Next-gen PS5 and next Xbox speculation launch thread |OT5| - It's in RDNA

What do you think could be the memory setup of your preferred console, or one of the new consoles?


  • Total voters
    1,051

Papacheeks

Member
Oct 27, 2017
3,132
Watertown, NY
The problem with VR is how expensive VR sets are. They will never become mainstream unless the cost factor comes down. It is easier to put big money on a project that will be available to tens of millions or hundreds (given the stance Microsoft has taken on PC) than it is to put money on an extension that may not see 10 million in sales.

Many have an issue paying full price for a console......how much more when they have to pay $400 on a device that does not gain as much in quality content?
I would argue Sony's investment in dedicated AA-AAA studios working on VR titles has shown the bar in terms of quality VR experiences software wise. Something PC side lacks a little in terms of big budget VR.

Them selling over 4 million in 2.5 years is nothing to scoff at either especially compared the the competition which had been out a lot longer, and they handidly almost sold more then all of them combined.

VR is going to get better, cheaper, and easier to use. The new CPU'S,GPU'S that are coming out will be able to handle VR ad more affordable prices, and the new consoles will give higher IQ fidelity for VR on consoles along with the newer tech for controls, sensors that come with it.

It's still niche for sure, but it's growing, and isn't going away. It's just still germinating until the headsets can be used more seamlessly without tons of wires and more accurate tracking/control.
 
Mar 23, 2018
1,541
I think "One More Time" Phoenix Splash has to do the OP I suppose ...
Lyrics by Daft Punk
That was a close call LOL

Really nice OP and the wordplay... that's good. That's really good. Thank you for making it!

I've been out of the Xbox loop for a while - Is it feasible they announce/reveal their new hardware next week at E3?
Yep, at least as a teaser like back at E3 2016 with Project Scorpio


Isn’t the thread title wrong? ”Next Xbox” is correct but “next gen PS5” is kindda wrong. It should be “Next gen PS”
PS5 alone stands for the next gen part, we don’t have a current gen PS5!
Now that you mention it... you have a point. I forgot about applying some kind of puntuaction like: "Next-gen: PS5 and next Xbox". You can put that one on me! LOL
 

Wandu

Member
Oct 27, 2017
1,617
Isn’t the thread title wrong? ”Next Xbox” is correct but “next gen PS5” is kinda wrong. It should be “Next gen PS”
PS5 alone stands for the next gen part, we don’t have a current gen PS5!
Now that I think about it, I will be mad if Sony names the next gen PS anything other than PS5. I want the consistency! Oh well, I will probably call it PS5 anyway.
 

Socky

Member
Oct 27, 2017
294
Manchester, UK
Just to return to the discussion about MS seeking to profit, or not, from Scarlett, I can't see how they can aim to profit from Anaconda, since doing so makes the PS5 much more likely to match it with a loss at launch.

For example, if the Anaconda total cost price is say $480, with RRP of $499 (I'm ignoring retailer mark-ups, etc for simplicity), giving them a slight profit of $19, Sony could potentially go for a loss of $81 with a cost price of the same $480, but an RRP of $399, which negates Anaconda's power lead and leaves it $100 more expensive than PS5.

There are many more variations possible of course, but any profit for MS on Anaconda makes PS5 more likely to match it in power levels, if Sony take a (likely) loss on it. I think it more likely MS will take a loss on both, since they can hardly make much profit on Lockhart without pushing it closer to the PS5 RRP. It's going to be interesting...
 

gremlinz1982

Member
Aug 11, 2018
1,111
I would argue Sony's investment in dedicated AA-AAA studios working on VR titles has shown the bar in terms of quality VR experiences software wise. Something PC side lacks a little in terms of big budget VR.

Them selling over 4 million in 2.5 years is nothing to scoff at either especially compared the the competition which had been out a lot longer, and they handidly almost sold more then all of them combined.

VR is going to get better, cheaper, and easier to use. The new CPU'S,GPU'S that are coming out will be able to handle VR ad more affordable prices, and the new consoles will give higher IQ fidelity for VR on consoles along with the newer tech for controls, sensors that come with it.

It's still niche for sure, but it's growing, and isn't going away. It's just still germinating until the headsets can be used more seamlessly without tons of wires and more accurate tracking/control.
No different to Kinect games that sold a lot of copies. It is a risk reward based business, and most publishers will want to put money in an area that allows them to canvas as many users as possible to try and minimize risk. Platform holders like Microsoft and Sony can take those risks by leveraging platform, royalties and services. That is not something that is open to all publishers.....and so long as VR headsets are damn expensive, it will remain a niche market.
 
anexanhume's performance and die-size estimation

anexanhume

Member
Oct 25, 2017
3,920
Reposting from last thread

Indeed. But is there any chance we can derive the number of stream processors from the die size? (estimation)
Or do we believe the GPU sports 3072 stream processors?
We can try!

We have two unknown variables unfortunately. Clocks and CU count. For reference, I’ll be using Strange Brigade benchmarks from here.

The knowns are:

RTX 2070 + 10% performance in Strange Brigade at 4K. This puts it within a few % of Vega 64, so let’s call them equal for simplicity’s sake.

Architecture gain of 1.25x per clock based on a suite of 30 benchmarks at 4K. This is a good comparison because it’s more likely to stress any memory bandwidth disparities.

Perf/Watt gain of 1.5x over GCN at 14nm. I’ll assume this is Vega 64 and immediately discard the metric. Why? Because we already know Vega 20 enjoys a 1.25x perf/Watt boost over Vega 64, so this is AMD admitting Navi is running at some clock where there are no additional perf/Watt advantages.

I think we should assume a minimum of 40 CUs based on the various leaks, and no more than 52 based on AdoredTV’s numbers.

Vega 64 has a 1250MHz base clock and 1550MHz boost. To draw equal, Navi must make up any CU deficiencies not overcome with the 1.25X factor by clocks. This boosts 40CUs to an effective 50, meaning a 64/50 ratio boost to clocks. 1,600MHz base clock. 1984MHz boost clock. These don’t seem totally far fetched given AMD says Navi clocks better, and Nvidia designs can clock that high.

44CUs: 1450MHz base, 1800MHz boost.
48CUs: 1333MHz base, 1650MHz boost.
52CUs: 1250MHz base, 1550MHz boost. (No change)

What’s also interesting in my mind is CU sizing. If CUs have grown a lot, it really speaks to a lot of architecture rework. Vega 20 fits 64 CUs and a 4096 bit HBM2 controller in 332mm^2. I think it has more negative die area than strictly needed, and my personal belief is because this could have been a hurried refresh as a pipe cleaner, as well as current and power density meant it couldn’t be shrunk further due to IR loss and heat density concerns. Navi is sporting a 255mm^2 die.

Navi has a 256-bit GDDR6 interface, and we know per 128 bits, it’s 1.5-1.75x larger than a 1024 bit HBM2 interface. Let’s assume Navi’s 256 bit and Vega 20’s 4096-bit are roughly equal, rather than GDDR6 being 20-25% smaller. I do this because I assume Navi will have less negative die area.

That means the rest of the area should be roughly equal, and so we can do an approximate CU sizing.

40 CUs: Navi CUs are 23% larger than Vega 20.
44 CUs: 12%
48 CUs: 2%
52 CUs: -5%

In any event, 255mm^2 is a good sign for consoles being able to include a full Navi 10 die along with 70mm^2 Zen 2 design, with some spare room for glue logic and misc. IO. If that leaked dev kit PS5 rumor is true (312mm^2), we’re clearly dealing with a cut down Navi 10 (or a LITE version with smaller CUs?)

Which outcome is better for consoles? I would argue the smaller CU is actually better for consoles, because it makes the clock situation a lot more favorable. I think the 52CU scenario is infeasible because there’s no way AMD would market a GPU with clocks that low and make the statements they did. I think we are likely looking at 44CUs for Navi based on it giving us a 1800MHz boost, which lines up perfectly with Radeon 7, and gives up the 0% perf/Watt advantage of Navi over Vega 20 that we expect. Of course, 40CUs is the best case if you believe the Gonzalo leak, because it tells us that console GPU clocks are actually 10% under desktop GPU boost clocks.

RX 580 clocks are 1257MHz/1340MHz, which means Xbox One X comes within 7% of desktop base clocks and 13% of boost clocks, and so I think Gonzalo clocks are completely believable for a 40CU Navi with ~2000MHz boost clock. They are far-fetched for anything beyond 44CUs.

And drum roll, teraflop time!

Given the clocks are scaled based on CU count, all above configurations have the same metrics. 8.2TF bass, 10.1TF boost. This puts us right in the TF band we expect for consoles (if a bit on the low side). I suspect the RX 5700 is not the top end Navi SKU though (I would expect an 8 or 9 in the name), and there’s probably a full die version with 4-8 more CUs enabled, meaning all the above calculations are going to move up. Given consoles will most likely disable CUs for yield, this may absolutely still be a comparable situation. Conclusion, I remain team 10TF, but they punch like 12.5TF.
 

Papacheeks

Member
Oct 27, 2017
3,132
Watertown, NY
No different to Kinect games that sold a lot of copies. It is a risk reward based business, and most publishers will want to put money in an area that allows them to canvas as many users as possible to try and minimize risk. Platform holders like Microsoft and Sony can take those risks by leveraging platform, royalties and services. That is not something that is open to all publishers.....and so long as VR headsets are damn expensive, it will remain a niche market.
Kinect was a dead device though. In terms of it's uses in strictly motion control games. If it was going to be used with hololens or something Microsoft was developing for VR we would have at least heard of something or rumors for developers making games for it. They didn't invest in VR, they went a different route, and wanted a Alexa type of console device.

Now VR is getting more accessible with all in one solutions from Samsung among others actually being good. Only a matter of time for the software to be there that makes a case for a casual buyer to jump in.

Playstation is at the for front of that with dedicated experienced developers that can make more refined/higher budget VR experiences compared to the competition.

Publishers have been putting their games into vr though? Bethesda, Square enix, Disney?

If Iron man VR does super well, your going to see more of that with big properties. Astrobot just proved you could do a mario type of game in VR. Would not be surprised if this year or going into next year we get Battlefield/Call of duty VR. Like play the entire game in VR.

It's coming for sure.
 

VX1

Member
Oct 28, 2017
3,678
Europe
anexanhume i know you are a big fan,but now with more Navi info we have, should we be more skeptical about possible HBM2 in next gen consoles/PS5?
 

gofreak

Member
Oct 26, 2017
2,194
It could just mean better performance at basic, but some people were thinking full RDNA will have ray tracing, which might makes it's way into PS5/X2.
I guess this is a matter of semantics, but I think we'll probably end up thinking about AMD's ray tracing tech as orthogonal to the base architecture, in the same way as it latterly turned out with Turing and RTX on the nVidia side. The base architecture on one side, the ray tracing tech as a bolt-on on the other.

If ray tracing was the only big difference with the 2020 chips, it wouldn't change much about the core RDNA microarchitecture's relationship to GCN.
 

Hoo-doo

Member
Oct 25, 2017
2,912
The Netherlands
No different to Kinect games that sold a lot of copies. It is a risk reward based business, and most publishers will want to put money in an area that allows them to canvas as many users as possible to try and minimize risk. Platform holders like Microsoft and Sony can take those risks by leveraging platform, royalties and services. That is not something that is open to all publishers.....and so long as VR headsets are damn expensive, it will remain a niche market.
Comparing VR to Kinect is some wild shit. They are nothing alike.
Kinect bled out because it never matured as a technology. It was an inferior input method compared to regular controllers in nearly every way and only got hold because of the novelty factor. The input lag and imprecision alone made sure that it was never going to be a peripheral that would support actual AAA games.

VR is completely different in that regard. The only thing against it is that it’s too expensive and still a bit cumbersome as of right now, but even that will change in couple of years.
 

Nobunny

Avenger
Oct 25, 2017
671
I dig the thread title.

May this new thread be filled with juicy non-pastebin based leaks.
 

RevengeTaken

Member
Aug 12, 2018
1,267
Your expectations are too freaking low, an 8-10 teraflops machine in late 2020 is pure disappointment when we already had that power in desktop GPUs for 3-4 years back. ~12tflops Navi + 24GB of super fast memory + 8c/16t Zen2 at 3.2Ghz + 500GB Solid State Drive is the minimum specs you've to expect from a console at 7nm node in late 2020 for $449 to $499!

But that's all good cause when MS or Sony announces over 12tflops console you're gonna shit in your pants in spite of your low expectations!
 

ClarkusDarkus

Member
Oct 27, 2017
1,684
Kinect was a dead device though. In terms of it's uses in strictly motion control games. If it was going to be used with hololens or something Microsoft was developing for VR we would have at least heard of something or rumors for developers making games for it. They didn't invest in VR, they went a different route, and wanted a Alexa type of console device.

Now VR is getting more accessible with all in one solutions from Samsung among others actually being good. Only a matter of time for the software to be there that makes a case for a casual buyer to jump in.

Playstation is at the for front of that with dedicated experienced developers that can make more refined/higher budget VR experiences compared to the competition.

Publishers have been putting their games into vr though? Bethesda, Square enix, Disney?

If Iron man VR does super well, your going to see more of that with big properties. Astrobot just proved you could do a mario type of game in VR. Would not be surprised if this year or going into next year we get Battlefield/Call of duty VR. Like play the entire game in VR.

It's coming for sure.
Don't know why people bring up Kinect, That was a motion sensor/controller at best, VR is a brand new way to play a game like RE7 as opposed too a flat TV, Kinect couldnt offer that as your still looking at the same screen. VR is and will be the most exciting tech in gaming.

Oculus Quest just released and it's completely wireless, And yes it sold out in many places. PSVR2 will do well to learn from the Quest as the only thing holding PSVR back now is cables and the moves, PS5 supporting PSVR is huge too. As PSVR will get cheaper and cheaper to purchase and jump into VR so the price argument is silly when people lash out hundreds/thousand for new GPU's or smartphones each couple of years.
 
Colbert's Next Gen Predictions

Colbert

Member
Oct 27, 2017
3,531
Germany
New thread - old habits

I repost my last recent prediction - bookmark it as I will update this prediction every once in a while ...



Remarks/Assumptions:
  • Clock speeds derived from TBP of Navi PC GPUs
  • RDNA Cache architecture allows for lower VRAM bandwidth
  • Both platform holders have the chance to wear the performance crown
  • Expecting an increase in game performance at the same computational power (TFs) just not by the number itself by at least 15%
Change Log:
Rev 9.0: Incorporated E3 Info from MS, AMD and some secret sauce from my side that will not be disclosed (14-06-2019)
Rev 8.2: Lower clock for the lower bounds . (27-05-2019)
Rev 8.1: Eliminate the difference between PS5/Anaconda but having 2 tier prediction for those consoles
Rev 8: New NAVI architecture makes CUs obsolete. Consolidated to ECC vote as more likely now after Computex info (27-05-2019)
Rev 7.1: Included the Endless Cycle Committee prediction vote into the table after a RfC (25-05-2019)
Rev 7: Completely changed prediction scheme: Lower and Upper bounds provide the range I expect the consoles will land (16-05-2019)
Rev 6.1: Changed from Baseline to Ballpark, Modified specs based on new information available to me
Rev 6: Consolidated to 1 baseline per console, adjusted specs & pricing for all consoles
Rev 5: Added 3rd tier to adapt on AdoredTV table of Navi GPUs
Rev. 4.3: SSD for Lockhart, Lockhart now $399 instead of $349
Rev. 4.2: Increased Memory clock instead of 448Gbps,
 
Last edited:

JahIthBer

Member
Jan 27, 2018
3,156
Your expectations are too freaking low, an 8-10 teraflops machine in late 2020 is pure disappointment when we already had that power in desktop GPUs for 3-4 years back. ~12tflops Navi + 24GB of super fast memory + 8c/16t Zen2 at 3.2Ghz + 500GB Solid State Drive is the minimum specs you've to expect from a console at 7nm node in late 2020 for $449 to $499!

But that's all good cause when MS or Sony announces over 12tflops console you're gonna shit in your pants in spite of your low expectations!
The problem is Desktop GPU's have slowed down to a crawl, GTX 1080 is still high end in 2019. This effects Consoles too.
 
Oct 27, 2017
2,803
Somewhere South
Given the clocks are scaled based on CU count, all above configurations have the same metrics. 8.2TF bass, 10.1TF boost. This puts us right in the TF band we expect for consoles (if a bit on the low side). I suspect the RX 5700 is not the top end Navi SKU though (I would expect an 8 or 9 in the name), and there’s probably a full die version with 4-8 more CUs enabled, meaning all the above calculations are going to move up. Given consoles will most likely disable CUs for yield, this may absolutely still be a comparable situation. Conclusion, I remain team 10TF, but they punch like 12.5TF.
So, you'd guess the RX5700 die has something like 48 CUs (or equivalent, 3072 SPs) with some 4 or 8 CUs disabled for yields. That's a pretty good guess, IMO. Leaves room for more one or two higher tier SKUs ( if 57K is 40CU, maybe 44 and 48; possibly with pushed clocks X versions).
 

gremlinz1982

Member
Aug 11, 2018
1,111
Kinect was a dead device though. In terms of it's uses in strictly motion control games. If it was going to be used with hololens or something Microsoft was developing for VR we would have at least heard of something or rumors for developers making games for it. They didn't invest in VR, they went a different route, and wanted a Alexa type of console device.

Now VR is getting more accessible with all in one solutions from Samsung among others actually being good. Only a matter of time for the software to be there that makes a case for a casual buyer to jump in.

Playstation is at the for front of that with dedicated experienced developers that can make more refined/higher budget VR experiences compared to the competition.

Publishers have been putting their games into vr though? Bethesda, Square enix, Disney?

If Iron man VR does super well, your going to see more of that with big properties. Astrobot just proved you could do a mario type of game in VR. Would not be surprised if this year or going into next year we get Battlefield/Call of duty VR. Like play the entire game in VR.

It's coming for sure.
The only way casuals jump in is if the price makes sense. It doesn't for the time being, not even to the hardcore gamer. Sony is possibly the biggest seller, and sales for PSVR considering the user base is nothing short of disappointing.

How many publishers are putting an effort into VR? The vast minority. And they mostly do not care because the numbers are low.

Comparing VR to Kinect is some wild shit. They are nothing alike.
Kinect bled out because it never matured as a technology. It was an inferior input method compared to regular controllers in nearly every way and only got hold because of the novelty factor. The input lag and imprecision alone made sure that it was never going to be a peripheral that would support actual AAA games.

VR is completely different in that regard. The only thing against it is that it’s too expensive and still a bit cumbersome as of right now, but even that will change in couple of years.
It doesn't matter what the tech is, what matters is whether you can sell enough and make money to justify the investment. Microsoft tried pushing it through in this generation and it failed.

So today, you have VR solutions that are ridiculously expensive, have low support.......it is a niche product. If someone invests in it well and good, but I wouldn't lose sleep if all Microsoft (and to a greatly lesser degree Sony) did was to have a way in which other VR headsets could work with their consoles.
 
Nov 12, 2017
2,877
Well a different thread is entertaining the rumour that Scarlet wouldn't have any first party next gen exclusives on release because all its games need to run on Xbox One hardware. But I guess we can leave that out unless it gets confirmed later.
Hint / That is different from saying those games will not be enhanced in their next gen version ;))))
 

Colbert

Member
Oct 27, 2017
3,531
Germany
So, you'd guess the RX5700 die has something like 48 CUs (or equivalent, 3072 SPs) with some 4 or 8 CUs disabled for yields. That's a pretty good guess, IMO. Leaves room for more one or two higher tier SKUs ( if 57K is 40CU, maybe 44 and 48; possibly with pushed clocks X versions).
From what I understood the RX5700 will be the top end card with RDNA v1 for 2019. The smaller card/GPU is rumored to be 160mm^2.
 

op_ivy

Member
Oct 25, 2017
3,451
New Thread Checklist:

Navi is Sony exclusive
SSD's are a game changer
Xbox is using Vega
2 SKU's is a mistake
Streaming won't take off
Pastebin is the law
This thread will tera-flip flop for the next year and a half

Did I miss anything? :)
how could you leave off my secret sauce of choice, xbox is using zen 3!
 

Gamechanger87

Member
Oct 25, 2017
444
I was wondering when the horseshit spreadsheet prediction would turn up.
Just because his predictions which he actually puts thought into don’t fit your narrative/trolling that PS5 will be the most powerful and the Xbox will be shit doesn’t mean you have to crap all over other people.

At some point it would be nice if your constant console warring would garner a warning.
 
Oct 27, 2017
1,575
That new thread smell.

Looking forward to getting more concrete hardware news, hopefully sooner rather than later.
 
Oct 26, 2017
4,146
United Kingdom
RDNA will evolve, like GCN did.

Maybe RDNA will be at 1.1 next year - but if 1.0 is actually a close-ish descendant of GCN (which wouldn't be too surprising), then I'm not sure 1.1 will necessarily differ substantially in that regard. It begs the question of what 'full RDNA' means - is it full when we're at 1.3 or 1.4 in a few years time? Or when we're at 1.1 in a year? Who decides?

(Which is all to say, these labels might be a bit arbitrary. And I think we kind of just have to take AMD's labeling and decisions about where to demarcate things at face value.)
Agreed. Your suggestion is much more likely.

Love dat new thread smell.

All hail team 18-20 TF!!
Lol... anarchist.
 
Oct 27, 2017
548
Just because his predictions which he actually puts thought into don’t fit your narrative/trolling that PS5 will be the most powerful and the Xbox will be shit doesn’t mean you have to crap all over other people.

At some point it would be nice if your constant console warring would garner a warning.
Oh please.

I've not said anything about Xbox in this thread.

It's not me that's attempting to portray something that isn't there in my messages.
 

BreakAtmo

Member
Nov 12, 2017
3,118
RDNA will evolve, like GCN did.

Maybe RDNA will be at 1.1 next year - but if 1.0 is actually a close-ish descendant of GCN (which wouldn't be too surprising), then I'm not sure 1.1 will necessarily differ substantially in that regard. It begs the question of what 'full RDNA' means - is it full when we're at 1.3 or 1.4 in a few years time? Or when we're at 1.1 in a year? Who decides?

(Which is all to say, these labels might be a bit arbitrary. And I think we kind of just have to take AMD's labeling and decisions about where to demarcate things at face value.)
On this subject, what was GCN 1.0 like compared to the final iteration of the previous architecture?
 

Colbert

Member
Oct 27, 2017
3,531
Germany
Colbert - Thank you for posting the change log. Pertaining to number 7, is there any info pertaining to how CUs are made obsolete in Navi?
I think there will be still CUs as the plot thickens RDNA in its first form will be a hybrid of RDNA and GCN. However when making the prediction this was not clear. It is anyway better to just focus on the stream processors as it makes it easier compare to NVIDIA GPUs as well. I can repost my TF table which also contains a CU to stream processor conversation.