Middle Generation Console Upgrade Discussion [Scorpio, 4Pro]

Status
Not open for further replies.
GDDR5X is currently single sourced with only Micron producing it. Getting above 5.5-6.0Gbit ranges, dependant on the device pushes it into a Quad Data Rate mode, which is not as efficient as DDR, which all GDDR5 devices use regardless of the speed.
This is something I'm a bit confused on, how does it actually work? I know that for example 5Gbps GDDR5 does work at 1250 MHz on some level, but it would then need QDR for 5Gbps, no?
 
For anyone wondering they are definitely going to support higher framerates in games.

"the best framerate". - brunette woman
"We can render at 60hz, we can render fully uncompressed pixels" - bald dude
"We are moving Fallout 4 to VR, and to have a console that can support that at the resolution and speed we really want, think its going to be magical " - Game Director at Bethesda Todd Howard
 
Why would sony or ms want to sell consoles for $100 or $200 when they can sell new ones at $400 or $500.

Because the margins on new consoles are certainly going to be much, much smaller. This is the nature of new technology, you want to sell it as low as possible without incurring a loss. The margins on older consoles are going to be fatter unless there is some weird legacy technology that doesn't scale well with time but neither current console have anything like this.

Why would EA or Ubisoft or whever want to throw hundreds of people at a project just to get looking a little nicer when they can throw it on a new machine and up the res , up the fsaa , up the filtering , up the texture quality and so on.

Publishers will continue to focus their effort on the platform with the largest install base. If there are five times as many people who One/PS4 consoles as Scorpio/Neo consoles then they need to get the games working well on the bigger install base becuae they stand to lose more sales through shoddy code putting people off buying the game.

Repeating what I said in the PS Neo thread, Microsoft face the same risk as Sony with Neo that it could be a damp squib if third party developers really don't make much of an effort to utilise the extra power except in ways which require very little effort. I'm sure Tod Howard at Bethesda is super excited at more teraflops but until 6Tf is the baseline for consoles and PC, they can't really target games to rely on that.
 
Because the margins on new consoles are certainly going to be much, much smaller. This is the nature of new technology, you want to sell it as low as possible without incurring a loss. The margins on older consoles are going to be fatter unless there is some weird legacy technology that doesn't scale well with time but neither current console have anything like this.

Current consoles have consistently received price cuts. I doubt the margins are much better at all esp in the case of something like the slim


Publishers will continue to focus their effort on the platform with the largest install base. If there are five times as many people who One/PS4 consoles as Scorpio/Neo consoles then they need to get the games working well on the bigger install base becuae they stand to lose more sales through shoddy code putting people off buying the game.

Repeating what I said in the PS Neo thread, Microsoft face the same risk as Sony with Neo that it could be a damp squib if third party developers really don't make much of an effort to utilise the extra power except in ways which require very little effort. I'm sure Tod Howard at Bethesda is super excited at more teraflops but until 6Tf is the baseline for consoles and PC, they can't really target games to rely on that.

Developers have supported new hardware in the past which is why we aren't all still on Atari systems. Developers are also targeting video cards of higher power already. So the dev support is there.

Your going to see current systems start to get shoddy ports by 2018 . Just wait till you see the neo versions of the ps vr games.
 
I overall agree with you
I want to add that we are reaching the limit for games budgets and the amount of man hours that can be sunk in games at their current retail price. There is also the matter of maximizing the addressable market. I think that there won't be massive improvements in the upcoming years and pretty much evey bits of extra power are going tonbe spent on resolution or improved framerate.
 
Last edited:
According to the info available Scorpio DOES seem to be more powerful than Neo.

Probably this is why Sony decided to not announce anything this E3
http://www.eurogamer.net/articles/digitalfoundry-2016-xbox-one-project-scorpio-spec-analysis


Humm... did Microsoft decide to announce the 6TFLOPs + 320GB/s soon enough for Sony to be in time to change the Neo's specs?
I imagine that if the Neo had rather conservative GPU and memory clocks (especially for FinFet on the GPU side), maybe they can still change the specs. For example, they could change the GPU clocks from 910MHz to Polaris 10's ~1200MHz and GDDR5 clocks to 8GT/s, effectively boosting its specs to 5.5 TFLOPs and 256GB/s.
Despite possibly creating a console with a substantially higher power consumption, the practical difference to Scorpio would be marginal. Remains to be seen if either of them is changing the CPU cores from Jaguar/Puma to Zen (I don't think they would, though).
 
I imagine that if the Neo had rather conservative GPU and memory clocks (especially for FinFet on the GPU side), maybe they can still change the specs. For example, they could change the GPU clocks

But then again, so could MS if they are launching Scorpio in Fall/Holiday 2017. If we believe the rumours Neo will be out later this year (before October?) so it should be in production by now.
 
Your going to see current systems start to get shoddy ports by 2018 . Just wait till you see the neo versions of the ps vr games.
and by then how long would the xo had been out for?
For example, they could change the GPU clocks from 910MHz to Polaris 10's ~1200MHz and GDDR5 clocks to 8GT/s, effectively boosting its specs to 5.5 TFLOPs and 256GB/s.
I think if Sony was in the habit of using the same cooler as xo, then maybe so, but I think it will be harder to just ramp it up so much without acoustic issues, and maybe cost of having to use less chips rated for it?
 
But then again, so could MS if they are launching Scorpio in Fall/Holiday 2017. If we believe the rumours Neo will be out later this year (before October?) so it should be in production by now.

Seems unlikely to do a launch this fall without an announcement at e3. I don't know. Are there any other major gaming events this summer, or some kind of special playstation event?
 
According to the info available Scorpio DOES seem to be more powerful than Neo.

Probably this is why Sony decided to not announce anything this E3
http://www.eurogamer.net/articles/digitalfoundry-2016-xbox-one-project-scorpio-spec-analysis

Probably why Sony essentially went to radio silence after MS leaked the specs of the scorpio.

From your article:
. Sony's technological advantage will be gone with the next wave of hardware - we already know that it cannot support true 4K resolution on cutting-edge games, because we've seen the internal documents that outline Sony's upscaling strategies for 4K display support (more on that soon). It's also unfeasible for Sony to produce a radically revised Neo - the silicon has been designed, developer kits have gone out. Matching Scorpio would require scrapping Neo's existing processor completely.
 
If DF is right in their speculation, I'd be disappointed. It looks like the resolution warriors won after all their complaining. A netbook CPU from 2013 paired with Vega technology releasing in 2017. Lovely.
 
But then again, so could MS if they are launching Scorpio in Fall/Holiday 2017.
They could, but whereas 4 TFLOPs seems rather conservative for a 14FF GPU, 6TFLOPs does not. That would make me think Sony has a bit more headroom than Microsoft.

If we believe the rumours Neo will be out later this year (before October?) so it should be in production by now.[/QUOTE]

Then again, the Neo wasn't shown at E3.
They wouldn't show the new console 3 months before its release? Maybe the fact that they haven't shown it means it's not coming in October after all?


I think if Sony was in the habit of using the same cooler as xo, then maybe so, but I think it will be harder to just ramp it up so much without acoustic issues, and maybe cost of having to use less chips rated for it?

If I was Sony, I'd probably prefer to launch my mid-gen console one or two quarters later to get time to change enclosure+cooling+power regulation+whatever than to live the whole mid-gen mid-cycle with a competitor that is ~50% faster, where the IQ difference in all multiplatform games are clearly visible (as proven by the recent generation).
Sony is already winning the hardware race by a lot. They are in a lesser rush to release their mid-gen, much less than Microsoft.




As for this statement:
It's also unfeasible for Sony to produce a radically revised Neo - the silicon has been designed, developer kits have gone out. Matching Scorpio would require scrapping Neo's existing processor completely.
Well not if all they're doing is upgrading the clocks. If Microsoft could change the Xbone clocks within weeks of launching the console, Sony would probably be good to do it half a year (or more?) before.[/QUOTE]
 
Well not if all they're doing is upgrading the clocks. If Microsoft could change the Xbone clocks within weeks of launching the console, Sony would probably be good to do it half a year (or more?) before.

That was specifically addressed in the article. Essentially, MS had the legroom to increase their clocks because of their cooling solution and their external PSU. In short, the very things that PS4 consumers hate would need to be implemented (larger console, external power brick), and even then, the article states the boosts wouldn't make any more of a difference in the end comparison as they did when MS did it for the XB1.
 
They could, but whereas 4 TFLOPs seems rather conservative for a 14FF GPU, 6TFLOPs does not. That would make me think Sony has a bit more headroom than Microsoft.

Unless MS is going with a bigger GPU, cut down Vega at very modest 800-850 MHz like DF suggested.

They wouldn't show the new console 3 months before its release? Maybe the fact that they haven't shown it means it's not coming in October after all?

Either way, it looks like they best can do right now is increase clockspeeds, something not out of reach for MS as well (even if they have a little less headroom). They could completely scrap it, but I don't see it happening. A $50-100 cheaper Neo will sell better than Scorpio regardless of tech specs, and launching almost a year before helps as well.

DigitalFoundry said:
It's also unfeasible for Sony to produce a radically revised Neo - the silicon has been designed, developer kits have gone out. Matching Scorpio would require scrapping Neo's existing processor completely.
 
The wording yesterday in the Scorpio reveal was carefully chosen to address all the concerns of Xbox supporters, power, compatibility, etc and also to emphasise the leaked weak points of the NEO.

All the games will be compatible, 4k and VR without compromises, and DEVS could take advantage of the extra power.

Will it force the hand of Sony?? to loosen its requisites??
 
Status
Not open for further replies.
Back
Top