Support NeoGAF

[FranXico]

Many have reading issues here, sadly.

If you read the quote from MS correctly, they outline the position/advantage least in their minds!!!

Quote
"Xbox Series X|S are the only next-generation consoles with full hardware support for all the RDNA 2 capabilities.

MS does not challenge the PS5 GPU not based on RDNA2 subset.

There position is the Series/ S are the only next gen consoles that fully hardware support all the features for RDNA2.

This quote seems to be right based on things we know. Since Sony themselves announced VRS not already there. They have a line in a blog post it be included later in a software update. It's not VRS hardware RDNA2.

Funny how you talk about others not being able to read, but the blog post was referring to VRR being enabled in a future firmware update. VRR, which has absolutely nothing to do with VRS.

All this to defend marketing material as fact.

I advise you to develop a little bit of critical thinking and not rush to insult others in the future.

 

[martino]

Well the more we look at it the more it seems to me we probably had all the infos all along....Youtube click baiter and loud fanatics from all side are just making things looking more complicated than they are.

The shader cores of the Xbox are also more suitable to machine learning, which could be an advantage if Microsoft succeeds in implementing an equivalent to Nvidia's DLSS (an advanced neural network solution for AI).

i think MS will more have to use it to compensate where they could lack down the road https://gamerant.com/microsoft-xbox-cloud-gaming-experimental-tech-machine-learning/

It's going to hit at every level of development from the early stages of development to the deepest layers of gameplay. So just an example, on the content creation side we already talked about Flight Simulator, but we also see machine learning models doing some pretty crazy stuff. We have a research project in our studios where they built a texture decompression and compression algorithm, where they trained a machine learning model up on textures. They can take a big texture, shrink it down till it's really ugly, and then use the machine learning model to decompress it again in real time. It's not exactly the same, but it looks really realistic. It's almost creepy how it works. There's a lot of areas where you can use AI as a creative aid.

 

[prinz_valium]

In fact, for me infinity cache is not RDNA2 related, but it's navy 21 related. I'm pretty sure that with smaller GPU using RDNA 2, you'll have less IC quantity or no IC.

They will have less IC, but not no IC.
It's needed for the lack of memory interface width

 

[longdi]

to sum up, rdna2 core differences are improved pipelining to reach higher clocks at lower watts and a RT core with the tmu.

other niceties include more efficient rops with hw vrs. improve ge with hw mesh shaders. improve tmu prt texturing with sampler feedback streaming hw. improv ML with more mixed precision hw.

 

[MonarchJT]

Well the more we look at it the it seems to me we probably had all the infos all along....Youtube click baiter and loud fanatics from all side are just making things looking more complicated than they are.


i think MS will more have to use it to compensate where they could lack down the road https://gamerant.com/microsoft-xbox-cloud-gaming-experimental-tech-machine-learning/

and to see how the ps5 own version of the rdna2 mesh shader perform compared to the amd one.(for the sake of god they aim to do the same things!)

But yes ML is a real deal

 

[onesvenus]

Great XSX and PS5 are both RDNA2, both Sony and MS officially stated that the two GPU's are "specialised configurations of RDNA2" / customised RDNA2 (actual quote from one of these tweets from James "SFS = 2-3x better than <unspecified baseline>" Stanard, and also what Sony and Mark "Pretty Cool, Right? Lend me your ears" Cerny also stated) .

Case closed?

Yes, of course.
Can you point me where I said PS5 GPU was not RDNA2-based?
I've only been against unsubstantiated rumors like IC cache or RDNA3 parts.

 

[Panajev2001a]

Yes, of course.
Can you point me where I said PS5 GPU was not RDNA2-based?
I've only been against unsubstantiated rumors like IC cache or RDNA3 parts.

I am not saying you said that, really trying to stop going in circles on this please, I am old I can only go in loops so much in a week .

 

[Old Empire.]

Funny how you talk about others not being able to read, but the blog post was referring to VRR being enabled in a future firmware update. VRR, which has absolutely nothing to do with VRS.

All this to defend marketing material as fact.

I advise you to develop a little bit of critical thinking and not rush to insult others in the future.

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

 

[Fess]

Ethomaz is the greatest poster on GAF.
I refuse to believe he is a functioning adult with a job.
Im just waiting for the reveal that he is actually a mental patient that stole someones phone and lives under a bridge.

Got to respect him for having some mighty balls to confront the console maker heads on.
Should've done it 37 pages ago though and used the time to play some games instead. I'm playing Amiga right now, zero CUs, 7MHz CPU, still higher and more stable fps on some games than what we have on XSX and PS5

 

[Superayate]

They will have less IC, but not no IC.
It's needed for the lack of memory interface width

That's depend of the GPU specs. With a smaller GPU than the navi one used in the RX5700XT, you don't need the IC. And at this performances level, that's clearly less usefull and too expensive to integrate that I think (smaller or not).

 

[Elog]

Please note: While we know that MS has ML capabilities, we do not know what Sony has. Looking at comments from various parties the jury is still out on that one. Locusa above has good arguments why ML probably is not there. Leviathangamer has good arguments for why it most likely is there.

Personally, I have no idea

 

[kyliethicc]

 

[Garani]

that you probably have not followed (i know you have, but i let it go) the succession of events nor understood what experts said about the ps5 dieshot nor what Stanard said to ethomaz , those users had spit fanboy nonsense without having a minimum of knowledge on the topic and of the facts , and I'm being kind

stop it here and follow what is saying panajev

Well, I have been following the whole shit show, but I am in a difficult moment and don't have the will to get into a constant rebutal war.

You have taken very stubbord positions, that are not substantiated, and the whole thread, but XBox fans, called you out. They have been very vocal but not unpolite (I have seen worse flame wars in life).

In fact, for me infinity cache is not RDNA2 related, but it's navy 21 related. I'm pretty sure that with smaller GPU using RDNA 2, you'll have less IC quantity or no IC.

Well, it was presented as RDNA2 by AMD (but I know what you mean and I agree with you).

Neither PS5 or Series/ S has infinity cache that's thoroughly known.

Microsoft never said the PS5 GPU is just a RDNA1 die. There believe is they got a GPU from AMD that hardware supports most of the capabilities that AMD provide with Q4 2020 RDNA2 release.

Right now that seems to be the case, since VRS not there to be used right away by devs teams making a third party version of the game for PS5.

There literally is no reason Sony would remove VRS functionality that some here suggested they did. MS waited a bit longer than Sony for the refresh upgrade. Will this pay off we just have to wait and see. RDNA2 gpus are not utilised fully by devs teams yet since the features are new upgrades and will take a bit longer to be selected. There used to the older techniques of building games, that's why we see minor differences visually this gen from last gen.

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

You keep on writing this stuff that, honestly, it's disheartening. You keep on wasting time on the VRS like it's the last argument that you have and you are trying to milk it to death. The sooner you accept that VRS is fake shit (we have the pictures to prove it) and that Sony decided it was better to use silicon to* cull triangles way before the rendering process instead of using VRS, the sooner you'll be at peace with yourself.

VRS, today, gives bad results and is needed only if you have no other optimization at the beggining of the rendering pipeline.

EDIT: * I have added the phrasing "to use silicon to" because it wasn't clear in the original message.

 

[KaitlyynDuff]

it seems like no one is expecting that PS5 will release this haha. It is also a good thing that the game reveals that something is missing even before the netizens can fully make a comment on it. It is also a good thing that they found a way to release the game even with the pandemic. What's interesting is that they even reported that more than 4.5 million units are already shipped within this month. I hope that it wont diasppoint us and live up to the expectations. I also hope that it will be offered in an affordable price.

 

[FatKingBallman]

I know that you are a good guy and you doing this to keep the situation calm in a convulsive thread. But you know better than me that things most likely aren't quite like that.

But if I no longer see people like John wick, assurdum, fat etc etc write nonsense I'll close it here too even though I'm and i was right all the time...and we will let the performances speak ... when the hw is exploited

case closed

First and foremost you should stop spreading nonsense. Everyone here reacted to you because you're spreading RDNA 1 PS5 GPU crap since beginning of this thread.

that you probably have not followed (i know you have, but i let it go) the succession of events nor understood what experts said about the ps5 dieshot nor what Stanard said to ethomaz , those users had spit fanboy nonsense without having a minimum of knowledge on the topic and of the facts , and I'm being kind

stop it here and follow what is saying panajev

And you are an expert here and it give you rights to calling out others and mock them? One of the member here, Locuza, the man whom you trust and you said he is never wrong, said that XSX isn't RDNA 2? Why you didn't call him out? So, what you will say to him because he offended XSX? Like i've said before, you have a DOUBLE STANDARDS.

 

[S0ULZB0URNE]

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

Why would MS clock the XSX lower than RDNA 2 spec?

They are custom.

 

[martino]

and to see how the ps5 own version of the rdna2 mesh shader perform compared to the amd one.

all we have to know about ps5 geometry engine was on road to ps5 :
what did Cerny said about it ?

PlayStation 5 has a new unit called the Geometry Engine which brings handling of triangles and other primitives under full programmatic control.
As a game developer you're free to ignore its existence and use the PlayStation 5 GPU as if it were no more capable than the PS4 GPU or you can use this new intelligence in various ways.
Simple usage* could be performance optimizations such as removing back faced or off-screen vertices and triangles.

* this is where people loose the thread and think it's about geometry engine when it's about primitive shader simple usage (or simple programmatic control example if you prefer). this is also the point where youtuber can make click bait speculation video. those speculation are laughtable to me because they talk about fixed function advantage when the feature is about overriding/replacing/not using them.
What describe here is also in line with amd and nvidia implementations and dx12u specs of primitive/mesh shader.
Amd ones struggling hard on vega/rdna1 and now working on rdna2
is that the result of a successful collaboration seen in a gpu launched roughly at the same as ps5 ? it seems like it.
Will one perform better ? ps5 has GMM and higher clock advantage and xsx has more units so it will depends

Also unlike what i though this trend to replace fixed hw function by shader workload is also true and give better result for vrs especially in deferred engine

 

[Max_Po]

REAL die shot

Spoiler

 

[PaintTinJr]

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

Everything that runs on a GPU in a modern console is effectively hardware accelerated, and the future of graphics is to move further and further away from fixed path features that confine the developer solutions. VRS in a fixed path feature has to offer whole multiplier number performance benefits over any evolving solution (deferred or async, etc?) to warrant restricting graphics programmers further. Which it almost certainly doesn't, and won't even match the saving already being made by the PS5 GE before the data is passed to those pipelines.

In fact, AFAIK, the greater the use of RT on games, the less relevant a fixed path VRS would be, because RT by its very nature only shades the pixels that are visible, and is VRS at its most optimal level. So we've already seen it in action in Spider-man's 30fps and 60fps RT modes. So failing to tick a box on DX RDNA2 spec sheet is pretty much a nothing issue, when the GE in the PS5 is going to be far more versatile, and thePS5 culls more effectively going by the Ex-Sony engineers' tweets, and the PS5 has more ROPs, and its RDNA2 RT and TMU can work simultaneously like the RDNA2 PC cards according to Cerny's words - saying they have the same RT BVH as the PC RDNA2 cards - unlike the XsX hotchips that says it is BVH or TMU functionality at any one time.

 

[M1chl]

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

Very simple, to save space, if they don't feel like needing it.

 

[longdi]

Very simple, to save space, if they don't feel like needing it.

or perhaps the hw vrs needed more time to finalize and will go past sony targeted launch timings.

 

[M1chl]

or perhaps the hw vrs needed more time to finalize and will go past sony targeted launch timings.

Possible, but size of the die, is always massive concern.

 

[Rea]

Cerny
"you can use this new
intelligence in various ways. "

why did Cerny called PS5 GE an intelligence . Also why he show the human brain picture to represent the GE, can anyone explain?

 

[martino]

Cerny
"you can use this new
intelligence in various ways. "

why did Cerny called PS5 GE an intelligence . Also why he show the human brain picture to represent the GE, can anyone explain?

because handling of triangles and other primitives are under full programmatic control
it opens new possibilities limited by devs imagination and capabilities to implement them.
ex of a dev thinking over what could be possibly done :

Having access to the topology in the mesh shader should enable us to calculate dynamic normals, tangents, and curvatures for a mesh that's deforming due to complex skinning, displacement mapping, or procedural vertex animation. We can also do voxel meshing, or isosurface extraction—marching cubes or tetrahedra, plus generating normals etc. for the isosurface—directly in a mesh shader, for rendering fluids and volumetric data.

Geometry for hair/fur, foliage, or other surface cover might be feasible to generate on the fly, with view-dependent detail.

3D modeling and CAD apps may be able to apply mesh shaders to dynamically triangulate quad meshes or n-gon meshes, as well as things like dynamically insetting/outsetting geometry for visualizations.

For rendering displacement-mapped terrain, water, and so forth, mesh shaders may be able to assist us with geometry clipmaps and geomorphing; they might also be interesting for progressive meshing schemes.

And last but not least, we might be able to render Catmull–Clark subdivision surfaces, or other subdivision schemes, more easily and efficiently than it can be done on the GPU today.

To be clear, a great deal of the above is speculation and handwaving on my part—I don't want to mislead you that all of these things are for sure doable with the new mesh and task shader pipeline. There will certainly be algorithmic difficulties and architectural hindrances that will come up as graphics programmers have a chance to dig into this. Still, I'm quite excited to see what people will do with this capability over the next few years, and I hope and expect that it won't be an NVIDIA-exclusive feature for too long.

 

[chilichote]

Please folks, let the RDNA2 or not RDNA2 discussion die and focus on the die. This annoys me and I start to report (whether it makes sense or not) if it doesn't stop.

Both have taken what they need from the AMD portfolio and customized it with their own needs/technology. The End.

 

[RaySoft]

Please folks, let the RDNA2 or not RDNA2 discussion die and focus on the die. This annoys me and I start to report (whether it makes sense or not) if it doesn't stop.

Both have taken what they need from the AMD portfolio and customized it with their own needs/technology. The End.

Exactly. Just goes to show why companies trademarks some random tech with a fancy name. People crave everything with a (tm) behind it;-)

 

[Loxus]

lol no you didn't!

He added one more piece.

So let me get this straight.
He is neither a GPU Engineer or the Engineer of XBSX. He just work at Microsoft.
Mark Cerny the Engineer of the PS5, provides break downs of the PS5, you don't believe him and says he's lying.
While when that guy is asked to provide information, he says he's an expert witness.
Lmao, how does he prove anything. He didn't even provide anything meaningful, just regurgitate the same PR from Microsoft.

 

[Lysandros]

Do we have any new insight about the location of Cache Scrubbers? Are they spread within the CUs/L0 or L1 caches, where are they exactly? Also why the big I/O area isn't being analysed, any idea about the size of SRAM etc.? Where the Tempest Engine might be situated if it's outside of the DCU clusters?

 

[chilichote]

Exactly. Just goes to show why companies trademarks some random tech with a fancy name. People crave everything with a (tm) behind it;-)

I can understand why many people want to know what technology is in the console of their choice. But this discussion of whether Series X is now more RDNA2 than the Playstation 5 or vice versa is slowly becoming absurd and ridiculous, simply because both companies have designed their consoles the way they need them. And to do this, they built on AMD's existing technology. This is communicated by Microsoft, Sony and also AMD. In this respect, we are simply should concentrating, at least in this thread, on the PS5 die. Obviously there is still a lot that is not easily understandable or unclear.

I would like to hear from Sony a few details about the cache scrubber. And to the GE. And how much performance gain is achieved as a result. Whether VRS is already taking place in the GE or is becoming unnecessary due to the GE, etc.

I don't have Twitter, but maybe someone there will ask Matt Hargrett specific questions that he may answer indirectly, like in the RGT video where he talked about the switch^^

 

[HoofHearted]

So let me get this straight.
He is neither a GPU Engineer or the Engineer of XBSX. He just work at Microsoft.
Mark Cerny the Engineer of the PS5, provides break downs of the PS5, you don't believe him and says he's lying.
While when that guy is asked to provide information, he says he's an expert witness.
Lmao, how does he prove anything. He didn't even provide anything meaningful, just regurgitate the same PR from Microsoft.

Eh... Just a thought - did you read his LinkedIn profile?

 

[ethomaz]

So ethomaz , will you stop saying XS GPUs are RDNA1 based now or will you dismiss his opinion because he is a Microsoft employee?

I never said the GPU ARE RDNA1.

I said and it is confirmed the GPUs have RDNA1 parts and are not full based in RDNA 2.

 

[HoofHearted]

Why would MS clock the XSX lower than RDNA 2 spec?

They are custom.

I keep seeing this discussion about a supposed "clock" spec for RDNA2... where is it?

 

[ethomaz]

lol no you didn't!

He added one more piece.

He still keep spinning the PR marketing because the silicon has old non-RDNA2 para jn GPU.

Dodged the question 3 times and trying to talk about SoC lol

 

[HoofHearted]

I never said the GPU ARE RDNA1.

I said and it is confirmed the GPUs have RDNA1 parts and are not full based in RDNA 2.

Well at least you got a definitive answer ....

 

[ToTTenTranz]

Sony PS5 gpu lacking hardware accelerated VRS too and you don't think that makes no difference? Why would Sony remove AMD core RDNA2 features? The explanation you have is likely a wetdream that they created something better than AMD could?

If you're claiming the PS5 lacks hardware accelerated VRS then you're right. The PS5 would never support a Microsoft DX12 Ultimate implementation.

If you're claiming the PS5 lacks acceleration for foveated rendering (where VRS is just DX12's name for it), then you're most probably wrong.
There are patents for foveated rendering implementations from Sony with applications dating back to 2016. It doesn't matter how subtle the differences might be on block arrangements, but Sony has been on top of this matter even before the PSVR came out.

I wish people would understand how worthless the "PS5 doesn't support VRS" statement is, and just move on from the subject for the nothing-burger it is.



Also, the performance boosts provided by VRS on non-VR games (and not the synthetic 3dmark test that is designed to maximize its effectiveness in an unrealistic scenario) have been questionable to say the least.
There's a reason Foveated Rendering has been around for many years, and it's always been part of SDKs for Virtual Reality apps and games. nVidia puts VRS inside their VRWorks Suite. Qualcomm supports it on their VRDK and it originally called it Adreno Foveation.
VRS / Foveated Rendering on monitors/TVs is an afterthought, and people expecting great things from it are going to be disappointed.

Foveated Rendering is meant to be used with eye-tracking, to reduce IQ wherever our eyes aren't focusing on. It's not great to be used in a single panel where the reduced shading rate will cause noticeable differences in IQ.

 

[HoofHearted]

He still keep spinning the PR marketing because the silicon has old non-RDNA2 para jn GPU.

Dodged the question 3 times.

?????

He was pretty direct in his response. GPU is RDNA2 - other portions of SoC are Zen2. What question did he dodge?

 

[ethomaz]

Well at least you got a definitive answer ....

Nope.

He dodged it.

 

[ethomaz]

?????

He was pretty direct in his response. GPU is RDNA2 - other portions of SoC are Zen2. What question did he dodge?

I asked him three times why parts of the silicon are non RDNA2... he said the three times the SoC have parts that are Zen2 totally unrelated to the question.

He never answered what I asked.

 

[ToTTenTranz]

Again:

I would be more conservative with the analyses being made on 9 Mpix photographs to interpret the lower level customizations that Sony or Microsoft applied in >12 Billion transistor chips.
Some things will be noticeable if there photos of previous chips with similar architectures to compare against, like approximate cache sizes, how many PHYs, number of WGPs and their redundancy, etc. Others will not be noticeable, because one pixel equates to over 1K transistors and the photo isn't perfectly focused.

That's not to say that Locuza and others aren't doing a spectacular job, but many things will be guesses.

 

[HoofHearted]

I asked him three times why parts of the silicon are mint RDNA2... he said the three times the SoC have parts that are Zen2 totally unrelated to the question.

He never answered what I asked.

Admittedly I haven't had my second cup of coffee yet and haven't read the full thread on Twitter... but since you're here... what exactly was your question to him that he "dodged"?

 

[HoofHearted]

Again:

I would be more conservative with the analyses being made on 9 Mpix photographs to interpret the lower level customizations that Sony or Microsoft applied in >12B chips.
Some things will be noticeable if there photos of previous chips with similar architectures to compare against, like approximate cache sizes, how many PHYs, number of WGPs and their redundancy, etc. Others will not be noticeable, because one pixel equates to over 1K transistors and the photo isn't perfectly focused.

That's not to say that Locuza and others aren't doing a spectacular job, but many things will be guesses.

Exactly - based on Mr. Stanard's role/position at MS - I'm fairly certain that he actually knows a bit more about what is or isn't on the SoC.

 

[ethomaz]

Admittedly I haven't had my second cup of coffee yet and haven't read the full thread on Twitter... but since you're here... what exactly was your question to him that he "dodged"?

His answers.



Another answer to another guy.



He never touched what was asked.
He keep saying the SoC has Zen2 and other things lol

 

[ToTTenTranz]

He answered your question: the Series X/S GPUs are 100% RDNA2, and your comment of the presence of RDNA1 parts is wrong by his (informed) account.
You said "parts of the silicon aren't RDNA2", to which he answered "yes, because parts of the silicon are not the GPU like the Zen2 cores".

There's no dodging here. The list of "RDNA compliance" you made is wrong for the Series consoles, and it may be equally wrong for the PS5.

 

[ethomaz]

He answered your question: the Series X/S GPUs are 100% RDNA2, and your comment of the presence of RDNA1 parts is wrong by his (informed) account.
You said "parts of the silicon aren't RDNA2", to which he answered "yes, because parts of the silicon are not the GPU like the Zen2 cores".

There's no dodging here. The list of "RDNA compliance" you made is wrong for the Series consoles, and it may be equally wrong for the PS5.

The talk was about GPU silicon not the whole SoC.
He did not answer why parts of the GPU are not RDNA2.

They are not 100% RDNA2.

The compliance guess is mostly like right... and I just included 4 big blocks... there are several others blocks that are not RDNA2.

 

[ToTTenTranz]

The talk was about GPU silicon not the whole SoC.
He did not answer why parts of the GPU are not RDNA2.

He said parts of the silicon are not RDNA2, because not all silicon is the GPU. The CPU isn't RDNA2, it's Zen2.
He even clarified that in a following tweet:

They are not 100% RDNA2.

At this point I think you either choose to believe a Microsoft Engineer or you don't. It's not a matter of interpretation.

 

[ethomaz]

He said parts of the silicon are not RDNA2, because not all silicon is the GPU. The CPU isn't RDNA2, it's Zen2.
He even clarified that in a following tweet:









At this point I think you either choose to believe a Microsoft Engineer or you don't. It's not a matter of interpretation.

He did not answer why there parts of the GPU silicon that are not RDNA2.
No body asked about the parts of the the silicon that are not GPU. He trowed Zen 2, non GPU parts, etc in the replies.

Maybe he got confused with the twitter thread he was replying so in the next question I included GPU silicon to make it clear and he replied again with Zen 2, non GPU parts, etc.

I realized I won't get a straight reply from him.

 

[onesvenus]

I never said the GPU ARE RDNA1.

I said and it is confirmed the GPUs have RDNA1 parts and are not full based in RDNA 2.

But he says it's 100% RDNA2.

 

[Knitted Knight]

Thread is getting spicy.

 

[ethomaz]

But he says it's 100% RDNA2.

That is not correct from what the GPU die shot and drivers showed.... that is the main point why I asked.

We know the GPU is 100% RDNA2 functional... in simple terms it can to everything RDNA2 card can... imo Infinity Cache doesn't affect that because it is only a performance feature and not a functional.

Said that the we are discussing the GPU silicon design that are not 100% RDNA2... that is what we were asking to him to understand why they choose old parts instead the new RDNA2 ones.

 

[Panajev2001a]

the PS5 has more ROPs, and its RDNA2 RT and TMU can work simultaneously like the RDNA2 PC cards according to Cerny's words - saying they have the same RT BVH as the PC RDNA2 cards - unlike the XsX hotchips that says it is BVH or TMU functionality at any one time

We need to get this clarified because it would be pretty interesting if so (although I am not sure how big the difference would be). Need to double check the HotChips presentation... that seems like an odd implementation of RDNA2 RT.