Predict: The Next Generation Console Tech

Status
Not open for further replies.
The problem with separate encryption units, and the reason there has been a shift away from them into encryption-aiding cpu instructions, is that in modern software, (typically, requests and responses in web browsers), you usually only encrypt quite small chunks at a time. When you are never encrypting more than a few kb at a time, the cost of moving data to the dedicated encryption unit can easily wipe out most of the gains, especially when you often could otherwise process the data in the same pass you de/encrypt it.

Coherent interconnects are going to be there in future consoles. Just make the encryption/decryption unit listen to that fabric.
 
function said:
Ivy Bridge graphics should be approaching console level and will be force fed to almost everyone.
Actually I7 SandyBridge would already be competitive with current consoles if not for certain API abstractions and the fact that people don't program for it as a platform.
And even so - it can be pushed to some pretty neat things.
 
AlphaWolf, the word homerdog was not using was "Enthusiast Single Chip GPU" and "Flagship model." So yes, there is a class above Mid Range and High Performance. Games are now even reflecting this where there is Low, Medium, High, and Ultra (and/or Enthusiast).

As for the NV 560 TI (448 shader edition), I wouldn't consider a mere 5-20% drop in performance from a 580GTX "mid range" especially when it nips at the heals of the AMD high end cards and sometimes bests them--how is that mid range? (And then there is the MSI version stocked overclocked that is like 8% faster yet than the stock 560 TI models). Of course the issue with the NV 560 versus the AMD models is power. Performance per watt is not where AMD is. Anyways, this is not a "mid range" product. Traditionally a mid range product is about half the performance of a top tier GPU (think 9800/9600 AMD products or the 6800/6600 series from NV where the upper middle tier GPUs were like half the functional units). i.e. Mid range is going to force you to cut resolution and quality nearly in half. A 560 TI, as Anand notes, a "3rd tier" upper echelon product right behind the Flagship 580 and 570.

On the CPU side, there is a pretty big drop off after the $200 range of CPUs. Your bang for buck drops quickly. Again, an i3 is not a low end CPU in the PC market. Maybe in the "Fall 2011 Gaming PC" segment but the PC market, which includes a lot of casual gaming, has a host of CPUs well below an i3 selling in huge volumes.

On another note, we could all dream of 20nm hitting some risk production in H2 2012 and volume by early 2013. We can dream, right? It does sound, on face value, TSMC thinks they can have volume 28nm in 2012. The cost and yields is probably a different story...

http://www.electronicsweekly.com/Articles/04/11/2011/52218/tsmc-ramping-28nm-3x-faster-than-40nm.htm

http://www.electronicsweekly.com/bl...log/2011/11/arm-tsmc-moving-fast-to-20nm.html
 
Last edited by a moderator:
AlphaWolf, the word homerdog was not using was "Enthusiast Single Chip GPU" and "Flagship model." So yes, there is a class above Mid Range and High Performance. Games are now even reflecting this where there is Low, Medium, High, and Ultra (and/or Enthusiast).

As for the NV 560 TI (448 shader edition), I wouldn't consider 5-20% the speed of a 580GTX "mid range" especially when it nips at the heals of the AMD high end. And then there is the MSI version stocked overclocked that is like 8% faster yet. Of course the issue with the NV 560 versus the AMD models is power. Anyways, this is not a "mid range" product.

On the CPU side, there is a pretty big drop off after the $200 range of CPUs. Your bang for buck drops quickly. Again, an i3 is not a low end CPU in the PC market. Maybe in the "Fall 2011 Gaming PC" segment but the PC market, which includes a lot of casual gaming, has a host of CPUs well below an i3 selling in huge volumes.

Thank you Josh, that's basically what I was getting at. You have a way of making points that I have yet to master. :smile:

I just have to reiterate that the i3 desktop CPUs are far from low end. They can deliver an enormous amount of performance in a very small and power efficient package. A true marvel of modern engineering.

The equivalent AMD CPUs are much larger, have twice or thrice the cores, and consume more than twice the power.

Since the consoles will be more in line with AMD's capabilities than Intel's I think we should temper our predictions of what a next gen console CPU can be. Low-midrange Intel desktop CPUs will be just as fast or faster than next gen console CPUs by the time they come out.

That is unless MS or Sony can get Intel Inside. That would kick ass.
 
If the transistor density stays the same but the die grows by %30 and the frequency grows by 25% and perf/watt stays flat while yields increase by 90% then you're probably going to be rubbing your hands together, dancing round the campfire naked drinking moonshine.*

Take a look at AMDs fastest 40nm lineup and see how things went in the first 12 months. Now take a look at the 480 and then the 580.

Dies growing is not something that is wanted, but required if you need a bigger chip to get performance. More silicon is more costs. Yes you want great yields with big chips and great yields can enable you to make a bigger chip, but for a console that has such a long lifespan, the size of the dies is a huge cost factor in itself and waiting a bit longer to get a bigger die is nonsense.

Fermi chip is 520mm2. Console relevance is low. Your comparisons still don't make too much sense, but I'll keep it short from now on. You have good points on the importance of mature process don't get me wrong, but imo you slightly confuse the differences between chip and console makers.

AlphaWolf, the word homerdog was not using was "Enthusiast Single Chip GPU" and "Flagship model." So yes, there is a class above Mid Range and High Performance. Games are now even reflecting this where there is Low, Medium, High, and Ultra (and/or Enthusiast).

As for the NV 560 TI (448 shader edition), I wouldn't consider a mere 5-20% drop in performance from a 580GTX "mid range"

I'm sorry but that example is little bit too much cherry picking. 560Ti 448-edition is a limited time special edition card, that is 560 in name only. Much proper name for it would be 570 LE or something like that. It uses the same 520mm2 die that is in GTX 580 and GTX 570, other 560 cards do not. It's more expensive than the regular 560s as well. It's just a handy way for nVidia to sell chips that didn't quite make it to be in the GTX 570. I'm sure no-one meant that model or Evga's GTX 560 2win card either, when they were talking about the GTX 560.

Imo the regular GTX 560 is an upper mid range card that has pretty good value proposition.
Or entry level high end...
 
Last edited by a moderator:
Performance per watt is not where AMD is. Anyways, this is not a "mid range" product. Traditionally a mid range product is about half the performance of a top tier GPU (think 9800/9600 AMD products or the 6800/6600 series from NV where the upper middle tier GPUs were like half the functional units). i.e. Mid range is going to force you to cut resolution and quality nearly in half. A 560 TI, as Anand notes, a "3rd tier" upper echelon product right behind the Flagship 580 and 570.
I've usually heard it is seen at most usually about a 50% jump going from gpus in the $200 range to gpus in the $300 range. We also have to keep in mind that iirc new gpus could very well have appeared at fall but were delayed to early next year. That could likely lower the price of a 560ti and its comparison to the new high end. Had new set of gpus made it to market this fall, we would have to consider in what range that would put a 560. Also we've to take into account that performance per watt is highly relevant in terms of console applicability, as well as area. IF amd gpus offer comparable performance with smaller gpus and lower power consumption, this relates to what's viable in the console space.
Since the consoles will be more in line with AMD's capabilities than Intel's I think we should temper our predictions of what a next gen console CPU can be. Low-midrange Intel desktop CPUs will be just as fast or faster than next gen console CPUs by the time they come out.
Aren't consoles likely to go with ibm power cpus? What's the rumored wiiU cpu core? how does a console viable power compare to intel's offerings? I mean, at things like SP floating points, aren't i7's at most close to the 5 year old cell? I would assume SP floating point is relevant to console cpu performance seeing as that was raised quite above what was possible in the pc cpu arena when these console cpus launched.
 
On another note, we could all dream of 20nm hitting some risk production in H2 2012 and volume by early 2013. We can dream, right? It does sound, on face value, TSMC thinks they can have volume 28nm in 2012. The cost and yields is probably a different story...

http://www.electronicsweekly.com/bl...log/2011/11/arm-tsmc-moving-fast-to-20nm.html

Well now, this changes the game quite a bit ....

"We expect A-15 to be sampling in the first half of next year, to be in full production in Q4 2012, and to be out in hand-sets by the end of next year,"

Compared to 28nm the 20nm process is expected to deliver a 25% improvement in power consumption, a 15-20% improvement in performance and a 1.9x increase in density...

I had no idea TSMC was being this aggressive!

It seems all the iphone and ipad rage might actually bring about some positives for the gaming industry afterall ... :cool:

At roughly same die size budget of xb360/ps3, we'd be looking at >16x :oops: transistors on 20nm or roughly 8billion!

If 2013 is really "the year" and this is the reason for the delay, I'd be ok with it ;)
 
Last edited by a moderator:
Mind you, there are different types of 28nm. Unless you're pining for a cell phone part in a console, how is this stuff related ?

Don't you think they should mention Nintendo as a customer? They're the ones who want the fabbing done.
 
Mind you, there are different types of 28nm. Unless you're pining for a cell phone part in a console, how is this stuff related ?

Don't you think they should mention Nintendo as a customer? They're the ones who want the fabbing done.

For the old technology Nintendo is going with, they don't need cutting edge manufacturing.

Just as they didn't need it for Wii's ancient tech.

As for this being "cell phone tech", granted, at the outset (q4 2012), it probably won't be suitable for high performance large chips, but as many of you guys have been saying, 2012 isn't the target date anyway.

So 2013 it is.

That leaves plenty of time (1 year) to refine and get it ready.

Just as 1 year from now is plenty of time for 2012 28nm.
 
I had no idea TSMC was being this aggressive!
This is not unheard of it foundry marketing, those numbers have been bandied about for previous transitions. I'm curious if these are for a single design target, often one can get each of those scaling factors along the power, density, and performance axes, but not at the same time.

Notably, the long-term trend for power has not changed overmuch. Power consumption improves by 25%, when there will be nearly double the number of transistors possible in the same area.
 
I've usually heard it is seen at most usually about a 50% jump going from gpus in the $200 range to gpus in the $300 range.

Not even close really. 6950 runs ~$250, 6970 about ~$350 and there is nowhere near a 50% performance gap between the two.

Aren't consoles likely to go with ibm power cpus?

Yes, and IBM is closer to AMD than Intel in terms of manufacturing prowess.

What's the rumored wiiU cpu core? how does a console viable power compare to intel's offerings? I mean, at things like SP floating points, aren't i7's at most close to the 5 year old cell? I would assume SP floating point is relevant to console cpu performance seeing as that was raised quite above what was possible in the pc cpu arena when these console cpus launched.

SP floats you say? Perhaps the GPU could deliver more of those than any CPU could ever dream of?
Besides, the i7 CPUs are like an order of magnitude faster than Xenon or Cell when it comes to running games, so SP floats must not be the largest factor in CPU performance.

This is why I don't see Cell in a next gen console. Cell's capabilities would overlap with a GCN or Fermi GPU.
Go with a smallish OoO CPU with predictable performance like the i3-2100 (or as close as IBM/AMD can get to it) and a larger, very flexible GPU. Something based on Kepler or GCN.
 
http://wiiudaily.com/2011/12/wii-u-has-quad-core-3ghz-cpu-768-mb-of-ram/

* Quad Core, 3 GHz PowerPC-based 45nm CPU, very similar to the Xbox 360 chip.
* 768 MB of DRAM “embedded” with the CPU, and shared between CPU and GPU
* Unknown, 40nm ATI-based GPU

That's the most important piece.

As is, developers should have their hands on the dev kits and should know whats under the hood in that regard ... and word up to this point, has been called roughly a Radeon HD4770.

960 GFLOPS, 826 million trans, 137mm^2, 80w TDP @ 40nm.

Not bad at all in comparison to xb360 and ps3 (~4X), but should be a good deal slower than true nextgen hardware.
 
As is, developers should have their hands on the dev kits and should know whats under the hood in that regard ... and word up to this point, has been called roughly a Radeon HD4770.

assuming those dev kits have final hardware, which may or may not be the case

(I expect it's unlikely at this point)
 
Not bad at all in comparison to xb360 and ps3 (~4X), but should be a good deal slower than true nextgen hardware.

Keep in mind, that does fly in the face of the +50% comment by a ton. A 4770 is worlds away from RSX or Xenos. 80-90W TDP for just the GPU in that chassis is a bit much don't you think? That's approaching original 90nm current gen consumptions.
 
2x x800 was a good indication of xenos? If you say so.

Indeed, MS was pushing the limits and an appropriate gpu with similar technology didn't exist.

All indications are Nintendo is looking for low power and dx10 (old tech).

No indication that they are pushing the boundaries, thus whatever they have in the dev kits, probably about as good as it gets.
 
Keep in mind, that does fly in the face of the +50% comment by a ton. A 4770 is worlds away from RSX or Xenos.

True, but it may have other handicaps which are limiting total system compute to 50%.

The CPU is one limit that would be holding it back if the report is accurate, and we have no idea what the ram bandwidth situation is aside from total amount (50% above xb360 and ps3)... might be 64bit bus to save money.
 
True, but it may have other handicaps which are limiting total system compute to 50%.

Well, what's making me really pause about this 4770 belief is that the 360 Slim already consumes 90W at load, so... it'd have to be some sort of dark nega-energy magic to put an 80W GPU inside the WiiU. That's not even considering the CPU!

I've mentioned it moons ago, but something like a 400 ALU part (e.g. 5670, 60W @ load) would make a lot more sense. Even downclocking it to 500MHz, you would still get something that can go far beyond Xenos' capabilities whilst still being thermally viable when you still have the supposed 360-like CPU to deal with.

And speaking of the CPU in relation to the "+50% on paper" comment, there will likely be larger caches (it'd be really really hard to not exceed 1MB L2 total *cough*) as well as OOOE/quad-threads (assuming any truth to Power 7 inheritence) that push up the system's capabilities and power consumption.
 
Status
Not open for further replies.
Back
Top