ???
"Microsoft's Xbox strategy boss Richard Teversham has quit the company after 15 years to assume a new role in Apple's European offices related to education. "
http://www.appleinsider.com/articles/09/04/30/apple_lands_xbox_boss_mac_os_x_10_5_7_999_macbook_ air_deal.html
I've seen the first indication that iphone 3.0 firmware supports OpenGL es2.0. If true this would mean that its designed to support a platform that has something other than an MBX 3D graphics accelerator, presumably that would be SGX.
"The final 3.0 firmware is said to support OpenGL ES v2.0"
http://www.boygeniusreport.com/2009/04/07/exclusive-apple-iphone-30-screenshots-leak-out/comment-page-3/#comments
Really ?
Not in their SDK .... the only change they did was to include stencil buffer support.
Really ?
Not in their SDK .... the only change they did was to include stencil buffer support.
Ummm isn't the 3.0 SDK beta supposed to be under an NDA ?
My posting cited BGR as a source, which was talking about the features that had been discovered by scanning 3.0 firmware, not documented feautres of the new SDK. Clearly if they want to play their cards close to their chest, Apple would not expose the new hardware features of an alledged new iphone in an SDK beta. I havn't seen anyone mentioning anything about video record/playback support in the 3.0SDK, and yet that is being very widely rumoured to be in a june iphone.
Or of course BGR could be totally wrong.
mark9white
02-May-2009, 03:09
I havn't seen anyone mentioning anything about video record/playback support in the 3.0SDK, and yet that is being very widely rumoured to be in a june iphone.
http://www.macrumors.com/2009/04/07/iphone-video-recording-interface-digital-compass-voice-control-and-auto-focus-camera/
http://www.macrumors.com/2009/04/07/iphone-video-recording-interface-digital-compass-voice-control-and-auto-focus-camera/
You merely prove my point, its been found in FIRMWARE 3.0 beta by investigation of the rom files, it has not been found/discovered/mentioned in the 3.0 SDK beta.
I don't think MBX has acceleration for stencils while SGX obviously does.
Simon F
07-May-2009, 09:05
I don't think MBX has acceleration for stencils
That is correct. It made OpenVG interesting. :-)
[EDIT: I should just qualify the OVG comment - it meant I got to develop a triangulation algorithm for "arbitrary" polygons]
Ailuros
07-May-2009, 10:41
That is correct. It made OpenVG interesting. :-)
OT: I know I shouldn't ask you, but why isn't there a VGX factsheet yet available?
Thought I'd resurrect this thread, as its the most pertinent to the subject matter.
With the usual madness surrounding the impending announcement next week of the apple "ithingy" ( which I suppose will technically be a handheld), does anyone have any concrete thoughts as to the graphics.
We know that Apple is an IMG shareholder and using SGX535 in the 3GS and latest itouches, but what do we exepct for the ithingy ?
The options would appear to be
1) Clock the existing iphone Soc quicker, giving a boost to both the CPU and graphics. Mixed with a possible reduction to 45nM this could result in a performance increase and a power reduction.
2) A new Soc with SGX545....545 would give roughly a x2-x3 graphics improvement alone without any clock change.
3) A New Soc with 543MP. This seems very unlikely to me as IMG only started to talk about it for the first time a year ago, and formally announced the IP in March last year.
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=449
And only started to ship finished IP to lead partners last month
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=497
Surely that makes it too early for chips.. even if they don't need to be delivered until April for a June product launch.
4) A Tegra solution... Nvidia have been talking up itablet type devices lately.
Something else entirely ?
OT: I know I shouldn't ask you, but why isn't there a VGX factsheet yet available?
Since the thread's been resurrected, there is now a factsheet for VGX available to Insider members.
Some speculation that it will be using the first product of the PA Semi acquisition.
If the device is a new form factor of a Mac and/or needed the full profile of OpenCL, a 545 would make sense. Otherwise, a 540 should be the only other alternative to the current 535.
Captain Chickenpants
21-Jan-2010, 09:52
Rys, I didn't realise you were one of us! When did you join?
If the device is a new form factor of a Mac and/or needed the full profile of OpenCL, a 545 would make sense. Otherwise, a 540 should be the only other alternative to the current 535.
With Apple being one of the leading lights of OpenCL, it might be a matter of policy to incorporate compliant cores.
Ailuros
21-Jan-2010, 10:26
Since the thread's been resurrected, there is now a factsheet for VGX available to Insider members.
It's been officially on their homepage for quite some time now. But since I'm in a pestering mood today: the October Insider newsletter got me all riled up waiting for the next one (read through it if you haven't already...). Waiting for the next edition to continue where the October issue has left is worse than watching 500 episodes of a soap opera and at the last 501 episode to have a power cut :sad:. It's January already....$%#@@@&^ (<ancient greek curse).
Captain Chickenpants,
I guess being slow is contageous inside IMG *snicker* :lol:
If the device is a new form factor of a Mac and/or needed the full profile of OpenCL, a 545 would make sense. Otherwise, a 540 should be the only other alternative to the current 535.
Why are you excluding the 2MP SGX543 option? Considering the increase in die area for the GPU between MBX Lite and SGX535 should be more than sizeable I don't see why. Depends really what Apple is targetting for and that's something unfortunately only Apple knows at the time.
Why are you excluding the 2MP SGX543 option? Considering the increase in die area for the GPU between MBX Lite and SGX535 should be more than sizeable I don't see why. Depends really what Apple is targetting for and that's something unfortunately only Apple knows at the time.
IMG stated in December past that they were only starting to ship production ready 543MP IP at the end of that month. Is it feasible to go from production ready IP to an Soc shipping in production quantities in 3 months (for March/April).
I see over on eetimes an analyst is quoted as saying:-
System processor: ''The key change in the component stack will be the application processor. Our checks indicate that Apple licensed the ARM Cortex core from Samsung but built out the rest of the silicon with the PA Semi design team. While Samsung will continue to supply the flash memory, it will be relegated to a foundry status for the application processor. We expect this silicon to populate future iterations of the iPhone.''
Nothing startling there regarding PA semi and Samsung but If the above is true (processor for itablet will be used in later iphones) then itablet is definitely IMG graphics. However surely whatever would be used in the itablet would be overkill in a phone ? (unless a slower clocked version was used in the phone).
Ailuros
21-Jan-2010, 11:21
I wouldn't say that the 3GS 535 is that much over 100MHz; if you'd compare that to (nearly) 4x times the frequency under 45nm for upcoming netbooks, there's quite a difference in raw fillrate between those too 535 variants.
Assuming Apple hasn't gone shopping elsewhere (which we'll find out only after products have been released as usual) and they're only using IMG graphics IP it comes down to what functionalities they want for each device (SGX545 obviously means the highest available featureset but at an obvious cost of perf/mm2 compared to other SGX variants), power consumption/die area for each device, if they want to have two different application processors for different markets or one and for the last only a frequency/fillrate difference (for higher resolutions-screen sizes) will come into play. But I'm just stating the obvious here.
545 by the way has additional GPGPU functionalities (due to it's full embedded GPGPU profile), but if those functionalities aren't needed, there's a lot more floating point power IMO you can squeeze out at the same frequency even from a single 543 than from a 545.
It's been officially on their homepage for quite some time now. But since I'm in a pestering mood today:
Freely available without an Insider login? I can't find it if so, although it wouldn't be outwith my abilities to not even find something on our own website :lol:
Your Greek curses have no effect here, Kristof has cast many spells around the office using his ducks and general wizardry....
Rys, I didn't realise you were one of us! When did you join?
Hello! Not that long ago actually, but I've found my feet now I think :runaway:
Ailuros
21-Jan-2010, 14:15
Freely available without an Insider login? I can't find it if so, although it wouldn't be outwith my abilities to not even find something on our own website :lol:
Your Greek curses have no effect here, Kristof has cast many spells around the office using his ducks and general wizardry....
*ahem* http://www.imgtec.com/powervr/VGX150.asp
Send the bastige my greetings I haven't seen him in a long time. As for my feline username I swear to Bastet that there's no pun intended in that pic:
http://users.otenet.gr/~ailuros/018.jpg
*ahem* http://www.imgtec.com/powervr/VGX150.asp
Send the bastige my greetings I haven't seen him in a long time. As for my feline username I swear to Bastet that there's no pun intended in that pic:
Yeah, but the factsheet linked from that page needs a login :wink:
As for the pic....I have a new desktop background now :lol:
In a 45/40nm 2011 product, I fully expect Apple to employ the 543MP in a dual, tri, or quad configuration.
According to Beyond3D's own archie4oz, Apple went all out to clock the 535 at 150 MHz in the 3G S (with the original iPhone's MBX Lite at the 103 MHz speed of the system bus as first speculated).
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.