Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]

Status
Not open for further replies.
Because you think if Games were made only for Xbox One X or even PS4 Pro they will look the same?
Yes they would pretty much look the same.
There not held back by the base consoles there held back primarily by there CPU's, that's the ball and chain.
But they was designed for specific tasks and fulfills it well.

Early cross gen games graphically won't be held back much either.
Where they would be held back is in scope, complexity, AI, physics, gameplay, etc.
So will cross gen games be held back, yes.
Will cross gen games be held back hugely graphically no, not at the start, not with how flexible most engines are.

Don't think they should be mandated to make cross gen games though. Wonder if Windows central heard there was a mandate, or it just happens that most are. Which wouldn't surprise me.
 
Last edited:
Yes they would pretty much look the same.
There not held back by the base consoles there held back primarily by there CPU's, that's the ball and chain.
But they was designed for specific tasks and fulfills it well.

Early cross gen games graphically won't be held back much either.
Where they would be held back is in scope, complexity, AI, physics, gameplay, etc.
So will cross gen games be held back, yes.
Will cross gen games be held back hugely graphically no, not at the start, not with how flexible most engines are.
If the engines are very flexible how could cross-gen games be held back graphically years later?
 
I wonder how much MS has changed.... if at all?
They only time MS won, or got close to "winning" a console generation was when they were perceived to have superior hardware.

I find it hard to think that they will not fall back on this approach for next-gen.
I know it's a very different MS, including leadership and inside the XBox div itself.
but it must be hard for a corporation not to go back to what works.
MS have historically had better services attached to their consoles, especially at the start of each gen.

Sony have typically been a bit less "fastest hardware" aware and often focused more on custom features.
If we apply the same logic to next-gen ( past performance is no indicator of future results*** - but this IS the baseless rumor thread)

I would expect MS to go all out on a faster/Bigger/Better GPU + memory+subsystems + services
While I would expect Sony to go for a console with reasonable GPU increase 8-10 TFLOPS range, and focus on the the "SSD" and loading times as a unique feature.


As for which approach is best I guess that the "biggest/bestest hardware" approach form MS will include some form of SSD like solution anyway.
How this compares to what Sony does, and IF Sony are able to offer unique games/features/functionality, then i would expect them to "win" next-gen.
So long as all next-gen consoles can tick the RT Raytracing Box, the RT performance wont matter. The details are very difficult to see, and beyond the Yes/No
Level of support most consumers wont care.

But sadly all the recent rumors appears to contradict, what i thought was good logic!!! Who knows :(
 
If the engines are very flexible how could cross-gen games be held back graphically years later?
Held back was extreme.

Tooling and engine development takes time though, so it's easier to just add those as supplements at the start.
Engine research, where you look at new ways of doing things wont happen overnight.
May decide to approach a graphical feature differently now that have a decent CPU. But again wouldn't expect that right out of the gate.

This time the GPUs are a lot more of an evolution than previous generations, but still lots for them to research, from gpu to system as a whole to get teeth into.
 
wonder how much MS has changed.... if at all?
They only time MS won, or got close to "winning" a console generation was when they were perceived to have superior hardware.

I find it hard to think that they will not fall back on this approach for next-gen.
I know it's a very different MS, including leadership and inside the XBox div itself.
but it must be hard for a corporation not to go back to what works.
MS have historically had better services attached to their consoles, especially at the start of each gen.
What makes you think their not trying to have the most powerful console?
 
If the engines are very flexible how could cross-gen games be held back graphically years later?
Flexibility means removing graphical features to run the game. So, um, how could they not be held back graphically if the engine is having to scale down? If Anaconda can raytrace realtime GI and Lockhart can't because it's not fast enough, and the engine scales down to a baked solution, it's being held back graphically. Most importantly, engines that scale don't do so optimally. You either have a low target and throw more effects on top to use higher spec hardware, or you have a high target and just remove features until you can run on the low end. But you don't take a machine that's not the target and then massage your engine to maximise its use.

It's the difference between a bad port and a good port to NSW. A bad port takes the game and hacks it down to fit. A good port rewrites the game to the fixed spec to run well.

I wonder how much MS has changed.... if at all?
They only time MS won, or got close to "winning" a console generation was when they were perceived to have superior hardware.
During the PS360, I don't think MS was ever perceived as having the superior hardware. PS3 had the power of the Cell and all that. And MS only really 'won' because Sony screwed up. The time MS properly went with the most powerful box, the OXB, they didn't win or even come close to competing.

'Most powerful console' rarely wins; it's a whole bunch of factors. Which is a discussion held many times before and isn't for this thread! Suffice to say, either company may or may not have the most powerful hardware. There's no particular precedent or business case why one should be more invested. We also have no idea what the prices of these boxes are. What if the console that's 10% weaker is half the price...
 
Last edited:
XB360 launched one year earlier. Then It had Oblivion one year before PS3. And some great exclusives like (the then innovative) Gears of wars and eventually the still fresh Halo games. And Live stuff.

But most importantly: it was like twice less expensive than PS3 ! :runaway:

No wonder MS is going to try that again with Lockhart. Well, without the "very powerful console" aspect. :nope:
 
If Anaconda can raytrace realtime GI and Lockhart can't because it's not fast enough, and the engine scales down to a baked solution, it's being held back graphically.
I tried to make this same example in the other thread, but realized it does not really hold. All RT approaches we have seen so far scale per pixel, so low res Lockhart == high res Anaconda.
Even for algorithms that do not work per pixel, like mine, it's the same principle - higher perf means just higher lightmap resolution / less lag, and lower screen resolution is fine with lower light map resolution as well.
The only example of holding back would be a PS5 game, running 1080p at 30 fps with magic path tracing or something. But even then you'll find ways to scale it down without falling back to entirely different tech like baked lighting.
There may be scaling issues with something like Unreal Engine that also targets mobiles, Switch etc, but any next gen engine should be able to scale from Lockhart to Anaconda/PS5 without compromise. There also is good automated asset processing now helping with the content.

Little extra work, huge benefit? I like the idea. Little boy begging for XBox for christmas is more likely to get one this way.
That's for sure, but 'holding back' may be just an assumption.
 
I tried to make this same example in the other thread, but realized it does not really hold. All RT approaches we have seen so far scale per pixel, so low res Lockhart == high res Anaconda.
RT may be an unrealistic example, but the idea is still the same. Control on consoles is 'held back' versus Control on an RTX2080Ti because it can't do the same things. Control on an RTX2060 is 'held back' versus the 2080 because you have to disable a lot of features to get a decent framerate. Although I guess if the difference is between Lockhart being 1080p and Anaconda being 4K, then they should be the same as long as everything else is. RAM could be a little less for smaller textures. CPU would need to be the same. If the CPU is half sized too, then the games will be compromised on one machine or the other.
 
If the CPU is half sized too, then the games will be compromised on one machine or the other.
Agree, but lower clock / disabled SMT would be 'acceptable'. I do not believe quad core rumors.

But how is Control held back for 2080 users, just because 2060 users need to disable stuff?
A working example would be if Control had full dynamic lighting, but having 2080 as minimum spec. But i'm totally convinced even this could scale, if done right.
 
i'm totally convinced even this could scale, if done right.

Everything can scale if done right, even doom runs on nsw. Halo infinite could be one of the best looking rt games of the next gen on scarlett, with a scaled down version on lockhart. HZD ps4 can probably be scaled to switch with some effort. If the market is wider it could be worth the effort/resources.
 
Everything can scale if done right, even doom runs on nsw. Halo infinite could be one of the best looking rt games of the next gen on scarlett, with a scaled down version on lockhart. HZD ps4 can probably be scaled to switch with some effort. If the market is wider it could be worth the effort/resources.
Will high-end consoles graphic be held back by low-end consoles?

Indeed most visual settings can be scaled down.
But the real problem is that if developers want good user experience in low-end consoles they will restrict art and graphical presentations in order to “looking good enough “ in low-end consoles.

The only way to produce “best-looking games” on both high-end and low-end consoles is to make two different versions with different art /presentation style. For example “ni no kuni another world” PS3 and NDS versions use complete different art style.
 
Lockhart being 1080p and Anaconda being 4K, then they should be the same as long as everything else is. RAM could be a little less for smaller textures. CPU would need to be the same. If the CPU is half sized too, then the games will be compromised on one machine or the other.
Specs I don't particularly believe any of the rumors.
But maybe I'm choosing to dismiss certain rumours out of hand?
Like Lockhart being quad core.
That would torpedo my views on scaling.

Little less ram, bandwidth slightly less.
Minor downclock on CPU possibly.
Possibly minor decrease in ssd speed.
Maybe around half the TF on gpu.

Only the GPU in my view can you have such a big discrepancy.

Little extra work, huge benefit? I like the idea. Little boy begging for XBox for christmas is more likely to get one this way.
That's for sure, but 'holding back' may be just an assumption.
This is my view.
The corner case of having a game that is to hard to scale, is outweighed when compared to the greater access to hardware for consumers.
I think this is more applicable to MS than Sony simply because of where they both sit in the market place. Sony would really have to mess up bigger than PS3 to change market reality.
 
Last edited:
Cell was hampered by making up for RSX shortcomings. Cell is still praised to this day https://www.tweaktown.com/news/6916...ll-cpu-far-stronger-new-intel-cpus/index.html

Quote in question:

VAN DER LEEUW: Even desktop chips nowadays, the fastest Intel stuff you can buy is not by far as powerful as the Cell CPU, but it’s very difficult to get power out of the Cell.

This is factually not true.
The Cell had a theoretical maximum of 25.6 GFLOPs * (1 PPE + 8 SPE) = 230.4 GFLOPs FP32.
The $1000 Core i9 10980XE with 18 cores at sustained 3.8GHz theoretically does 18 cores * 4 instructions-per-cycle * 2 FMAD * 3.8 GHz = 547.2 GFLOPs, and this is without even taking advantage of its two AVX 512 units.
Let alone the fact that Intel's out-of-order + massively larger L2 and L3 + massively larger bandwidth to system memory + hyperthreading + branch prediction + micro-op cache + other architectural benefits would always make it much more likely to get closer to its theoretical throughput than Cell could ever do.


Regardless, the Cell's SPEs in the PS3 weren't really praised as a CPU cores. IIRC they were praised as co-processors that were instrumental to perform pixel shading operations due to RSX's weakness in that field.
In a modern architecture with CPU cores that are very fast at single-threaded tasks and GPUs that excel in floating point throughput with much lower power-per-flop, something like the Cell has little reason to exist.


Regardless, the point was Xenon's 3-core PPE was arguably better at being a CPU than Cell, considering Xenos was pretty effective at doing GPU stuff.
 
Where is this talk about the Lockhart CPU being half sized coming from? Tom Warren just said it was downclocked.
 
I'm not a reader or follower Kotaku but there's no denying Jason Schreier has solid access to top tier developers so I'd recommend folks listening to the Split-screen podcast - I did this morning on the way into work. The interesting discussion starts around 25 mins in.

As he said in his tweet, he is hearing that devs are concerned about Lockhart holding back Anaconda (and presumably PS5 because cross-platform games) but he's also hearing that the new consoles are graphically very powerful and, confessing he's not a tech guy, performance of the 2080RTX was used as a frame of reference.

Like others, I find this difficult to believe but he's had a series of insider scoops and while I wouldn't take technical info from Kotaku, I would from devs they talk too.

In six months time (E3) we'll know. If Sony repeat the timeliness of their PS4 reveal, we may high-level technical information about PS5 as early as February. Microsoft's 2013 Xbox conference was in May.
 
Status
Not open for further replies.
Back
Top