Predict: The Next Generation Console Tech

[liolio]

I am the last person in the world to doubt STI's abilities to create a monster CPU, as many of you already know However, I am a lot more conservative with regards to the specifications on the PS4 because of a few things.

1. The success of the Wii, things will be different going into next generation. Expect to see Sony focus on some type of revolutionary(or not) interface, perhaps utilizing built in cameras that can track and diffirentiate between several bodies. This is why I have dubbed the PS4 CPU the Sense Engine.

2. Kutaragi is(largely) out of the picture, he was the major drive for the most cutting edge hardware possible.

3. They will not launch at anything higher than 400 US dollars and I do not see them taking massive hits on hardware in the beginning.

4. 32 or more cores is just impractical IMO. Think of the SPE's you'll have to reserve for redundancy, sustained performance, die size, etc. It will just be a nightmare for developers, it's best to jump up clock speed and make the necessary architectural improvements for a 16-24 SPE beast, rather than throwing 32+ less powerful cores on there to inflate floating point power.


Things have changed, and while the PS4 will feature an incredible hardware spec; we will not see Sony pushing things to the max. A CPU with 1-1.5 TFLOPS computing power and a complimentary Cell based GPU will be nothing to scoff about. You'll be looking at a next gen Uncharted game looking roughly on par with the CGI in the commercial.

1 I agree Wii succes will have consequences, but not only on the controller ( I think that in most case an improve wii mote would be enough). Sony may be forced to launch a system earlier than they planed.

2 & 3 agreed.

4)It depends on when Sony will have to launch the system, 32nm has to be widely available, 22nm will could be difficult to reach (Intel is in a league of its own) Sony would have that many room for costs reductions via chip shrinking.
@32nm the cpu you describe is OK even with larger LS improved executions units etc.
But waiting till 32nm could be dangerous to Sony, Sony opponents are likely to launch their next system earlier!
could Sony face the 360 andthe Wii on the low end and 720 and Wii2 on the high end with the ps3 alone?

They could be let for some years with very few sales left!

I think it would be better to think more marketing terms like "what are our goals?" than technical one like "what could be done?"

 

[liolio]

I read that yesterday:
http://www.lostcircuits.com/cpu/amd_phenom/

It goes pretty much in Graham sense memory system can make up for less efficient core!

It should be the main MS goal for RD instead of craming to many cores all together.

Devs will always be in the learning curve as far as multithreaded development is concerned, on the PC side SMP with not that many core will also be the mainstream.

Ireally hope Marketing Flop won't pull MS in the wrong direction.

 

[Panajev2001a]

I am the last person in the world to doubt STI's abilities to create a monster CPU, as many of you already know However, I am a lot more conservative with regards to the specifications on the PS4 because of a few things.

1. The success of the Wii, things will be different going into next generation. Expect to see Sony focus on some type of revolutionary(or not) interface, perhaps utilizing built in cameras that can track and diffirentiate between several bodies. This is why I have dubbed the PS4 CPU the Sense Engine.

2. Kutaragi is(largely) out of the picture, he was the major drive for the most cutting edge hardware possible.

3. They will not launch at anything higher than 400 US dollars and I do not see them taking massive hits on hardware in the beginning.

4. 32 or more cores is just impractical IMO. Think of the SPE's you'll have to reserve for redundancy, sustained performance, die size, etc. It will just be a nightmare for developers, it's best to jump up clock speed and make the necessary architectural improvements for a 16-24 SPE beast, rather than throwing 32+ less powerful cores on there to inflate floating point power.


Things have changed, and while the PS4 will feature an incredible hardware spec; we will not see Sony pushing things to the max. A CPU with 1-1.5 TFLOPS computing power and a complimentary Cell based GPU will be nothing to scoff about. You'll be looking at a next gen Uncharted game looking roughly on par with the CGI in the commercial.

Paul, I really doubt we will see a GPU that is CELL based and not nVIDIA (more likely for PS4 and less likely for PSP2) or PowerVR (less likely for PS4 and more likely for PSP2 [it would make a lot of sense there]) designed in PS4.

 

[Paul]

Pana.. re-read my earlier post!

On the GPU side, I do see nVIDIA having part in it again. However, I do not see the GPU just being an off the shelf rip off again, as was sadly the case with RSX. It's going to leverage off the Cell architecture TO A DEGREE, think Visualizer but not with that many cores.

 

[Panajev2001a]

Pana.. re-read my earlier post!

My answer is the same: I do not think PS4's GPU will leverage CELL in its own micro-architecture... meaning no SPE's in it .

 

[kyetech]

My thoughts on the future of console tech.

I’m having a really hard time imagining next gen consoles with 4GB of ram and 400 - 500GB of bandwidth etc....I think, the next lot of machines are not going to max specs out like PS3 did this time around. I cant comment on architectures, since they haven’t shaken out enough yet to even guess and 2009 is going to be the year when architectural designs become clearer to us.

Here are my key reasons for thinking we wont get maxed out monsters for next gen like we did with PS3

- WII / DS is proving you don’t need cutting edge to win.

- Software houses cannot cope with demands of production. Graphics is killing the industry.

- PS3 has given sony a rough ride / headaches and gambles in execution, and it prob represents the last of the behemoth consoles.

- Even Microsoft doesn’t want to lose another $4 billion in a gen despite having bottomless pockets, they still have concerned shareholders, and they still have to make business sense in what they execute.

- Processes for fabing chips after 22nm is very unclear, meaning that their path to cost reduction is less sure and higher risk, especially if walls are hit after 22nm.

- No one will ever risk a console that will need cost more than $399 on launch ever again. It simply won’t happen. Sony has only just got away with the huge risk/cost, and still are not out the water yet. Price brackets are very sensitive no matter how strong the brand. Something Sony learnt the hard way.

therefore:

Sorry, but the writing is on the wall. we wont see the same massive jumps, and tech squeezing like today ever again. Any one wishing for 3Billion transistor discreet graphics chips, with a 2billion transistor CPU monsters etc etc etc... will be disappointed.

I’m a graphics whore as much as the next, but I believe we will need to look to new spaces for it, in the shorter term: PC games and tech demos are a given i guess, but longer term I think it will be more likely be found in the hardware rendering in graphics houses, for special effects and CGI sequences (not necessarily real time at first, but much more exciting uses and tech and rendered in hardware (as opposed to software offline rendering))

Think about it realistically, if we had the hardware to render FFTSW in real time(hell even just toy story 1 would be a nightmare) how many software houses could afford to build engines that could use them? In fact at that level the issue is not just about art (even though that is a MASSIVE problem), but more about having good enough programming that can truly exploit the function of the hardware on all levels... physics / lighting / effects / animation etc. whilst being highly mathematically optimised and built for multicore.

Im not saying things wont progress over time, Im just saying that I believe the next 10 / 15 years are gonna slow right down while software has a chance to catch up, and hardware starts hitting new walls. something I believe is on the cards.


However here are my guesses for some basic specs based on what I have read and what I logically deduce.

a *single* 1.5 - 2 billion (tops) transistor system on chip. with super smart domain clocking and super smart cache sharing. with embedded dram and large / intelligent internal bandwidth.

a respectably clocked 256bit external memory bus.

2GB ram.

Around the same performance (or just below) as an SLI rig in 2009 (new gen parts)

Time will tell i guess.

/* end rant */ ;-)

 

[liolio]

Kietech I agree with You!
The only I disagree with is RAm 4gb may be cheap and while saving RD costs it's an easy way to improve perfs.

did you agree with my post about the Wii2 and how MS and Sony will have to fight to not be caugh the pant down?

 

[archangelmorph]

Kietech I agree with You!
The only I disagree with is RAm 4gb may be cheap and while saving RD costs it's an easy way to improve perfs.

did you agree with my post about the Wii2 and how MS and Sony will have to fight to not be caugh the pant down?

if we were talking about DDR2 then sure 4GB of ram would be quite cheap, however i'm not sure GDDR4(/5?) or XDR ram will be anywhere close to cheap enough by then to warrant such large quantities..

They *could* always stick with GDDR3 however which could possibly provide a cheaper solution by then with a larger capacity (maybe 3-4GBs) providing the increase in memory bandwidth over increasing the clock but who knows..

I do agree however that the next round of platforms will be as economically friendly on the companies' bottom lines as possible..

 

[hoho]

What about two memory pools? 512/1G fast (video) RAM and bunch of slower DDR2/3 for other stuff? It won't be simple for developers, though.

 

[kyetech]

Kietech I agree with You!
The only I disagree with is RAm 4gb may be cheap and while saving RD costs it's an easy way to improve perfs.

easy to improve perf... /on the flip side/.... easy way to reduce costs.

did you agree with my post about the Wii2 and how MS and Sony will have to fight to not be caugh the pant down?

can you link me to that post, def sounds about right from what you have put.

if we were talking about DDR2 then sure 4GB of ram would be quite cheap, however i'm not sure GDDR4(/5?) or XDR ram will be anywhere close to cheap enough by then to warrant such large quantities..

They *could* always stick with GDDR3 however which could possibly provide a cheaper solution by then with a larger capacity (maybe 3-4GBs) providing the increase in memory bandwidth over increasing the clock but who knows..

agree,

Its not impossible to have 4GB of ram but you would be looking at much lower perf ram. and I dont think its worth having too much disparity between size and bandwidth. I think 2GB and higher bandwidth would be better than 4GB and lower bandwidth.

 

[Acert93]

One thing to consider is that there remains an active and aggressive market of "gamers" who do place some importance on graphics and technology. These gamers also tend to buy a significant amount of games.

If 2 of the 3 companies go the "Wii route" the third company stands a high degree of capturing this audiance. Currently, in the US, Microsoft appears to have captured a healthy portion of this market. When you capture the "gamer" who plays Madden and the like religiously they can inturn evangelize to their friends. As the console prices drop I think we will see this effect.

So while we may see some return to sanity next generation in terms of console pricing with more of a focus on early cost reduction and an eye on the process shrink issues, I also don't believe companies will automatically avoid investing in performance. That said I do see the technology shift moving toward content creation and maximizing development resources for at least one of the companies with another taking the "leverage last gen tech" as a means to addressing cost concerns.

 

[AlNom]

Couple questions given recent disc format events:

• Can Microsoft feasibly go with an otherwise dead format, becoming the sole consumer of HD DVD?
  • how costly would it be to keep producing HD DVDs? (Assuming the remaining exclusive HD DVD movie studios switch over to BD exclusively)
• Given the hurdles of asset/content creation, can Microsoft get away with HD DVD's lower space?

Several things to consider

• How much more diverse can we expect games to get?
• Advanced compression algorithms reduce storage footprint, which is already a nice candidate for multi-threaded decompression
• As the sole user of HD DVD, it would be fairly difficult to pirate games through disc copying.
• Extra space allows:
  • data redundancy to reduce the random access, seek times
  • high def cut-scenes to hide loading
• Multiple layers
  • HD DVD --> SL = 17GB, DL = 34GB, TL = 51GB
  • BD --> SL = 25GB, DL = 50GB etc

There are speed penalties associated with more layers, and keep in mind how data transfer speed has scaled poorly with the increase in system memory available to consoles. Per 1x speed, data transfer rates for the blue laser drives are only an improvement of ~3.4x compared to DVD.

 

[V3]

Couple questions given recent disc format events:

• Can Microsoft feasibly go with an otherwise dead format, becoming the sole consumer of HD DVD?
  • how costly would it be to keep producing HD DVDs? (Assuming the remaining exclusive HD DVD movie studios switch over to BD exclusively)

• 
  
  They could, but dead format is bad for business, you don't want to be associated with dead format. It's kinda pointless to discuss IMO. Once HD DVD is really dead, I expect MS to erase all the existance of the add on. It's just bad for business.

 

[V3]

One thing to consider is that there remains an active and aggressive market of "gamers" who do place some importance on graphics and technology. These gamers also tend to buy a significant amount of games.

If 2 of the 3 companies go the "Wii route" the third company stands a high degree of capturing this audiance. Currently, in the US, Microsoft appears to have captured a healthy portion of this market. When you capture the "gamer" who plays Madden and the like religiously they can inturn evangelize to their friends. As the console prices drop I think we will see this effect.

So while we may see some return to sanity next generation in terms of console pricing with more of a focus on early cost reduction and an eye on the process shrink issues, I also don't believe companies will automatically avoid investing in performance. That said I do see the technology shift moving toward content creation and maximizing development resources for at least one of the companies with another taking the "leverage last gen tech" as a means to addressing cost concerns.

I reckon both Sony and MS will still invest in high end stuff, just not so soon. That is I am not expecting a new gen within 3-4 years, and it will last even longer.

For Nintendo, they'll milk Wii for all its worth, they'll probably the last company that'll launch their next gen consoles (not counting the possible scenario of Wii HD).

 

[Shifty Geezer]

Couple questions given recent disc format events:

• Can Microsoft feasibly go with an otherwise dead format, becoming the sole consumer of HD DVD?
  • how costly would it be to keep producing HD DVDs? (Assuming the remaining exclusive HD DVD movie studios switch over to BD exclusively

Why even consider HD DVD? Wouldn't it make far more sense to use BRD with it's mass-produced drive and add BRD playback to your console?

 

[liolio]

I agree, why consider hd-dvd by now?

If Ms sells 360 with built in hd optical drive it will be for movies playback, I expect MS to stick to dvd for Games for quiet some times.
If it was for games I guess the chinese format related to hd-dvd would be the best choice for cost saving purpose.

 

[AlNom]

It would make more sense to go BD because everyone else is, but I'm still curious to know the alternatives. I'm just trying to keep an open mind here. MS will be paying royalties for using BD, despite it being inherently cheaper. How much would it cost them to go HD DVD

I was ignoring the BD playback though.

 

[Acert93]

Any company going with optical media is going to have to address the transfer rate issues. On the hardware side a HDD (for content install) and a fairly large flash based intermediate cache seem like no brainers, although HDD costs will be a concern. On the software side the trend toward content generation created at loadtime as well as dynamic technologies may help alleviate some of the issues.

It is interesting seeing how some areas of technology progress at different paces. Addressing the correct issues to get the best end result for your buck is going to be interesting to watch as it unfolds.

 

[Megadrive1988]

I had hoped that optical disc technology would be eliminated from the next round of consoles, but that seems increasingly unlikely.

 

[Blackraven]

I had hoped that optical disc technology would be eliminated from the next round of consoles, but that seems increasingly unlikely.

Unlikely because not all countries (unfortunately) even have 1mbps internet speed.....lest even broadband connection at all for that matter (shitty dialup).

So yeah, until then, I can't see game consoles moving to internet downloads (or game streaming) for that matter.