[ergem]

they're expected increase in profit margin might come from the matured ps suite business.. i believe kaz hirai was quoted stating he expects the PSN that's including the ps suite platform to be a 3-billion dollar business in the future..

I also heard of a similar ps suite platform for their tv which will be hardware nuetral as well as playable on the ps4..

ps suite android + ps vita = touchscreen casual gaming with optional traditional button support
hdtv google tv + ps4 = camera-based casual gaming with optional move support..

my point is, ps suite has the potential to be a cash cow for sony in the future..

[Rangers]

Hasn't PS Suite been pretty lackluster? Seems like a big leap of confidence to see it churning huge profits in 2-3 years.

[IllusionistK]

Xbox Live doesn't even rake in $3bln/year.

Perhaps by PS suite they mean Playstation as a whole(Hardware+Software)

[ergem]

It is yet to be released, it was only recently released for open beta. Though I agree Sony must execute faster than the current rate, but the potential is still there.

Sony has the games to challenge the popular titles on smartphones with their own casual games like patapon, loco roco and echochrome, and of course their thousands of ps1 back catalog.

edit: let find the quote i'm not too sure now..

edit2: i find it hard to quote the article, i'm only using my ipad.. but i found one article regarding the psn purportedly to gain 3 billion sales.. http://news.softpedia.com/news/Kaz-H...d-174951.shtml

[Rangers]

Kaz saying it means little.

For the purposes of this discussion, imo it's a decent indicator they're not going to go big time loss leading on PS4 hardware. Occams razor. I think there's some wiggle room there for others to argue against this reasoning, but it's still pretty likely.

Not that I even dont think you cant build a nice, pitcairn class machine at a profit. Also, Sony projecting future profits is a fools errand anyway, they have been projecting a big turnaround every single year.

[Shifty Geezer]

Originally Posted by IllusionistK

Xbox Live doesn't even rake in $3bln/year.

XBox Live is limited to the number of enabled devices.

Perhaps by PS suite they mean Playstation as a whole(Hardware+Software)

PSSuite is a software platform bringing 'PlayStation' to non-PlayStation hardware. If PSS sees wide adoption on Android phones (big if!), that's currently 250 million (?) Android devices and growing, plus who knows how many Google TVs. This forecast reckons 1 billion Android devices in five years. Sony would be wanting maybe $10 per Android device average to make their wished-for turnover. Perhaps $20-30 for higher end devices only.

[TheWretched]

I don't see it that way... I think all options are open, depending on your viewpoint. Since the Playstation stuff is one of the few businesses at Sony that actually do make a profit, lowballing it might not be in Sonys best interest. On the other hand, pulling a Wii might be even better for Sony... I dunno.

[N2O]

Originally Posted by Shifty Geezer

XBox Live is limited to the number of enabled devices.

PSSuite is a software platform bringing 'PlayStation' to non-PlayStation hardware. If PSS sees wide adoption on Android phones (big if!), that's currently 250 million (?) Android devices and growing, plus who knows how many Google TVs. This forecast reckons 1 billion Android devices in five years. Sony would be wanting maybe $10 per Android device average to make their wished-for turnover. Perhaps $20-30 for higher end devices only.

^you know many non-Xbox hardware will get Xbox Live later this year right?actually it's already in many non-Xbox hardware if you include mobile phone

[3dilettante]

There were several areas of overreach with the PS3's development that probably won't happen with the PS4. At least some of what bled Sony during the early parts of this generation of consoles will have no need to be repeated.

The Blu-ray roll out isn't happening twice, and that was a chunk of the early problems and costs.

Cell was a technological curiousity that Sony and Toshiba (IBM not as much) saw filtering into a wide variety of electronic products, not just the PS3. This was part of the justification for the engineering effort, and a probably source for some of the design parameters from Sony and Toshiba that have turned out to be miscues.

That's probably not going to happen again either. As a product, Cell has gotten its ass handed to it everywhere not inside a PS3.
There were some probable overextensions in expanding production capability that Sony and Toshiba went through for this chip that added to costs.
If the rumors are correct, Sony is going with the paradigm that beat it handily last time.

The packaging and integration research is something that might be budgeted as a more global project. Unlike Cell, this is not a technical direction everybody else is avoiding, and it can save money in manufacturing from mobile to PC.

[hoho]

Originally Posted by N2O

^you know many non-Xbox hardware will get Xbox Live later this year right?actually it's already in many non-Xbox hardware if you include mobile phone

Do you know how big of a player is WP on phone market? I'd be surprised if it doesn't die in next couple of years.

[McHuj]

Originally Posted by TheWretched

I don't see it that way... I think all options are open, depending on your viewpoint. Since the Playstation stuff is one of the few businesses at Sony that actually do make a profit, lowballing it might not be in Sonys best interest. On the other hand, pulling a Wii might be even better for Sony... I dunno.

The problem for Sony is that it's hard to pull off a Wii. Yes, the Wii was cheap hardware, but that wasn't the only reason (or a big one imo) for it's success. It had Wii Sports and an new control mechanism that people loved. Does anyone think that Nintendo would have had the same success if the Wii didn't have motion controls and was just an upgrade to the Gamecube? I don't.

If Sony doesn't provide an uber console that makes the core gamers drool and they decide to "Wii-it", they have to provide something else like the Wii did. Unfortunately, I don't see that capability at Sony to innovate that. Bundling Move or a stripped down Vita controller isn't going to cut it. I want the uber console from Sony.

[bgassassin]

Originally Posted by Acert93

I personally offended a number of MS folks last generation with my signature and lengthy rants about how there is no such thing as a MS rumors, only MS leaks. Surprisingly no one has PM'd me this generation yet with hot leads like last time

I think someone was a little upset that people were passing the Xbox 360 block diagram around 2 years before the launch. But it made for sooo much fun. The various, unsubstantiated, contradictory rumors + fear that all the console makers are turning out Wii-rehashes isn't as fun as, "OMGWTFBBQ Cell broadband 1TFLOPs on 90nm before Intel!!!111 Free Cybord inside."

At least this time we have developers like Crytek, DICE, Epic, etc speaking up quite vocally about wanting a lot of cores (dozens even), beefy 2TFLOPs+ GPUs, 8GB of memory, etc.

But I take that as a bad sign: They probably were hearing whispers of 4 core, 1 TFLOPs GPUs, and 2GB of memory and were disappointed.

So I can blame you for all this "trouble" with MS then.

I think those vocal devs were just being greedy and wanted more beyond what were targeted. The way it's starting to sound Sony and MS were in a staring contest with them. MS blinked (recent rumors that they are going "above and beyond" a normal jump) and Sony didn't even flinch. Early on I had heard it was possible Xbox 3 would be more powerful, but that was before IGN's 6670 rumor muddied the waters. Though I still trusted the person I heard that from over IGN's rumor.

Originally Posted by Rangers

But I'm afraid BG's cred is going to take a tumble when it's inevitably confirmed Wii U is PS360 level only...

The way this GPU is sounding, that may end up being a push.

Originally Posted by Mianca

My bet is on a Vishera derived CPU + Pitcairn successor + 4GB of unified RAM for the PSNext. It will incorporate some APU-ish functionality so they can basically turn the thing into a SoC later in the lifecycle.

4GB is seeming very unlikely IMO with PS4. Xbox 3 may be the only with 4GB. And if they do that without a density increase, I look forward to seeing how they pull that off.

[liolio]

Originally Posted by Acert93

An AMD 6850 was released 10/2010 on the 40nm process node. Retailing at $179 it is a 255mm^2 chip (127W max TDP) with a 256bit bus to 1GB of GDDR5 memory.

I am not sure I would agree that $110 for BOM is "pricy" as a video card includes * the GPU chip * Memory * PCB with PCIe bus * various smaller chips * fan and casing * Ports for DVI, HDMI, mini display port, etc. A "console" is going to need to enlarge a few of those features (e.g. PCB, fans) but you are mainly adding a CPU, industrial design, optical drive, possibly a HDD, networking (WiFi, Bluetooth, etc), USB ports, etc. An expensive CPU would be a mistake and if that GPU setup is expensive (which if you look at what was shipping when the 360 shipped it surely is not) it really doesn't require much of a CPU. And the rest is affordable/anticipated costs.

Also consider that as 28nm ramps up and high end production moves to 20nm and memory densities improve it is hard to imaging that a GPU that was released in October 2010 at $179 could not see SIGNIFICANT cost reductions 3 years later in October 2013. In fact the 7850 (which is already seeing its price drop at retail from the MSRP of $249) has 2GB of GDDR5 and based on densities and the smaller die size (255 vs 212) it is likely as 28nm matures it will be cheaper than the 6850 was at launch.

I use this one because there was no estimate for the 7xxx series.\
It was the closest. Bart is bigger whereas 7850 are tinier on more expansive wafer.
Same RAM amount (a bit slower for the 6850) still there must not be that much a difference between both cards. AS you pointed out RAM (gddr5) is pretty costly if anything and considering the statement from Nv and GF as well as slow launch for AMD part I would say that the 7850 is a bit more expansive.
May come even when 28nm is master (for the chip alone). I suspect that early series cost a lot though.

[Shifty Geezer]

Originally Posted by N2O

^you know many non-Xbox hardware will get Xbox Live later this year right?actually it's already in many non-Xbox hardware if you include mobile phone

Sure. What has that got to do with Live not making 3 billion a year now?

[N2O]

Originally Posted by hoho

Do you know how big of a player is WP on phone market? I'd be surprised if it doesn't die in next couple of years.

you know iOS is big right?
also,Windows 8

Originally Posted by Shifty Geezer

Sure. What has that got to do with Live not making 3 billion a year now?

because if PS suite can do it,XBL can do it either(no offense,just guess)

[ergem]

Originally Posted by Rangers

Kaz saying it means little.

of course.. that's just projection on his part. But surely that would amount to the big chunk increase we see on this chart regarding Sony's "projected" profit margin.

Whether hirai's expectation which we now can see reflected on that chart shall come to frution will depend on a lot of ifs. But it doesn't change the fact that Hirai has been quoted to believe that he sees the PSN to grow larger in the near future. It should be logical to believe that that previous prediction is being reflected on this chart.

[tunafish]

Originally Posted by N2O

you know iOS is big right?
also,Windows 8

iOS and android are the only smartphone platforms that are actually growing. All the new releases have managed to the WP7 is reduce the rate at which it's share is shrinking. It's still losing users, not just share.

esr plots the comscore numbers (US-only, collected from the carriers, reliable numbers on the total amount of smartphones connected to the networks, as opposed to the amount of smartphones sold) nicely, and the mobile sector is rapidly turning into a two-horse race.

http://www.catb.org/esr/comscore/

[Heinrich4]

Originally Posted by liolio

The big advantage I see for Sony with a SoC is cost.
Especially if they were to go with pretty much off the shelve Kaveri.
They already reduce (supposedly) the R&D by going with AMD which provide complete solution CPU+GPU.
Prior APU was not tempting but kaveri is another matter. By the way I'm impress that they plan to have such a jump in GPU power and still keeping the TDP at 100W.

I can't answer you on the benefits of an APU vs a two discrete part of the same power. I've read here and there (like pretty much every body can) that low latency communication between the CPU and the GPU could have positive effects on perfs but I can't tell more. For all I know it could be misunderstanding from my side of theirs.
I mean there is priority in job submitted to the GPU but as I see a GPU it could be a while before you get the result. Still if they were way to save the round trip to memory (GPU writing to it and then the CPU when informed than data are there as to read for the RAM...) but will Kaveri do that? (will the GPU L2 be coherent with the cores ones? I would say no).
Intel seems do be able to pull that to the benefit of their driver team (... don't laugh ).

There are benefits in power consumption you avoid to duplicate the memory controller for example. But it's not free you have to share bandwidth.
I believe that in the PC world AMD efforts are hold back by the memory type, APU are bandwidth starved.
You have only one chip to test etc. it makes sense.

How much consume an Athlon II X4? I would say quiet a lot, Let say they have a TDP of 65 Watts.
Then you have not a hd 5850, it consumes a lot I guess. a HD7850 may be a better base (or a 6850 using the 45nm lithography). That's plenty of Watts. That's a gpu with a 256 bits bus. that's two memory pools. etc.

Say a fairer (but hypothetical) comparison would be a Athlon III X4 same speed but using 32 nm lithography. Let say it consumes 45 Watts.
Then you have a 7770 or something with more CU enable but clock lower so its power consumption is in between the 7770 and the 7750 (45 and 75) so let say 60 watts.
The whole thing consumes 105 watts and there is 4 chips of gddr5 and 4 chip ddr3 consuming extra power. you need an a bit more complex board. You need two cooling solutions, etc.

Now what a kaveri with gddr5 would come close of that but cost most likely less. I beileve that there is no magic the GDDR5 might up kaveri power consumption, there might be trade off, in CPU speed, in mem controller speed, they may want to let the GPU clock untouched though but who knows (and it's not like 10% would change the big figure).
It might perform worse but from a cost pov, It may lead, may be not an awesome difference but on million units.

For the bug I'm may be overly dark as the early phenom incident is old now and Stream roller may either tweak K10 architecture or the BD cores v3, in both case AMD should would not be in unknown territory. It still clearly a risk, with Kaveri release no one know when in 2013.

Honestly Kaveri is not an amazing proposal but played right at the righ price (so pretty cheap) it may do Sony a lot of good, especially if they (/rumors) are right about missing 2013.
Kaveri will keep with pC games for a long while, no matter what MS aims at, could it be the moon.

EDIT

WRT to the hd 6670 / 7670 by AMD own numbers HD 5670 costs 75$ so it's cheap but not that much more like margins are terrible in that price range.
Edit 2 not amd number but looks like serious estimate. Link might come tomorow

Again I'm agree with you largely, especially on the APUs are the costs in the option, but that this would be the way or most "correct" to withstand/resist competition from tablets,smartphones etc in a cycle of five or more years?

I'm still thinking inside of my ignorance on the subject that may cpus "old" and Athlon II X4 customized (with some thweaks for consoles) and die shrink to 32nm,could be more interesting for their reliability,especially if they put something custom and with HD 5850 shrink at 28 nm (which dissipates "only" 151 watts under 40nm process) could offer more raw power (2.1 Tflops + Cpu) even if have any problems even with two memory controller (DDR3 and GDDR5 = 8 modules) for 4 GBs and still remain below 200 watts.

Half off the subject ...many rumors about next gen console sdks have a A8-3850 with 400 SIMD / shaders / Stream Processors + HD 6670/480SPs (in general gives something like 5*ps360 shader power) ... maybe they are primary setting SDKs in early stages and in fact Sony and MS are targeting for next year twice the performance as developers wish (Crytech, ID / Carmack etc.),if not would be a tremendous disappointment to an "old" gamer as I see that the next generation will come with hardware lagged very much before launch.

I'm still thinking ... on these Sdks and in the realm of pure speculation and foolish ... and if they (sony and ms) are showing signs they intend to use APUs seems clear that they are focusing efforts on a 1.25 Tflops on same die or at least praying here they are counting performance addition 25/33% to reach something like 1.6/1.7 like they did in past with old sdks for ps360 (ps3 have cell with 2.4GHz early stages and x360 gpu Radeon 9800/10800 etc).

(Sorry for so many "if","preaching" and crying is that I was just stunned despite suposedly efficience of APU see how sony and ms would be setting so low specs for the next generation ...)

Edit:
Kaveri
http://news.softpedia.com/newsImage/...ri-APUs-3.png/
Athlon II review
http://www.legitreviews.com/article/1297/1/

About HD 5850:
http://en.wikipedia.org/wiki/Compari...ocessing_units

Edited 3 :

This would be an issue that needs to be answered by the developers ... they would rather deal with a hardware "Athlon II X4" (could be another model) with 2GB DDR3 + Radeon HD 5850 with 1 or 2 GB GDDR5 with all its limitations in dealing with two memory buses latencies etc., but with 2+ TFlops more power or use an APU with all the supposed improvements and facilities(Cpu and gpu "talking" each other) bus on same unified die "on the fly" and 2GB GDDR5 etc etc,but only 1.25 Tflops rely on brute force?

Why i'm ask this? Cause sony (and probably MS) suposedly would be spending a lot of resources in PS4 (heard sometthing like US$1 billion around this years) and wants make life easier for developers (like psvita development model).

[Shifty Geezer]

Originally Posted by N2O

because if PS suite can do it,XBL can do it either(no offense,just guess)

No-one said otherwise. The topic had drifted onto how Sony could make $3 billion when Live! currently can't. One significant advantage of Sony over MS though is that they aren't tied to an OS. They can release Sony content on Android devices, or Windows. But this is all going OT. The only particular relevance I see is confirmation that Sony would like a hardware platform that's a good fit for their software platform, hence either ARM or x86, and a normal GPU.

[N2O]

Originally Posted by tunafish

iOS and android are the only smartphone platforms that are actually growing. All the new releases have managed to the WP7 is reduce the rate at which it's share is shrinking. It's still losing users, not just share.

esr plots the comscore numbers (US-only, collected from the carriers, reliable numbers on the total amount of smartphones connected to the networks, as opposed to the amount of smartphones sold) nicely, and the mobile sector is rapidly turning into a two-horse race.

http://www.catb.org/esr/comscore/

actually i mean iOS have Xbox Live,even it only have very few feature right now

Originally Posted by Shifty Geezer

No-one said otherwise. The topic had drifted onto how Sony could make $3 billion when Live! currently can't. One significant advantage of Sony over MS though is that they aren't tied to an OS. They can release Sony content on Android devices, or Windows. But this is all going OT. The only particular relevance I see is confirmation that Sony would like a hardware platform that's a good fit for their software platform, hence either ARM or x86, and a normal GPU.

what MS really tied just Windows(because it's dominant),but they don't tied WP because it still doesn't have enough big market

[3dilettante]

Originally Posted by Heinrich4

I'm still thinking inside of my ignorance on the subject that may cpus "old" and Athlon II X4 customized (with some thweaks for consoles) and die shrink to 32nm,could be more interesting for their reliability,especially if they put something custom and with HD 5850 shrink at 28 nm (which dissipates "only" 151 watts under 40nm process) could offer more raw power (2.1 + Cpu TFlops) even if have any problems even with two memory controller (DDR3 and GDDR5) for 4 GBs (8 modules) and still remain below 200 watts.

The closest thing approaching an Athlon X4 on 32nm is Llano.
In fact the Athlon II X4 638 and 641 are 32nm Llano chips, albeit most likely salvage product.
Despite it being a power-optimized design, we see Llano pulling about 100W when approaching 3 GHz, and most blame probably falls on the CPU part.

A straight shrink of an X4 to 32nm would most likely have been non-functional, since shrinks aren't that simple anymore.
An X4 tweaked and shuffled about to make it remotely acceptable at 32nm with AMD's newest design methodologies is what Llano is.
The TDPs for the chip were way more sensitive to CPU clocks and power, and Llano's yields were initially terrible and are rumored to be not all that great even with AMD and GF stating the process problems have been significantly improved.

If Sony wanted a 32nm X4 that's not Llano, it would pay more money for AMD to dust off the pipeline it abandoned for good reasons. Llano isn't acceptable for a console now, and it is far better equipped to working at 32nm than an X4.

[liolio]

Heinrich it's not even about raw power there is a significant difference in marketing FLOPS and the really world.
Putting some architectural difference away we have 3 type of FLOPS in the GPU world.
count FLOPS with:
vliw5
vliw4
scalar (Nv and GCN).

As I was saying to KB-somker about site that swallow marketing cool aid wrt to diminishing FLOPS peak figures in late AMD GPU and speak of efficiency because I'm iffy if they understand the difference in design.
The difference is not really efficiency it's a sound architectural difference. You can literally remove 20% (possibly a bit more on average) of the peak of a vliw 5 design vs scalar so from 2.1 for example you go down to 1.68TFLOPS for example.

It's not efficiency it's design.
In previous AMD VLIW 5 GPUs, the base of the design is not the Stream processor as touted by marketing materials... but a group of 5 ALUs.

All those ALUs are not equal, you have 4 std ALUs and the special one (in charge of trigonometry, etc. the RYS unit as people call it here ).

A SIMD is not 80 SP acting in vectorized fashion, not at all.
It's indeed 16 5 wide units acting in a vectorized fashion, hence that's why hardware.fr /behardware.com calls them 16 Vec5 (16 Vec4 for cayman).

In fact those VLIW blocks are organized in bigger blocks, quads , of VLIW5 units wrt regard to the register files (massive amount of register files).

So now those 5 wide block are MIMD units "adressed" 'operated' (can't find the proper saying) in a VLIW fashion. So it's up to the compiler to extract parallelism and make sure that those units are busy.
So all the 16 blocks in a SIMD receive them VLIW instructions but work on different data.

At this point you will notice that most of the work done by GPU is on four elements basically the Fifth alus the Transcendental one is there not be used all the time but to make sure specific operations execute fast. It was a cheap way to achieve that (the fifth ALU) as ALU are cheap.
On average utilization of a vliw5 units is 3.8 instructions per cycle I believe (should check but that a bit below 4).
There are reason in graphic workload forthat matter of fact but that are also architectural one, you five ALU and register can be accessed by all of them at the same time. This is complicated but there are plenty of posts in the forum that explains why it set a limitation to the design.

That the reason why AMD moved first to a VLIW4 design. It removes headache for the register ports design (that's if I understand properly... ) as well as the pressure it creates on the compiler.
There was a trade off trancendental operation are slower. Thing is on graphic workload the IPC is mostly the same as in a VLIW5 design (and still below 4)

that's where marketing FLOPS kicks in, the FLOPS were always given including the T unit and based on the SP numbers as if they were equal, BS so called hardware sites don't give a shit they care for clicks not accuracy. It was an irrelevant figure. Now AMD moved from it and using marketing parlance for people that are not interested in tech ( I can understand some people are just gamers and is not a sin). In marketing parlance it's "more efficient".
Whereas this diminishing number of SP and FLOPS as no impact or really marginal on the design but when you fed people SP and FLOPS as metric for performance you have to come with something.

For me it's not efficiency the compiler can't extract more ILP (instruction level parallelism) with the new design, it's just easier to avoid conflict in register access, neither can the hardware VLIW is by design as dumb as can be.

Then why AMD moved to Scalar / pure SIMD design as Nvidia? Because in some situation (not happens much in graphics) the IPC that can be extracted by the compiler is way below 3.8.
By design it can go as low 1 or 2. In effect using marketing parlance you don't have 80 SP in your SIMD but respectively 16 and 32... massive hit in efficiency.

That's whay AMD moved to a scalar?plain SIMD design. Extracting ILP in graphic were pretty easy but as compute get relevant it turns into a double edge sword. there are case where there is simply niot that much ILP to be extracted when it happens the architecture (even refined VLIW4) fails.
AMD gave up with GCN on leveraging Instruction Level parallelism, that;s it.

A nice effect is that comparing a CU so 64 ALUs and a cayman SIMD 64 ALUS too, the former always achieve 64 operation per cycle (it's an incorrect way to put it but that's pretty much the figure looking from the distance) whereas on average Cayman SIMD will do 3.8*16 in graphic and can end up well below in other cases. Even a VLIW 5 design would not beat GCN it would push 3.8*16 pretty much as cayman.
EDIT ALUs utilization is a more correct way to put it as instructions are likely to take more than one cycle to execute. So you have 100% ALUs utilization on one side (plain SIMD) amd3.8/5*100 or 3.8/4*100 on the other/EDIT

I've no clear understanding of those low level stuffs but that's the best description I can give you about "not taking marketing and FLOPS figures... at their face value". Hope it helps.

Others members can correct approximations /or things I would get wrong if they want or provide even more information.

[Heinrich4]

Originally Posted by 3dilettante

The closest thing approaching an Athlon X4 on 32nm is Llano.
In fact the Athlon II X4 638 and 641 are 32nm Llano chips, albeit most likely salvage product.
Despite it being a power-optimized design, we see Llano pulling about 100W when approaching 3 GHz, and most blame probably falls on the CPU part.

A straight shrink of an X4 to 32nm would most likely have been non-functional, since shrinks aren't that simple anymore.
An X4 tweaked and shuffled about to make it remotely acceptable at 32nm with AMD's newest design methodologies is what Llano is.
The TDPs for the chip were way more sensitive to CPU clocks and power, and Llano's yields were initially terrible and are rumored to be not all that great even with AMD and GF stating the process problems have been significantly improved.

If Sony wanted a 32nm X4 that's not Llano, it would pay more money for AMD to dust off the pipeline it abandoned for good reasons. Llano isn't acceptable for a console now, and it is far better equipped to working at 32nm than an X4.

Excellent information you send us always.

But if there is somehow the Athlon II X4 32nm in approaching the shape on the Llano cpu cores then perhaps best to be part of this hypothesis cpu cpu + gpu apart.

And indeed your statements as always accurate Athlon II X4 may not be the best choice, but as I said before ("can be another cpu"), perhaps another cpu (Bulldozeer, Bobcats, "Pill something"etc.) could be interesting in this paradigm "non APU like".

[liolio]

Originally Posted by Heinrich4

Excellent information you send us always.

But if there is somehow the Athlon II X4 32nm in approaching the shape on the Llano cpu cores then perhaps best to be part of this hypothesis cpu cpu + gpu apart.

And indeed your statements as always accurate Athlon II X4 may not be the best choice, but as I said before ("can be another cpu"), perhaps another cpu (Bulldozeer, Bobcats, "Pill something"etc.) could be interesting in this paradigm "non APU like".

I think what he means is that AMD moved to bulldozer. It's not like AMD have that much resources.
AMD focus on two architectures Bulldozer in the high power CPUs and the Brazos/whatever they are named low power CPUs.

And that implementing a chip on silicon is getting tougher and tougher as well as more and more expansive. Sony would ahve to found them a hell lot of money to push an architecture they may have left behind for go reasons, no matter how BD v1 performs.

Either CPu or APU (if not llano) his pov is that bulldozer on cpu is to be expected. So for me he already answered your question

[Heinrich4]

Originally Posted by liolio

Heinrich it's not even about raw power there is a significant difference in marketing FLOPS and the really world.
Putting some architectural difference away we have 3 type of FLOPS in the GPU world.
count FLOPS with:
vliw5
vliw4
scalar (Nv and GCN).

As I was saying to KB-somker about site that swallow marketing cool aid wrt to diminishing FLOPS peak figures in late AMD GPU and speak of efficiency because I'm iffy if they understand the difference in design.
The difference is not really efficiency it's a sound architectural difference. You can literally remove 20% (possibly a bit more on average) of the peak of a vliw 5 design vs scalar so from 2.1 for example you go down to 1.68TFLOPS for example.

It's not efficiency it's design.
In previous AMD VLIW 5 GPUs, the base of the design is not the Stream processor as touted by marketing materials... but a group of 5 ALUs.

All those ALUs are not equal, you have 4 std ALUs and the special one (in charge of trigonometry, etc. the RYS unit as people call it here ).

A SIMD is not 80 SP acting in vectorized fashion, not at all.
It's indeed 16 5 wide units acting in a vectorized fashion, hence that's why hardware.fr /behardware.com calls them 16 Vec5 (16 Vec4 for cayman).

In fact those VLIW blocks are organized in bigger blocks, quads , of VLIW5 units wrt regard to the register files (massive amount of register files).

So now those 5 wide block are MIMD units acting in a VLIW fashion. So it's up to the compiler to extract parallelism and make sure that those units are busy.
So all the 16 blocks in a SIMD receive them VLIW instructions but work on different data.

At this point you will notice that most of the work done by GPU is on four elements basically the Fifth alus the Transcendental one is there not be used all the time but to make sure specific operations execute fast. It was a cheap way to achieve that (the fifth ALU) as ALU are cheap.
On average utilization of a vliw5 units is 3.8 instructions per cycle I believe (should check but that a bit below 4).
There are reason in graphic workload forthat matter of fact but that are also architectural one, you five ALU and register can be accessed by all of them at the same time. This is complicated but there are plenty of posts in the forum that explains why it set a limitation to the design.

That the reason why AMD moved first to a VLIW4 design. It removes headache for the register ports design (that's if I understand properly... ) as well as the pressure it creates on the compiler.
There was a trade off trancendental operation are slower. Thing is on graphic workload the IPC is mostly the same as in a VLIW5 design (and still below 4)

that's where marketing FLOPS kicks in, the FLOPS were always given including the T unit and based on the SP numbers as if they were equal, BS so called hardware sites don't give a shit they care for clicks not accuracy. It was an irrelevant figure. Now AMD moved from it and using marketing parlance for people that are not interested in tech ( I can understand some people are just gamers and is not a sin). In marketing parlance it's "more efficient".
Whereas this diminishing number of SP and FLOPS as no impact or really marginal on the design but when you fed people SP and FLOPS as metric for performance you have to come with something.

For me it's not efficiency the compiler can't extract more ILP (instruction level parallelism) with the new design, it's just easier to avoid conflict in register access, neither can the hardware VLIW is by design as dumb as can be.

Then why AMD moved to Scalar / pure SIMD design as Nvidia? Because in some situation (not happens much in graphics) the IPC that can be extracted by the compiler is way below 3.8.
By design it can go as low 1 or 2. In effect using marketing parlance you don't have 80 SP in your SIMD but respectively 16 and 32... massive hit in efficiency.

That's whay AMD moved to a scalar?plain SIMD design. Extracting ILP in graphic were pretty easy but as compute get relevant it turns into a double edge sword. there are case where there is simply niot that much ILP to be extracted when it happens the architecture (even refined VLIW4) fails.
AMD gave up with GCN on leveraging Instruction Level parallelism, that;s it.

A nice effect is that comparing a CU so 64 ALUs and a cayman SIMD 64 ALUS too, the former always achieve 64 operation per cycle (it's an incorrect way to put it but that's pretty much the figure looking from the distance) whereas on average Cayman SIMD will do 3.8*16 in graphic and can end up well below in other cases. Even a VLIW 5 design would not beat GCN it would push 3.8*16 pretty much as cayman.
EDIT ALUs utilization is a more correct way to put it as instructions are likely to take more than one cycle to execute. So you have 100% ALUs utilization on one side (plain SIMD) amd3.8/5*100 or 3.8/4*100 on the other/EDIT

I've no clear understanding of those low level stuffs but that's the best description I can give you about "not taking marketing and FLOPS figures... at their face value". Hope it helps.

Others members can correct approximations /or things I would get wrong if they want or provide even more information.

I'm always enjoy your posts immensely Liolio.

I agree entirely with the design have a decisive influence on the efficiency (pipes full or processing much more,less latencies, less cycles etc) and these GPUs and APUs, but despite this plethora of "flops" VLIW5 (5850 if not my mystake), VLIW4 (6950) and scalar (NV gpus ,AMD HD 7970..5850/6970 only double precision) i'm still be interesting with his "5850 2.1 TFLOPS marketing" VLIW5/Scalar(is just a simble... eficiencie reach something like 74% on these archtecture according Stanford) against one of Apu about "1.25Tflops marketing".

Talking here in a hipothesis Sony/MS doesn't much care (but not going bleeding ressources like ps360 before...) about development cost close box console... We may don't know (talking about me) for sure the full real pros and cons of an APU with about 2.5 times more power than a A8-3850 (A8-3850 + HD 6670 it seens sony and Ms setting for your APU console),but perhaps if we look at tests,benchmarks,discounting eficiencies of the universe closed box i'm still keep thinking APU probably less Interestingly than a set with Cpu (Bobcat,Enhanced Bobcat, Bulldozzer, whatheaver pile etc) and a custom GPU GCN (less clock etc) apart at same TDP...but its just my only 2 cents..