• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
  • The Politics forum has been nuked. Please do not bring political discussion to the rest of the site, or you will be removed. Thanks.

Will ML be a differentiator between the consoles?

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
Insecure? You're citing 2 racing games as system sellers? Gears 5? 50m sales for the powerful console that couldn't even pass the portable machine with its simple looking games?

So 3 titles in the entire gen for the most powerful console? Can you understand what I'm trying to say?

I was just trying to illustrate how power is nothing without control.

MS needs talent. That talent has to be cultivated. Do 343 look talented? Look at the 360 era.
You are cherry picking to push a narrative.
Is 343 the only MS game studio?
Turn 10 is top of the food chain developer, and make one of the best car sims available on any platform in Forza.
Playground make THE best car game in Horizon series. They now have a second team making fable. PG are an amazingly talented studio.
The Coaltion are an excellent studio and Gears 5 was a super looking game.
Ninja Theory are an amazing studio.
Obsiden are one of the best WRPG makers going around.
Id are the OGs of shooter games.
Arkane are an amazing AAA studio.
Bethesda is the big daddy of WRPGs.

Shit has changed, and the "Xbox has no games" troll is over.
You come across as a bit rattled by the improvement MS has made.
What's the saying? "The lady doth protest too much, methinks"

As a Sony fanboy you should be happy to see MS step up their games studios, as it will most likely mean Sony will do the same. Then you win as well.
 
  • Like
Reactions: MrFunSocks and Riky
Mar 7, 2017
2,897
6,106
520
Again, the context is shader cores hardware MORE suitable to machine learning between the consoles.

Source 1
From https://wccftech.com/xbox-series-xs...ning-powered-shader-cores-says-quantic-dream/

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).

Yes. He’s referring to Rapid Packed Math for Integer OPs which lives in the shader cores. XSX is confirmed to have it. PS5 is not confirmed to do so. A lack of confirmation is not a confirmation of lack.

Wccftech is merely speculating as he references no source other than himself.

In all likelihood, RPM for int is a RDNA2 staple that PS5 includes, e.g. just like the RT cores in the TMUs.
 

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
Yes. He’s referring to Rapid Packed Math for Integer OPs which lives in the shader cores. XSX is confirmed to have it. PS5 is not confirmed to do so. A lack of confirmation is not a confirmation of lack.

Wccftech is merely speculating as he references no source other than himself.

In all likelihood, RPM for int is a RDNA2 staple that PS5 includes, e.g. just like the RT cores in the TMUs.
There's level to RPM.
PS4 Pro had FP16 for instance which is RPM.
4 And 8 bit Int is not a staple of RDNA 2 at all. There is no confirmation that AMD cards have 4 and 8 bit Int as does the XSX. In fact, MS confirmed they added the hardware required to add it of their own desire, over and above stock RDNA 2.
 
  • Like
Reactions: MrFunSocks

J____Av

Member
Jun 12, 2020
235
328
260
As a Sony fanboy you should be happy to see MS step up their games studios, as it will most likely mean Sony will do the same. Then you win as well.

No one in gaming should be happy when a company buys a massive established studio and picks up games THEY WERE ALREADY GETTING! They gained no games XBOX fans couldn't already play. They added nothing to their gaming catalog. That's not stepping up your game. Bringing games in you wouldn't get is stepping up your game and forcing the comp to step up. A move like what Microsoft did is more likely to push Sony to buy up gaming studios to get games they already were getting. Its terrible for the industry. I am PS first, but I would be PISSED at them if they spent a shitload of money to by someone like Take Two

And take away Bethesda and XBOX first party is back to being horrible
 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
Yes. He’s referring to Rapid Packed Math for Integer OPs which lives in the shader cores. XSX is confirmed to have it. PS5 is not confirmed to do so. A lack of confirmation is not a confirmation of lack.

Wccftech is merely speculating as he references no source other than himself.

In all likelihood, RPM for int is a RDNA2 staple that PS5 includes, e.g. just like the RT cores in the TMUs.
That statement is false.

From https://wccftech.com/david-cage-on-quantic-dream-future-next-gen-gaming-and-not-ruling-out-sequels/

Wccftech Question: As developers, what do you think about the PlayStation 5 and Xbox Series X? Which unique aspects of each platform excite you the most?

CEO David Cage of Quantic Dream's Answers: 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).

---------------

Baseline NAVI has a level of Rapid Pack Math.




Baseline NAVI has 2X rate INT16/FP16 rapid pack math.

4X rate INT8 and 8X rate INT4 are optional features for NAVI. MS requested AMD to include the optional NAVI dot4 and dot8 feature for XSX.
 
Last edited:

SCWTRAIN82

Neo Member
Oct 30, 2020
9
8
80
One or two paragraphs in a multiple pages interview == “talked it up to no end”... sure... :rolleyes:.


So, did they then drop FP16 with the proof for it being they did not talk about it at all (they did not call it out) ;)? What kind of point are you trying to make really?

You are the same one that is going about the SSD in the XSX thread claiming that it is surely customised in ways we do not even know even though nothing was mentioned about it specifically yet for some mysterious reason if Sony does not explicitly mentions something they have jack? Almost forgot that as far as MS is concerned “hope springs eternal”.


Oh yeah, this is the point you were building up to :LOL:.
The dudes a joke, you can’t take him seriously.
 

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
There's level to RPM.
PS4 Pro had FP16 for instance which is RPM.
4 And 8 bit Int is not a staple of RDNA 2 at all. There is no confirmation that AMD cards have 4 and 8 bit Int as does the XSX. In fact, MS confirmed they added the hardware required to add it of their own desire, over and above stock RDNA 2.
From https://github.com/llvm/llvm-project/commit/9ee272f13d88f090817235ef4f91e56bb2a153d6

Big Navi/GFX1030/Navi 21/Sienna Cichlid/RX 6800 series does support those Instructions
case GK_GFX1030:
Features["ci-insts"] = true;
Features["dot1-insts"] = true;
Features["dot2-insts"] = true;
Features["dot5-insts"] = true;
Features["dot6-insts"] = true;
Features["dl-insts"] = true;
Features["flat-address-space"] = true;
Features["16-bit-insts"] = true;
Features["dpp"] = true;


dl-insts = deep learning instructions.

RX 6800 series will double as a workstation GpGPU SKU on Apple's Mac Pro!

CDNA is not a GPU since it's missing rasterization and raytracing hardware.
 
Last edited:
  • Like
Reactions: longdi

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
No one in gaming should be happy when a company buys a massive established studio and picks up games THEY WERE ALREADY GETTING! They gained no games XBOX fans couldn't already play. They added nothing to their gaming catalog. That's not stepping up your game. Bringing games in you wouldn't get is stepping up your game and forcing the comp to step up. A move like what Microsoft did is more likely to push Sony to buy up gaming studios to get games they already were getting. Its terrible for the industry. I am PS first, but I would be PISSED at them if they spent a shitload of money to by someone like Take Two

And tak away Bethesda and XBOX first party is back to being horrible
Taken away all the companies Sony bought and they would be horrible as well.
Companies build up their studios with both aquisitions and through organic growth (creating new studios). MS has done both, as has Sony.
The only difference is MS is a big dog, and Sony is a smaller dog.
It's obvious that alot of this "MS has no games" yap is coming from a fear that MS might soon have too many games.

MS haven't finished buying yet. There will be more. I think it's a good thing. I think it's good if Sony follow suit and buy some more studios. That's because I will get both consoles and won't miss out on anything I want to play.
The thing that amazes me is when people say "Its all about the games", "Games are more important than hardware", but yet say "I wont buy an Xbox or PC to play Microsoft games". So in reality, it's not about the games at all. Its about your team.
 
Last edited:
  • Thoughtful
Reactions: Riky

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
From https://github.com/llvm/llvm-project/commit/9ee272f13d88f090817235ef4f91e56bb2a153d6

Big Navi/GFX1030/Navi 21/Sienna Cichlid/RX 6800 series does support those Instructions



dl-insts = deep learning instructions.

RX 6800 series will double as a workstation GpGPU SKU on Apple's Mac Pro!

CDNA is not a GPU since it's missing rasterization and raytracing hardware.
It shows 16 bit. That is FP16. I didn't say RDNA didn't have FP16, I said it didn't have 4 and 8 bit at this stage.
AMD haven't confirmed this, and I would have expected them to.
 

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
It shows 16 bit. That is FP16. I didn't say RDNA didn't have FP16, I said it didn't have 4 and 8 bit at this stage.
AMD haven't confirmed this, and I would have expected them to.
"16-bit-insts " = 16 bit instructions

NAVI 10 (GK_GFX1010) is missing GFX1030's dot1, dot2, dot5, dot6 instructions.


case GK_GFX1030: //Radeon RX 6800 series
Features["ci-insts"] = true;
Features["dot1-insts"] = true;
Features["dot2-insts"] = true;
Features["dot5-insts"] = true;
Features["dot6-insts"] = true;
Features["dl-insts"] = true;
Features["flat-address-space"] = true;
Features["16-bit-insts"] = true;
Features["dpp"] = true;
Features["gfx8-insts"] = true;
Features["gfx9-insts"] = true;
Features["gfx10-insts"] = true;
Features["gfx10-3-insts"] = true;
Features["s-memrealtime"] = true;


case GK_GFX1010: //Radeon RX 5700 series
Features["dl-insts"] = true;
Features["ci-insts"] = true;
Features["flat-address-space"] = true;
Features["16-bit-insts"] = true;
Features["dpp"] = true;
Features["gfx8-insts"] = true;
Features["gfx9-insts"] = true;
Features["gfx10-insts"] = true;
Features["s-memrealtime"] = true;


case GK_GFX906: //Radeon VII has dot1 and dot2 instructions
Features["dl-insts"] = true;
Features["dot1-insts"] = true;
Features["dot2-insts"] = true;
 
Last edited:

Shubh_C63

Member
Dec 6, 2016
1,112
870
420
India
Nothing is groundbreaking anymore.

Level with me that these tech in some form were always used. Just having a name and obvious power increase in console to better utilise those tech doesn't make it a system seller.

Reality is often disappointing. It will be more of the same (regarding development) with a nudge to some small areas.
 

BeardGawd

Member
Dec 16, 2019
794
1,431
485
Yes. He’s referring to Rapid Packed Math for Integer OPs which lives in the shader cores. XSX is confirmed to have it. PS5 is not confirmed to do so. A lack of confirmation is not a confirmation of lack.

Wccftech is merely speculating as he references no source other than himself.

In all likelihood, RPM for int is a RDNA2 staple that PS5 includes, e.g. just like the RT cores in the TMUs.

Yet Sony confirmed RT hardware. Why won't they do the same for rpm int8/int4?

There's no reason to keep it secret. Anybody with common sense can see why they won't confirm these "standard rdna2 features".
 
  • Like
Reactions: rnlval and Riky

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
"16-bit-insts " = 16 bit instructions

NAVI 10 (GK_GFX1010) is missing GFX1030's dot1, dot2, dot5, dot6 instructions.


case GK_GFX1030: //Radeon RX 6800 series
Features["ci-insts"] = true;
Features["dot1-insts"] = true;
Features["dot2-insts"] = true;
Features["dot5-insts"] = true;
Features["dot6-insts"] = true;
Features["dl-insts"] = true;
Features["flat-address-space"] = true;
Features["16-bit-insts"] = true;
Features["dpp"] = true;
Features["gfx8-insts"] = true;
Features["gfx9-insts"] = true;
Features["gfx10-insts"] = true;
Features["gfx10-3-insts"] = true;
Features["s-memrealtime"] = true;


case GK_GFX1010: //Radeon RX 5700 series
Features["dl-insts"] = true;
Features["ci-insts"] = true;
Features["flat-address-space"] = true;
Features["16-bit-insts"] = true;
Features["dpp"] = true;
Features["gfx8-insts"] = true;
Features["gfx9-insts"] = true;
Features["gfx10-insts"] = true;
Features["s-memrealtime"] = true;


case GK_GFX906: //Radeon VII has dot1 and dot2 instructions
Features["dl-insts"] = true;
Features["dot1-insts"] = true;
Features["dot2-insts"] = true;
You are assuming dot 1, 2, 5 and 6 are something that they arnt.

It's very simple.
MS has demonstrated that THEY added hardware to include 4 and 8 bit. MS didn't say THEY added BVH to their GPU. That was an AMD RDNA inclusion.

We have Sony not mentioning it in their tech showcase. Not only that but we have a Sony software engineer say specifically that the PS5 doesn't have ML.

It's really silly to try and make out it does at this point.

It is what it is. Who knows if it will ever manifest in anything tangible, just like FP 16 didn't on the PS4 Pro.

Let's just wait and see.
 
  • Like
Reactions: SoloKingRobert

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
You are assuming dot 1, 2, 5 and 6 are something that they arnt.

It's very simple.
MS has demonstrated that THEY added hardware to include 4 and 8 bit. MS didn't say THEY added BVH to their GPU. That was an AMD RDNA inclusion.

We have Sony not mentioning it in their tech showcase. Not only that but we have a Sony software engineer say specifically that the PS5 doesn't have ML.

It's really silly to try and make out it does at this point.

It is what it is. Who knows if it will ever manifest in anything tangible, just like FP 16 didn't on the PS4 Pro.

Let's just wait and see.
It's very simple, NAVI 10 is missing support for Dotx instructions while it still has Rapid Pack Math's double rate FP16 and INT16.

DXR only uses BVH accelerated structure. LOL

Dot5 is the closest to dot4
Dot6 is the closest to dot8

Radeon VII and Radeon Instinct MI50 have support for dotx instructions and AMD marketed GFX906 "Vega 20" for Deep Learning segments i.e. read https://www.amd.com/en/products/professional-graphics/instinct-mi50#product-specs

Radeon Instinct MI50 (60 CU) has Peak INT8: 53 TOPS.

BiG NAVI GFX1030 has exceeded Radeon Instinct MI50 in dotx instructions support.
 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
Yet Sony confirmed RT hardware. Why won't they do the same for rpm int8/int4?

There's no reason to keep it secret. Anybody with common sense can see why they won't confirm these "standard rdna2 features".
Make a large social media noise. The large social media noise has forced Sony to confirm PS5's hardware raytracing. Sony's transparency is worst than MS's.
 

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395
It's very simple, NAVI 10 is missing support for Dotx instructions while it still has Rapid Pack Math's double rate FP16 and INT16.

DXR only uses BVH accelerated structure. LOL

Dot5 is the closest to dot4
Dot6 is the closest to dot8

Radeon VII and Radeon Instinct MI50 have support for dotx instructions and AMD marketed GFX906 "Vega 20" for Deep Learning segments i.e. read https://www.amd.com/en/products/professional-graphics/instinct-mi50#product-specs

Radeon Instinct MI50 (60 CU) has Peak INT8: 53 TOPS.

BiG NAVI GFX1030 has exceeded Radeon Instinct MI50 in dotx instructions support.
One has nothing to do with the other.
MI50 isn't a games card. Its also Vega based.
So because MIMI5i0s Vega and has Int 8 abilities, the Xbox One and PS4 did as well?

Show me anywhere where AMD has listed Int 4 & 8 in the new big Navi GPUs.

Show me where in it's specs sheets it has it listed. Just like the MI50 has it listed, because it has it, just like Nvidia lists it for their RTX cards, because they have it, and just like MS lists it for the XSX, because it has it.

AMD has only listed FP32 and FP16 for their RDNA 2 cards. Trust me, if they had Int 8 & 4, they would have it in the spec sheets.

It is what it is. It just is.
 

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
One has nothing to do with the other.
MI50 isn't a games card. Its also Vega based.
So because MIMI5i0s Vega and has Int 8 abilities, the Xbox One and PS4 did as well?

Show me anywhere where AMD has listed Int 4 & 8 in the new big Navi GPUs.

Show me where in it's specs sheets it has it listed. Just like the MI50 has it listed, because it has it, just like Nvidia lists it for their RTX cards, because they have it, and just like MS lists it for the XSX, because it has it.

AMD has only listed FP32 and FP16 for their RDNA 2 cards. Trust me, if they had Int 8 & 4, they would have it in the spec sheets.

It is what it is. It just is.




Major differences between Rapid Pack Math vs Dot Math features

Rapid Pack Math's FMA two FP16 resolves into FP16

Dot's FMA two FP16 resolves into FP32
Dot's FMA four INT8 resolves into INT32
Dot's FMA eight INT4 resolves into INT32


Vega 10 (RX Vega 56/64, GFX900) only has Rapid Pack Math's double rate FP16.

PS4 Pro only has Rapid Pack Math's double rate FP16.


PS; GFX908 = CDNA
 
Last edited:

Valued Gamer

Banned
Oct 26, 2020
489
1,055
395



Major differences between Rapid Pack Math vs Dot Math features

Rapid Pack Math's FMA two FP16 resolves into FP16

Dot's FMA two FP16 resolves into FP32
Dot's FMA four INT8 resolves into INT32
Dot's FMA eight INT4 resolves into INT32


Vega 10 (RX Vega 56/64, GFX900) only has Rapid Pack Math's double rate FP16.

PS4 Pro only has Rapid Pack Math's double rate FP16.


PS; GFX908 = CDNA
Again, you are using an unrelated card to RDNA 2.
You are saying that Dot 1, 2, 5 & 6 equates to FP32, FP16, Int 8 and Int 4, when it doesn't.

Let's just stick to AMDs own info.
As you can see, it only lists FP32 and FP16.
That's all it has.
You are trying to convince yourself that this isn't something extra the MS added to the XSX that isn't present on PS5 and RDNA 2 cards.
You won't take Sony's Principle Software engineers word for it. You won't take AMDs word for it, and you won't take Microsoft's word for it.
So if you won't take theirs, you're not going to take mine.

 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
Again, you are using an unrelated card to RDNA 2.
You are saying that Dot 1, 2, 5 & 6 equates to FP32, FP16, Int 8 and Int 4, when it doesn't.

Let's just stick to AMDs own info.
As you can see, it only lists FP32 and FP16.
That's all it has.
You are trying to convince yourself that this isn't something extra the MS added to the XSX that isn't present on PS5 and RDNA 2 cards.
You won't take Sony's Principle Software engineers word for it. You won't take AMDs word for it, and you won't take Microsoft's word for it.
So if you won't take theirs, you're not going to take mine.

From https://www.amd.com/en/products/graphics/amd-radeon-vii#product-specs
AMD only listed TFLOPS (FP) numbers for Radeon VII (Vega 20, GFX906) and it's the same "Vega 20" ASIC design as Radeon Instinct MI50.

The LLVM Compiler Infrastructure project is not a joke.

From https://github.com/llvm-mirror/llvm/blob/master/lib/Target/AMDGPU/AMDGPU.td


def FeatureDot1Insts : SubtargetFeature<"dot1-insts",
"HasDot1Insts",
"true",
"Has v_dot4_i32_i8 and v_dot8_i32_i4 instructions"
>;

def FeatureDot2Insts : SubtargetFeature<"dot2-insts",
"HasDot2Insts",
"true",
"Has v_dot2_f32_f16, v_dot2_i32_i16, v_dot2_u32_u16, v_dot4_u32_u8, v_dot8_u32_u4 instructions"
>;

def FeatureDot3Insts : SubtargetFeature<"dot3-insts",
"HasDot3Insts",
"true",
"Has v_dot8c_i32_i4 instruction"
>;

def FeatureDot4Insts : SubtargetFeature<"dot4-insts",
"HasDot4Insts",
"true",
"Has v_dot2c_i32_i16 instruction"
>;

def FeatureDot5Insts : SubtargetFeature<"dot5-insts",
"HasDot5Insts",
"true",
"Has v_dot2c_f32_f16 instruction"
>;

def FeatureDot6Insts : SubtargetFeature<"dot6-insts",
"HasDot6Insts",
"true",
"Has v_dot4c_i32_i8 instruction"
>;

You're wrong! I gave you hints before I debunked your narrative.
 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
At the cost of consuming CU's themselves, big difference.
1. CU that supports dot4 and dot8 features has less cost than NAVI baseline Rapid Pack Math (RPM) double-rate FP16/INT16.

2. There's a difference in the final results between dot and RPM features i.e. dot function resolves to 32bit datatype results while RPM FP16/INT16 remains in FP16/INT16 datatype results.
 
Last edited:

rnlval

Member
Jun 26, 2017
1,296
1,047
460
Sector 001
gpucuriosity.wordpress.com
Again, you are using an unrelated card to RDNA 2.
You are saying that Dot 1, 2, 5 & 6 equates to FP32, FP16, Int 8 and Int 4, when it doesn't.

Let's just stick to AMDs own info.
As you can see, it only lists FP32 and FP16.
That's all it has.
You are trying to convince yourself that this isn't something extra the MS added to the XSX that isn't present on PS5 and RDNA 2 cards.
You won't take Sony's Principle Software engineers word for it. You won't take AMDs word for it, and you won't take Microsoft's word for it.
So if you won't take theirs, you're not going to take mine.


BiG NAVI supports dot8 INT4 and dot4 INT8.

PS; 128 value is the baseline rate for INT32 and FP32

OPERAND / RESULTMODEOPS/CYCLE/CU
FP16/FP16Packed256
FP16/FP32Mixed Precision256
FP32Native128
FP64Native8
Int64Native32
Int32Native128
Int16/Int16Packed256
Int16/Int32Mixed Precision256
Int8/Int32Mixed Precision512
Int4/Int32Mixed Precision1024
 
Last edited:

RockOn

Member
Feb 18, 2021
107
127
250
UK
DirectML short for Direct Machine Learning is AMD equivalent answer to Nvidia's DLSS short for Deep Learning Super Sampling, Enabling higher performance levels by using machine learning to upscale images to higher resolutions without visual downgrades.
AMD has an answer to Nvidia's DLSS, and that answer is DirectML-powered Super Resolution. AMD plans to utilise Machine Learning to improve the visual quality of games, & AMD's solution will have the backing of Microsoft.

Microsoft confirmed that both of their next-generation consoles would support Machine learning for games with DirectML. Through their collaboration with AMD when creating their new consoles.
A component of Microsoft's DirectX feature set DirectML isn't a Radeon-only technology, and its applications extend far beyond Super Resolution functions. Over the coming years, future PC and Xbox games will bring Machine Learning into games in several new and innovative ways, impacting all gamers with supported hardware.

Microsoft has already showcased the potential of machine learning in gaming applications, with the image below showcasing what happens when Machine Learning is used to upscale an image to four times its original resolution (basically from 1080p to 4K) to generate a sharper final image with reduced aliasing. The image below is a comparison between ML Super Sampling and bilinear upsampling.

PS5 is confirmed to not use DirectX feature so even though AMD made their CPU and GPU for the PS5 it looks like PS5 will not have DirectML
so will Sony come up with alternative method to compensate for not having this feature cause Machine Learning is the future of gaming.
While DirectML hasn't received as much attention as DirectX raytracing, you can be sure that developers are looking at the new API closely. As screen manufacturers are starting to push beyond 4K, AI upscaling technologies like Machine Learning will continue to increase in popularity. 4K gaming is already a challenge for modern gaming hardware, and 8K is going to prove to be even more problematic for game makers and hardware vendors.
Technologies like DirectML will become vital for future games, both on PC and on consoles. The application of machine learning will allow developers to deliver higher levels of graphical fidelity without the insane hardware costs of traditional computational methods.
AMD open ML is open source and will run on any platform. Thats the reason AMD super resolution is cross platform. Don't get why others think that PS5 Custom APIs GNM/GNMX won't get update to support PS5/AMD custom hardware features
 

RockOn

Member
Feb 18, 2021
107
127
250
UK
That statement is false.

From https://wccftech.com/david-cage-on-quantic-dream-future-next-gen-gaming-and-not-ruling-out-sequels/

Wccftech Question: As developers, what do you think about the PlayStation 5 and Xbox Series X? Which unique aspects of each platform excite you the most?

CEO David Cage of Quantic Dream's Answers: 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).

---------------

Baseline NAVI has a level of Rapid Pack Math.




Baseline NAVI has 2X rate INT16/FP16 rapid pack math.

4X rate INT8 and 8X rate INT4 are optional features for NAVI. MS requested AMD to include the optional NAVI dot4 and dot8 feature for XSX.
Wrong, mixed integer int4/int8 etc are all standard basic features of RDNA 2 enhanced CUs. As PS5 is RDNA2 its in PS5