The Console Arms Race: Is This What Console Gamers Want?

Do you like the idea of half-cycle (tick-tock) upgrades and forwards compatibility?


  • Total voters
    75
Sony at this game has an advantage though, if we are talking about strict 100% hardware BC: its 1.84tflops GPU. They can easily put 2 of them and overclock them like supposedly they did. Hopefully they won't forget the Mhz starved Jaguars this time. Also the true unified memory is a huge advantage for 4K gaming.

Microsoft GPU is already weaker and already overclocked at 1.3tflops (so they won't be able to beat Sony by using a SLI setup + overclock) and esram bottlenecked which won't be ideal if devs want to display their games at 4K. It will add complexity in an already complicated setup.

2 PS4 GPUs overclocked at 1000Mhz = 4.6Tflops (less efficient than a normal GPU obviously because SLI config)
2 XB1 GPUs overclocked at 1000Mhz = 3Tflops + added complexity of esram setup.
I don't see either company going for a crossfire or SLI type setup though.

It would be cheaper to use a brand new SOC for both of them with the smaller 14nm process. I would think they would go with a GPU similar to the 390 in ALU count to get better performance and MS would change to either GDDR5 or HBM.

With everything ms has been working on with UWP and forwards compatibility they definitely wouldn't be limited to keeping the same exact Soc architecture or memory setup as the Xbox one.

Who knows, but I doubt either company will just slam an extra duplicate SOC on a new board and call it a day.
 
The risk of process shrinks not hitting when you need to launch a new console is good reason to switch regular-ish refresh model instead. The big whammy model is just broken for all parties, including consumers.

Yes, there's new console excitement but worthwhile next gen games take years to arrive. Spending 600currency (console+games+bits) is just a luxury disposable income spaff. It's rubbish value.

I'd hope that having active new customers who aren't just using their consoles for Netflix might increase diversity in AAA titles. Not sure why I think that though!
 
The risk of process shrinks not hitting when you need to launch a new console is good reason to switch regular-ish refresh model instead. The big whammy model is just broken for all parties, including consumers.

Yes, there's new console excitement but worthwhile next gen games take years to arrive. Spending 600currency (console+games+bits) is just a luxury disposable income spaff. It's rubbish value.

I'd hope that having active new customers who aren't just using their consoles for Netflix might increase diversity in AAA titles. Not sure why I think that though!

No it's not. You buy a console and you know it will play everything released for the next 5 ~ 6 years and that everything released runs the same on your launch models as it does on somebody who bought a model 2 or 3 years in.

Refresh models are going to mean 2 years in that new game you want to play might not run or look that great on your "old" console vs the refreshed model. So what do you do? Buy another 400 YFC console? That is going to get expensive quickly because nobody is going to buy your old console. Sony and MS would probably dumb stock cheap and who wants the old model anyway?

The only positive outcome I can think of is for Sony and MS who might be able to lock you to their ecosystem better because there is no big market reset every couple of years. I'm sure thats good for them, but for me? If you want power buy a pc, that's what the master race is for. Consoles are meant to be a relatively cheap plug and play experience. Having PS4 v1 ~ v5 all offering a different experience isn't cheap nor plug and play.

The whole idea is just dumb. Like I said, want power? Buy a pc.
 
Refresh models are going to mean 2 years in that new game you want to play might not run or look that great on your "old" console
Which is already the case, except everyone's talking about extreme times here instead of realistic times. The notion that a new refresh appears two years after launch and considerably more powerful than the old version is silly. New hardware after 2 years won't be powerful enough or a large enough target to warrant notably improved games. The real refresh cycle is more like 3 years, and at 4 years, where the old console is starting to struggle to run the games devs are wanting to make, you have the option of upgrading. It's absolutely no different to what we have now, other than the transition becoming softer. Early adopters can stay on top. The general populace can pick the hardware that meets their price and performance preference.

In fact it's a major win. Presently poorer folk waiting for console hardware to drop to <$200 are buying an EOL product. A progressive model means their machine can still run the newest games if at reduced quality. Consider something like FIFA, or Diablo 3. The same game on PS3 is actually a different game to that on PS4. The devs have to keep up two software paths, and of course they can't be arsed with the older machine (in the case of D3, Blizzard gave up and don't update any more) and it gets a weaker game. Diablo 4 will run on PS4, and then PS4.5 will run it without the framerate drops when things get really busy, and then PS5 will also run it and Diablo 5, which PS4.5 will also run. If you only have a PS4, you have the option to buy a $400 PS5 to play the best D5 experience, or buy a $200 PS4.5 and still play it.

Consoles are a complete outlier! The more I talk about, the more ridiculous consoles seem. Again, they have their own little universe outside the realms of normal product cycles simply because the nature of technology required it. I can't think of any other product that has these 5 year reboots. The console experience that people are fearing will be ruined isn't really what you think. The console experience is put in disk, play game. People weren't buying NES or Saturn because they could be sure no-one else was playing the same games in better quality - that's a 'reason' those resisting the idea are pulling up just as a counter point. Gamers bought the boxes to play the games. That'll still be the case! They'll just have a couple more options. Which is no different from having the choice of a $300 console from Company A or a $500 better console from Company B.

I can only conclude that the reservations are just basic psychology and fear of change. They really don't add up, and the arguments against the notion are appeals to things that really don't matter. If the boxes play the games, it's good, especially if it makes the job of providing games that much easier. But at this point it's all getting repetitive so I'll refrain from more arguments. ;)
 
The whole idea is just dumb. Like I said, want power? Buy a pc.

Sorry this argument really gets on my tits. I have a PC, it's a laptop / tablet, it sits on my desk at home (I WFH) and I do my work and word processing on it. I really have no inclination to finish work and remain at my desk and play games too.

I like consoles because they sit under my TV, update themselves automatically and I don't have to worry about drivers, or patching games, or mods.

I built my own PCs between the late 90s and sometime in the early 00s. Up until I started office work really. I've had a couple of gaming laptops since, but I now can't justify buying any more. I never get the same satisfaction from games, just spending time getting optimal settings. Now I can just play.

Yes, I know there are guys that love playing with settings and fair play to them. It's just not for me. That time has long since passed. And no, I don't want an HTPC or anything plugged into my TV.

The last computer I got for games was the Raspberry Pi 3 and I've loaded Retropie onto it with a couple of SNES pads. That's the closest I get to gaming on a PC these days and it's perfectly sufficient.
 
I don't see either company going for a crossfire or SLI type setup though.

It would be cheaper to use a brand new SOC for both of them with the smaller 14nm process. I would think they would go with a GPU similar to the 390 in ALU count to get better performance and MS would change to either GDDR5 or HBM.

With everything ms has been working on with UWP and forwards compatibility they definitely wouldn't be limited to keeping the same exact Soc architecture or memory setup as the Xbox one.

Who knows, but I doubt either company will just slam an extra duplicate SOC on a new board and call it a day.


Crossfire would be good to bundle with your VR unit. Use an external second gpu that way you don't need to reduce the quality of your games compared to non vr games.

MS has a partnership with Oculus for xbox one remotes. They also put out a picture of the hololens with oculus touch controllers. Oculus uses Constellation tracking which is just an IR camera so you can use a newer Kinect or just an ir camera.

Sony's early success with preorders at $400-$500 would make a 2017 cv1 on a new xbox easily possible
 
Maybe Microsoft is waiting for 10nm. TSMC is agressively pursuing 10nm bulk production and is targetting 2017 and 2018 for 7nm.
 
The transition from 16/14 to 10nm should be significantly faster than 28nm to 16/14nm. As I think the biggest issue was developing FinFet transistors. Going to 10nm will mean a traditional type of shrink without the need to develop a whole new transistor variety.

However, it won't be that fast for high performance parts (2017). I think you expect some 10nm parts in the mobile space at the end of next year and high performance GPU/CPU type chips following in a year or two after that (~2019).



Perhaps this is one of the contributing reasons to a half-gen upgrade. The time that it would take to offer a traditional console upgrade (10x) would simply be too long for the market to support. 10nm even in 2019/2020 might not be able to provide the traditional performance jump (maybe 4X over what we have now) so instead of waiting for a new gen, upgrades will come out when 2X the performance can be delivered (2017 16nm, 2021: 10nm, 2025: 7nm).
 
Crossfire would be good to bundle with your VR unit. Use an external second gpu that way you don't need to reduce the quality of your games compared to non vr games.

MS has a partnership with Oculus for xbox one remotes. They also put out a picture of the hololens with oculus touch controllers. Oculus uses Constellation tracking which is just an IR camera so you can use a newer Kinect or just an ir camera.

Sony's early success with preorders at $400-$500 would make a 2017 cv1 on a new xbox easily possible
Sony's early success is based on 40 million already having a ps4, millions already having the camera, even 15 millions having moves, and the DS4 being fully trackable. This means millions have a 399 entry price into high end VR.

A new xbox starts from scratch, CV1 is $600, you add motion controllers "below $200", and you need a new more expensive xbox... Cost of entry at launch is not comparable. Those with that amount of disposable income get a PC VR instead. (which is exactly what you did)

Phil Spencer recently said there's is no xbox 1.5 in development, the xb1 is doing well, next gen hardware will have to be substantially more powerful, etc.... It would be a pretty stupid comment to make if they launch a mid-gen power upgrade next year.

http://www.techradar.com/news/gaming/there-won-t-be-an-xbox-1-5-says-phil-spencer-1318199
 
Reasonable posts on this page. More I think about more it appears that this new parrrel/ refresh may actually be very good to consoles. Slowing die shrinks which can't provide generational jumps after six years at that price anymore , publishers not wanting longer than six years gens, few studios closing every time when usebase is reset to zero, mass riots on internetz every time when supposedly there is few frames or pixels in open world games and in reality situation is quite good comparing to past gens, cynicly amplified by so called gaming/technology journalists. Options in console world will help with all of that.
I'm mostly lurking since god knows how many years but I gotta say that there are still some places on the internet where discussion doesen't look like event horizon movie :D
 
Its all in how you read his statements. To be honest, an XO 1.5 would more or less just be a PS4 on the gpu front. While they could double the xbox one GPU and end up roughly 33% more than a PS4. This seems doable while fitting within his statement of major system upgrade or jumps -- its doubling of the GPU of the XO!

XO 12CU * 1.5 = 18CU (same as PS4)
XO 12CU * 2.0 = 24CU (33% more than PS4)
 
Back
Top