Next gen PS5 and next Xbox launch speculation - Secret sauces spicing 2019

When will the first 'next gen' console be revealed?

  • First half of 2019

    Votes: 399 14.8%
  • Second half of 2019(let's say post E3)

    Votes: 1,029 38.3%
  • First half of 2020

    Votes: 1,178 43.8%
  • 2021 :^)

    Votes: 82 3.1%

  • Total voters
    2,688
Oct 27, 2017
191
crazy suggestion incoming:

now we have a rumor about anaconda (Arden) having Non-Transparent Bridge, which supposedly is used to connect 2 processors right?

what if, WHAT IF, anaconda has some sort of an SLI like solution, and that anaconda's GPU is basically 2 lockhart GPUs stacked together?
this means less R&D on creating this and will probably be much easier for developers.
GPUs aren't independent processors, they need a host CPU to function. Also, multi-GPU in a console remains a terrible idea at this time.
 
Nov 12, 2017
1,985
Arden has this as well.

Edit:Additionally Arden has this which is really interesting:

162bArden PCIe Non-Transparent Bridge

In short this is a means to connect two independent processors across PCIe. So, if Arden is the Xbox Scarlet SoC, what other processor exists in Xbox Scarlet for this bridge to connect to?
Raytracing!
 
Oct 27, 2017
1,150
Nah. Raytracing probably wouldn't be done with an independent processor with its own memory either.

MS may just add a secondary low-power processor to run background tasks. Similar to what the PS4/Pro has, but not implemented in such a....."creative" way.
I mean why not. honestly to do raytracing you need a chip for it, just look at the rtx chips where nvidia takes only a small part of the chip. what happends, and I agree ms would never do this, but have a full raytracing chip? that would allow them to dpo stuff no pc can, but its zero chance that happends.
 
Nov 12, 2017
1,985
Nah. Raytracing probably wouldn't be done with an independent processor with its own memory either.

MS may just add a secondary low-power processor to run background tasks. Similar to what the PS4/Pro has, but not implemented in such a....."creative" way.
Yeah ..very likely

Ops sorry double ..guys
 
Oct 27, 2017
191
I mean why not. honestly to do raytracing you need a chip for it, just look at the rtx chips where nvidia takes only a small part of the chip. what happends, and I agree ms would never do this, but have a full raytracing chip? that would allow them to dpo stuff no pc can, but its zero chance that happends.
It's the type of chip. The type of connection enabled by that bridge device is what you would use to connect two CPUs not a CPU and specialized processor unless that specialized processor had the ability and the need to operate independently of the main CPU.
 
Oct 25, 2017
5,001
Make little sense to have raytracing hardware in only one high end SKU of one platform. It won’t get the traction to make it worthwhile.
 
Nov 12, 2017
1,985
Do you have a source for this ?
Yeah

About the reddit leak.

  1. The hardware is partially true
  2. Storage is true
  3. Raytracing is true
  4. Lockhart is not Streaming box
  5. Xbox SoC codename is Anubis, check AMD's plan
  6. MS AI is not a part of the hardware, in other words, never heard of TPU or ASIC like it
  7. How to implement Raytracing? See GDC 2019
  8. Why they make a decision like Lockhart? See GDC 2019
  9. Why there still no DevKit? After GDC 2019
 
Feb 19, 2018
4,037
Scarlett fever and Princess Ariel are perfectly acceptable code names
It's the type of chip. The type of connection enabled by that bridge device is what you would use to connect two CPUs not a CPU and specialized processor unless that specialized processor had the ability and the need to operate independently of the main CPU.
So this is honestly just a secondary processor to offload secondary tasks and downloads just like ps4 has with a secondary ARM. Not anything huge
 
Oct 25, 2017
2,282
I imagine that, if they're doing split pools, they'd rather use them hierarchically - like cache levels - rather than having system and OS pools. If that's the case, it would make sense to go with more than 4GB for the upper level (cheaper) pool.



Ariel has two busses listed, A and B, and two separate bridges. It's probably doing the thing where the processors can access memory directly or through each other's cache, hence the crazy bridging.



I don't know if I grasp the differences between these two topologies well enough to explain it, lol.
Hypervisor/system manager/background task ARM core, mayhaps? Similar to the PS4 Pro.
 
Oct 25, 2017
2,943
crazy suggestion incoming:

now we have a rumor about anaconda (Arden) having Non-Transparent Bridge, which supposedly is used to connect 2 processors right?

what if, WHAT IF, anaconda has some sort of an SLI like solution, and that anaconda's GPU is basically 2 lockhart GPUs stacked together?
this means less R&D on creating this and will probably be much easier for developers.
SLI is never as good as it needs to be.

It Dragon Ball terms it's the opposite of fusion. This means that you end up getting less power than you should. Instead of 100 + 100 = 200 you get more like 100 + 100 is more like 150.
 
Mar 6, 2018
178
What do you make of all this Ariel thing?
I'm your average Joe here, I read the list and don't think much of it.

Do HBCC or StoreMI rely on it?
As far as I understand it the HBCC works self contained on Vega10 without a IOMMU interaction direct CPU page table access.
If you have a console chip with different memory pools then there might be interaction happening with different memory management units.

StoreMI or better FuzeDrive from Enmotus might exploit the IO communication capabilities of recent processors, it does also support Intel CPUs.
 
Oct 25, 2017
3,579
I think secret sauce will be at a minimum.
Nah i expected there will be a good amount for both Sony and MS .
They working with AMD for years on this .
Now i don't expected some crazy TF number just that MS and Sony will have stuff that would suit there needs more .
That AMD won't bother to put in there desktop version cause there is no need to .
 
Oct 27, 2017
191
Scarlett fever and Princess Ariel are perfectly acceptable code names

So this is honestly just a secondary processor to offload secondary tasks and downloads just like ps4 has with a secondary ARM. Not anything huge
No idea. That's just one possibility as to why that bridge device exists. That's why I said I wonder what it could be bridging to. I was hoping somebody could come up with something I didn't think of (or someone who knows would drop a hint).
 
Last edited:
Oct 25, 2017
133
I think it's best not to expect any surprises in terms of the hardware at this point. After Navi is revealed, we'll essentially know what kind of performance we're going to get by looking at the mid-range option (which might still be the fastest Navi GPU they reveal, mind you).

I think what people are going to be impressed by, and this has been the case for every console generation for a while, is the innovation on the software side. Sure, you can't get any more realistic than PBR in terms of authoring content, but all those little things devs wish they could put into their games today will become commonplace and more brutally optimized than we've ever seen them before. Just think of all the beautiful techniques for indirect lighting, shadows, and complex materials you've already seen (sans raytracing) being used together and at higher settings.

And let's not even get started on how insane the artwork is going to look, given the right talent is involved. Actually nevermind, let's. We can already use far more polygons than devs are typically using on current hardware, and a lot of the reason artists aren't going higher is because of the futility of making an overly smooth model without a detailed enough normal map to complement it. Also, many are afraid of vertex skinning costs, but this limit is going to be lifted in a tremendous way with these new CPUs. So from day one, we'll see many of these rendering techniques used together with less pressure for artists to optimize, and without a fundamental shift in authoring tools, leading to an impressive leap in visuals without any new innovations. After all, there are many realtime techniques that have been available for years that devs have been eager to use which the PS4 just couldn't support without aggressive optimization and absurd man hours. Many of these things were only just barely out of reach on contemporary console hardware.

Now, on top of all of that, consider the additional techniques and research that will be born out of an 8-12 TF GPU being the lowest common denominator (excepting the Switch since it requires so much extra work and many games already skip it). While we could certainly use the crazy performance some are daydreaming about, I think a lot of developers aren't even going to know what to do with 10 TF especially when it comes to their artwork. 'Easy mode' would be making the games of today with PC-like settings and more density, which isn't exactly a bad thing, but if a developer really wants to make use of the hardware, they'll be able to blow our socks off.

The primary limitations on artists from this generation will be gone. There is so much short of raytracing that we're going to be happy to have.
 
Nov 12, 2017
2,205
SLI is never as good as it needs to be.

It Dragon Ball terms it's the opposite of fusion. This means that you end up getting less power than you should. Instead of 100 + 100 = 200 you get more like 100 + 100 is more like 150.
On PC yes, where few use it and the devs likely don't have very much time to program for it efficiently. Couldn't it be different in the case of a console that has it built-in, with every single game being programmed to use it properly?
 
Oct 27, 2017
1,150
On PC yes, where few use it and the devs likely don't have very much time to program for it efficiently. Couldn't it be different in the case of a console that has it built-in, with every single game being programmed to use it properly?
dont think its worth it, but if sony or ms were to do it, I would rather have a gpu+raytracing chimp, than gpu+gpu. the reasson is its never 1 to 1, at best you get maybe 80% of the second gpu.
 
Nov 12, 2017
2,205
dont think its worth it, but if sony or ms were to do it, I would rather have a gpu+raytracing chimp, than gpu+gpu. the reasson is its never 1 to 1, at best you get maybe 80% of the second gpu.
Yeah, though 80% could well be enough when 2 smaller chips are cheaper to make than one larger chip due to better yields.

Also, I now want a ray-tracing chimp.