Support NeoGAF

[Three]

Announcing DirectX 12 Ultimate - DirectX Developer Blog

It is time for DirectX to evolve once again. From the team that has brought PC and Console gamers the latest in graphics innovation for nearly 25 years, we are beyond pleased to bring gamers DirectX 12 Ultimate, the culmination of the best graphics technology we’ve ever introduced in an...

devblogs.microsoft.com

I've already told you what you're not explaining and not understanding. sampler feedback isn't unique to MS. It is not even all that different to the the feedback used in megatextures. You still cannot show this and we go round in circles. I can guess what will happen next even: Somebody will point to 'custom hardware' that is about keeping more textures in memory as a failsafe for failing to load that texture in the frametime.

 

[Bernkastel]

You keep appealing to authority on those but they don't describe what you think it's describing.
SFS is streaming textures you need, so is UE5, so is idtech 6, the question you're not answering because you can't is how you believe this to be 2x more efficient in determining what is and isn't needed compared to current hardware. The answer is it isn't. That may upset you but this number comes from previous gen games on Xbox one as clearly stated by MS. Previous gen games held a lot more in memory because they had a slow ass drive so ALMOST ALL games and engines kept textures in memory that were not visible in the scene because the player could turn or move faster than you could load them in.

Then why are taking authority on a non-literal interpretation of what he said ?

 

[THE:MILKMAN]

I think a few here are just setting themselves up for disappointment unnecessarily. I'm sure the custom texture filters are a useful win in any case.

 

[funking giblet]

Yes. By loading only textures that are needed, as they are needed, with fine granularity.

But I guess what we do need to ignore is the fact that they also mentioned the XSX has custom hardware, that sampler feedback is a DX12U spec but sampler feedback streaming isn't, the fine granularity part, and most importantly, that you couldn't be accurate at all in the past with similar methods (basically killing the benefit of it).


That a disk is slower than RAM does not mean that transferring from disk to GPU would not be faster than transferring from disk to RAM to GPU. And that was the whole point.


They said nothing close to it? How come both these people bring up how the NES works as a reference?


As for that person not being a 'source'... Obviously, not an official one. Whether credible info is provided by this person or not is up for debate, although some people are very keen on stopping that debate completely. If what Ronaldo8 at Beyond3D said in this post was considered credible, and somehow that other one is considered complete nonsense because it aligns with what Kirby Louise said, then I don't know what else there is to talk about.

This thread was fine, but again we have people coming in here dictating what is allowed and not allowed to be talked about. So yeah. Another thread that was going fine with great sharing of information, ruined because of ego boosting.

I'm out.

FUCKING BANK SWITCHING?

Stop just stop dude. You are embarrassing yourself now.



edit: Ok look, I don't mean to be harsh. I would rather help you understand, but we cannot go around in circles. Pulling bank-switching out of the air is nowhere near what Ronaldo8 was saying, and then his solution doesn't offer a solution at all. The really important thing isn't that Virtual Memory is slow, we know that. The new hotness is that a good enough SSD, good enough latency and some clever techniques bring it back into focus around how the fetch assets in the first place. Bank switching is something we did in single threaded, tiny memory, tiny data environments to extend memory addressing into a ROM, it has been superseded by ... Virtual Memory. RAM : SSD ratio is too high to do anything else. And that's leaving out that these ROMs were authored to support such a thing. Assets these days are not anywhere near the data structures they need to be to be directly addressable (hell, they are compressed first of all!).

Actually speaking of which, did you ever play SFA2 on the SNES? It used bank switching and compressed assets. SFA2 for the SNES had load times! You couldn't dynamically switch banks, otherwise you lose access to the other banks!

 

[Dodkrake]

For the Facts:
1. That is not a fact
2. incorrect, and bullshit to be honest
3. yes
4. Incorrect.

what to expect:
1. Possibly, we don't know
2. Possibly, we don't know
3. Possibly, we don't know
4. Bullshit.
5. Possibly, we don't know

I'm going to quote myself, as it looks like this thread is heavy on shooting first and asking questions later:

Facts are, with what we have so far:

1. Sony invested a lot more in Custom IO to reduce bottlenecks
2. Microsoft preferred to invest in SW applications
3. HW, unless stupidly poor engineering, will beat SW in most applications where the usage is similar
4. Most techs being presented as "Xbox exclusives" have had similar implementations patented by Sony

Please pay special attention to my wording. I said "with what we have so far". Special emphasis on so far.

So yes, these are the facts:

1. It is a fact. More custom IO = more R&D + higher costs with said IO = more investment
2. Since they are pushing the software angle, and Sony has barely provided any info on their specific SW implementations, if any of them are as relevant as Microsoft's ones, this is indeed a fact.
3. Glad we agree on something
4. I'm not going to scout this thread for implementations, but suffice to say that Sony has their own VRS implementation, for example. So yes, most techs (again, wording is important) being pushed by MS have a similar patent from Sony's side. If said tech will be used is a completely different subject.

As for the next action items, please pay attention to my wording. I'm not going to discuss all 5 because you said possibly, but again, what I said is important. I'm going by the numbers, not the actual real world implementation, because that will become apparent when we see the games, especially 3rd party non-exclusive ones.

What to expect, with what we have so far

1. Xbox will have the edge in RT and resolution
2. PS5 will have the edge in sound and asset streaming
3. Third party games will probably look roughly the same
4. Game design in first party games will likely improve way more in Sony's camp, as Xbox needs to scale game design to the Xbox One and PC
5. MS will likely have an edge in graphics for their first party games (not by much)

My 2 cents.

Addressing 4 - You must understand that while MS is going with a strategy of supporting all consoles since the Xbox One, game design (not to be confused with graphics) will likely (again, I bolded this just so that you get the proper context) improve way more in Sony's camp. If we want to be pedantic, I concede that this is for games that will be available in all platforms and not Series X only games (which seem to be none, at the moment).

The facts

1. Game design will, as stated by many devs, improve substantially as the throughput of SSD is orders of magnitude higher when compared to HDD's
2. MS is investing on multi gen games which are constrained by HDD's
3. Therefore, game design will not drift much from said games

Example: The flying scene in the Unreal 5 tech demo VS devs having to scrap flying in Horizon Zero Dawn. Going by Sony's exclusives, you wouldn't design the same game for 2 platforms around one having flight (PS5) and the other not having flight (PS4).

 

[OptimistPrime]

You keep appealing to authority on those but they don't describe what you think it's describing.
SFS is streaming textures you need, so is UE5, so is idtech 6, the question you're not answering because you can't is how you believe this to be 2x more efficient in determining what is and isn't needed compared to current hardware. The answer is it isn't. That may upset you but this number comes from previous gen games on Xbox one as clearly stated by MS. Previous gen games held a lot more in memory because they had a slow ass drive so ALMOST ALL games and engines kept textures in memory that were not visible in the scene because the player could turn or move faster than you could load them in.

• Isn't idtech 6 running on current generation hardware?
• You claim I'm relying on faith in order to not assume Microsoft's 2X-3X claim is for all texture streaming systems. Where is your proof that the 2-3X claim does not apply to other texture streaming systems?
• No one is claiming that the 2-3X is in comparison to other next generation systems, but you seem to be setting up that straw man to use the XBox One X baseline as an argue against.

Here's a little one act play I wrote to illustrate the absurdity of this conversation:

Microsoft: "Hey guys we made something that let's you use 2X-3X less memory and bandwidth."
Me: "Oh cool, that sounds impressive. I guess I'll see how accurate that is when more information and demonstrations are available."
Three: "FOOL! They made NOTHING! What they were secretly saying is that just like Rage, Doom 2016. Titanfall 2, and to some extent nearly every game of the last 7 years you can stream texture information from disk on our new consoles. They were just reminding everyone how cool that is and that you can do it on Xbox Series X. You have to be an IDIOT and a crazy person to think that Microsoft would be giving you any kind of expectation about how their new product is new or better in any way."

 

[Segslack]

The interesting thing is that the same user is describing what Kirby Louise said as a real possibility in this post.

One important difference between MS and Sony's I/O solution is MS's claim to be able to transfer data directly from the SSD to the GPU. The claim of 100 GB of NAND SSD being instantly available is brought to mind. The questions are then:
(1) What does the qualifier "instantly" means in this context?
(2) What is exactly being made "available" ?
(3) For what purpose?

The careless observer will just wave it away by saying that this is just good old virtual memory paging; that is not in fact the direct transfer of data from the SSD to the GPU . However, the idea of virtual paging does not stand up to scrutiny in this case.
Reason 1: There is nothing particularly instantaneous about virtual memory paging. It describes a tortuous circuit whereby the CPU will have to acknowledge a page fault, look through the filesytem to find the requested page on the SSD, find an empty frame in main memory or evict a stale page to create one and then swap in the correct page from the SSD. Yea, nothing to brag about in terms of instantaneousness.
Reason 2: Phil Spencer, in an otherwise mundane interview in December 2019 drops an absolute bombshell: the SSD of the upcoming Xbox can be used as virtual RAM. Now this can either mean a matching of a page on the SSD to an address in the physical memory address space which remains unchanged (virtual memory paging) OR the memory mapping of a portion of the SSD (100 GB) of it and its addition to the physical memory address space contiguous with system RAM. Phil Spencer specifically mentions that the SSD will act virtually as RAM by significantly increasing the physical memory address space, comparing it to the 32 to 64 bit transition for good measure. Thus, it becomes highly probable that MS has succeeded into making a part of an NVME SSD byte-addressable which cuts down significantly on the CPU overhead associated with virtual memory as the CPU likely can't differentiate between system RAM and the SSD.

This type of technology is not unprecedented in consoles, that's how the ROM cartridge of the good old NES functioned. Nowadays it finds an echo in a field far removed from gaming: big data and AI systems. The addressable SSD is what can be described as persistent memory, a technology now ubiquitous with dual socket servers being used for RDMA. Tom Talpey of Microsoft is actually a good source for the ongoing effort to develop a new filesystem API for presistent memory when in memory mode. This is it for the term of art 'instantaneous'.
Now what is this data available for? I speculate that it is available to be duplicated back to another portion of the physical memory space which is system RAM (the CPU will view it just as a duplication of data from one RAM address to another) and/or streaming of textures from the SSD to the GPU as part of SFS. One interesting result of this aspect of the XVA is that it doesn't actually requires the use of coherency engines or GPU scrubbers.

Maybe a dumb question, but is she implying that only Xbox Series X can transfer data from the SSD directly to the GPU?

How she came to this conclusion without have access to the PS5 Dev kit?

I thought UE5 was streaming 8k textures directly from the SSD.

 

[Three]

Then why are taking authority on a non-literal interpretation of what he said ?

Because you're applying two separate and explained things and relating them, whereas I'm not.

• Isn't idtech 6 running on current generation hardware?
• You claim I'm relying on faith in order to not assume Microsoft's 2X-3X claim is for all texture streaming systems. Where is your proof that the 2-3X claim does not apply to other texture streaming systems?
• No one is claiming that the 2-3X is in comparison to other next generation systems, but you seem to be setting up that straw man to use the XBox One X baseline as an argue against.

A) yes it is but ultimately still relies on a HDD on these systems.

B) you're the one making the claim not me. All MS have said is that on xbox one games 2x 3x the textures were in memory but not visible in the scene. I know for a fact that this is the case in almost all current engines and games because of the HDD speed meaning needing to store a lot more than currently visible. On top of that some didn't even use megatextures like streaming. Look at any GDC talk about streaming.

C) you must be kidding me. people in this thread linked that to custom secret hardware for next gen that will close the gap specifically referring to that figure. Just look at Lady Bernkastels comments for one and that other banned user Rtnago or whatever his name was which was what I originally replied to, or even you yourself going in circles when I mentioned where I stand on this argument by clarifying that my main point of contention is the 2x figure and linking that to custom hardware. you kept arguing round in circles pointing to MS implementation in SFS and saying 'but we don't know if it has it' though so I know where you stand. Instead of coming to an agreement that yes we can load only textures we need in games like Rage but in most games we need to load more due to HDD speed you go off on SFS and custom hardware and not knowing if PS5 has it. Never providing any information on how you think it provides double the efficiency, even pointing to SSD bandwidth eating features for some reason to get the 'custom' in there.

.

Here's a little one act play I wrote to illustrate the absurdity of this conversation:

Microsoft: "Hey guys we made something that let's you use 2X-3X less memory and bandwidth."
Me: "Oh cool, that sounds impressive. I guess I'll see how accurate that is when more information and demonstrations are available."
Three: "FOOL! They made NOTHING! What they were secretly saying is that just like Rage, Doom 2016. Titanfall 2, and to some extent nearly every game of the last 7 years you can stream texture information from disk on our new consoles. They were just reminding everyone how cool that is and that you can do it on Xbox Series X. You have to be an IDIOT and a crazy person to think that Microsoft would be giving you any kind of expectation about how their new product is new or better in any way."

Thanks for proving my point that you think MS 'made' something that gets you the 2x 3x efficiency. Still shows you haven't understood what you are reading.

 

[Panajev2001a]

So conjecture. Got it.

That is not how it works, you cannot make something up/provide no proof for your own conjecture or interpretation of reality and if people do not bite and provide reasons for that put their reasoning on equal footing with your conjecture. If you want to make a point, support it, the burden of proof is on those that say that this 2-3x improvement is on top of PRT/Virtual Texturing.

You are taking a statement from MS, implying that the baseline for the benchmark is 2-3x higher than what it is (thus making the achieved number greater too) and expecting people to take it as is.

Bernkastel nobody is saying that SF is not an improvement over PRT or SFS an improvement over SF for the texture streaming case, people are arguing on the magnitude of this improvement and you and others are recycling the same arguments every X pages with the same kind of arguments as if you were reading off a playbook. Add something new at least...

 

[Panajev2001a]

Announcing DirectX 12 Ultimate - DirectX Developer Blog

It is time for DirectX to evolve once again. From the team that has brought PC and Console gamers the latest in graphics innovation for nearly 25 years, we are beyond pleased to bring gamers DirectX 12 Ultimate, the culmination of the best graphics technology we’ve ever introduced in an...

devblogs.microsoft.com

Am I supposed to take it as evidence for 200-400% improvement over PRT / software driven Virtual Texturing? This is like hidden Xbox One dGPU all over again, one would have thought that winning the TFLOPS war would have made you happy enough not to engage in that stuff again... .

 

[OptimistPrime]

...even you yourself going in circles when I mentioned where I stand on this argument by clarifying that my main point of contention is the 2x figure and linking that to custom hardware. you kept arguing round in circles pointing to MS implementation in SFS and saying 'but we don't know if it has it' though so I know where you stand. Instead of coming to an agreement that yes we can load only textures we need in games like Rage but in most games we need to load more due to HDD speed you go off on SFS and custom hardware and not knowing if PS5 has it. Never providing any information on how you think it provides double the efficiency, even pointing to SSD bandwidth eating features for some reason to get the 'custom' in there.

Let's just contrast your claims to my claims.

Here is what you're claiming as far as I can tell:

• Texture streaming systems are perfect at knowing precisely which MIP level or tile to request from disk. No meaningful improvement in accuracy is possible.
• Sampler Feedback is not a new hardware feature, and if it somehow is, it provides no efficiency improvement to texture streaming systems.
• Sampler Feedback does not make texture sampling cheaper in any way - if Sampler Feedback allows you to sample more then it uses SSD more bandwidth.
• Sampler Feedback Streaming is a fancy marketing dress on entirely existing techniques, only benefitting from an SSD and nearly worthless new filters.
• Microsoft's 2X-3X efficiency improvement claim about SFS is actually applicable to all consoles and has nothing to do with SFS.

Here is what I'm claiming:

• It doesn't seem that crazy that there is room for efficiency improvements in texture streaming.
• Microsoft and nvidia are probably not lying when they refer to Sampler Feedback as a new GPU hardware feature that improves texture streaming.
• Microsoft is probably not lying when they say that Sampler Feedback allows "a shader to efficiently query what part of a texture would have been needed to satisfy a sampling request, without actually carrying out the sample operation."
• Microsoft probably didn't just make up Sampler Feedback Streaming purely as marketing drivel to say that "SSDs make streaming better", and there's probably actually some substance to this featureset.
• Microsoft intended to claim that SFS gives them a 2X-3X improvement in efficiency.

Thanks for proving my point that you think MS 'made' something that gets you the 2x 3x efficiency. Still shows you haven't understood what you are reading.

This is what Microsoft said. Are you pretending they didn't say this, or are you saying that Microsoft is lying?

Sampler Feedback Streaming (SFS) - A component of the Xbox Velocity Architecture, SFS is a feature of the Xbox Series X hardware that allows games to load into memory, with fine granularity, only the portions of textures that the GPU needs for a scene, as it needs it. This enables far better memory utilization for textures, which is important given that every 4K texture consumes 8MB of memory. Because it avoids the wastage of loading into memory the portions of textures that are never needed, it[Sampler Feedback Streaming] is an effective 2x or 3x (or higher) multiplier on both amount of physical memory and SSD performance.

 

[Dory16]

Wow. People taking personal offense at the idea that Microsoft may be even remotely correct about its own product description. The 2-3x efficiency multiplier talking point came directly from Microsoft's slides and even from an interview if I'm not mistaken. And you have 40 pages worth of people fighting it off with their lives. It went from "Prove that they said it" to actually dissecting every component in the unreleased and unseen velocity architecture to "Every system can do it" . All this with the dogma that "Microsoft's I/O has to be slow enough for me to satifyingly live with the fact that the PS5's processors and bus are slower" and then making their arguments from there.
This has been pretty pathetic, folks. Especially from some of the same people who take everything stated during the "Road to PS5" presentation at face value.
You will all be fine regardless of what systems performs better and it's very likely that both consoles will amaze in their own way.

 

[Jaguar Victory]

I thought they got a new PR/Marketing team and strategy? I guess not, same ole Microsoft in action.
By August it will already be game over in terms of specs. Everyone would have pre-ordered and made up their mind. If you are getting a PC and latest Nividia/Intel/Amd you have already made up your mind, If you are getting a XSX or PS5 nothing will change it either.

While people criticism Cerny. He did a brilliant thing giving that talk and laying out what they did for their SSD, I/O solution and how complex it is.
Making everything plain. Its not surprising that SSD has taken over the discourse. If that talk didn't exist it wouldn't have.
Of-course MS being MS will take forever to respond and actually will never respond. Knowing MS from the past behavior in tech (not just consoles). The corporates are still running the show.

If a real down to earth small company were in charge. They would move up parts of that Hot chip presentation like yesterday and do what Cerny did including a Tech demo of their own specifically to outline their SSD/IO solution.

Its hilarious that people have to read between the lines on BCPACK, SFS and heck what the hell is Direct Storage anyway. Sure we know the general description but nothing tangible. When is it coming to PC. How will the PC market utilize it and how will it improve the current crop of SSD NVME in PCs. Simple stuff. But it will take months to get approval due to red tape. Just goes to show you that the same leadership is still in charge unfortunately.

It's cute that you think more than 0.01% of customers want to think about the underlying tech so much. We're an anomaly.

 

[thicc_girls_are_teh_best]

Wow. People taking personal offense at the idea that Microsoft may be even remotely correct about its own product description. The 2-3x efficiency multiplier talking point came directly from Microsoft's slides and even from an interview if I'm not mistaken. And you have 40 pages worth of people fighting it off with their lives. It went from "Prove that they said it" to actually dissecting every component in the unreleased and unseen velocity architecture to "Every system can do it" . All this with the dogma that "Microsoft's I/O has to be slow enough for me to satifyingly live with the fact that the PS5's processors and bus are slower" and then making their arguments from there.
This has been pretty pathetic, folks. Especially from some of the same people who take everything stated during the "Road to PS5" presentation at face value.
You will all be fine regardless of what systems performs better and it's very likely that both consoles will amaze in their own way.

This always happens. There's a distinct double-standard applied to any claims by Sony and MS when it comes to the next-gen consoles. All claims from Sony (specifically Cerny) are to be assumed true until proven otherwise. All claims from Xbox engineers are to be assumed false until proven otherwise.

When certain theories are posed regarding rationale explanations on how MS might be able to justify their claims, they are quickly and routinely shot down by the same few people, sometimes viciously so as we can see some few posts near the top of this very page, which seem to only speculate on reasoning related to much older hardware with much older technologies (let alone different architectures altogether that have their own documented bottlenecks which could've impacted other implementations including, yes, bank-switching on cartridges, depending on how the specific game was programmed and what limitations there were with its cart co-processor).

Meanwhile, in any efforts to repudiate aspects related to PS5, many of them just tend to repeat what Mark Cerny himself has said, rather than saying these things in their own words. Which at least would show one has put time into their own research on the topic, because regardless how smart or clever a person can be, they are not infallible (nor may their methodology be the only correct one to implement).

It's kinda become a wolf-in-sheeps'-clothing thing at this point; I have some evolving concepts on the systems in a multitude of areas, but probably gonna need some time to gather them all into an organized state. Would help if we had more concrete information on the more unknown aspects of these systems but hey, beggars can't be choosers xD.

Exodia

There's some rumors going around they might speak more on XvA in June, but they should be taken with a giant grain of salt.

Can't speak much on the management side of things; personally I wouldn't of mind a small gameplay teaser for Halo: Infinite early next month ahead of the July event to show off some genuine next-gen gameplay for a game actually in development, visually as good (at least) with UE5 demo and showing off the investments in SlipSpace Engine and a taste of what it can do. Because personally I think waiting until July to show off more gameplay is probably going to hurt MS especially if Sony's June event turns out as good as it's sounding it'll be (got my fingers crossed for Tomba 3. Hey, I can dream alright!).

One thing I hope MS avoids is getting into a "Yeah well we can do that too!" sort of tit-for-tat with Sony. They won't get anywhere basically trying to show they can do the same thing Sony does because that cedes control of the conversation to Sony. If they want to do that to quell doubts, it has to be paired with them showing or doing something that's aside from the current discourse that puts them in a position of controlling the conversation.

That way there's a bit of acknowledgement on their part but also a lot of willingness to display leadership, because only the company that best displays leadership will do the best this coming generation.

JareBear: Remastered Yeah I saw Redtech Gaming's vid (also watching the BRAP podcast they were on; pretty good stuff and a bit edgier) and I hope they can confirm some of the other PS5/XSX rumors they have heard from their sources too, particularly regarding APU customizations on the CPU/GPU front.

I still don't think XvA will close the SSD I/O performance gap between XSX and PS5, but I do think that delta will be a lot smaller than some think, provided devs actually target XvA's feature set. Both systems are going to have very good SSD I/O solutions with their own methods that work best for their respective platform holders and their market interests in mind.

 

[Three]

Let's just contrast your claims to my claims.

Here is what you're claiming as far as I can tell:

• Texture streaming systems are perfect at knowing precisely which MIP level or tile to request from disk. No meaningful improvement in accuracy is possible.
• Sampler Feedback is not a new hardware feature, and if it somehow is, it provides no efficiency improvement to texture streaming systems.
• Sampler Feedback does not make texture sampling cheaper in any way - if Sampler Feedback allows you to sample more then it uses SSD more bandwidth.
• Sampler Feedback Streaming is a fancy marketing dress on entirely existing techniques, only benefitting from an SSD and nearly worthless new filters.
• Microsoft's 2X-3X efficiency improvement claim about SFS is actually applicable to all consoles and has nothing to do with SFS.

Here is what I'm claiming:

• It doesn't seem that crazy that there is room for efficiency improvements in texture streaming.
• Microsoft and nvidia are probably not lying when they refer to Sampler Feedback as a new GPU hardware feature that improves texture streaming.
• Microsoft is probably not lying when they say that Sampler Feedback allows "a shader to efficiently query what part of a texture would have been needed to satisfy a sampling request, without actually carrying out the sample operation."
• Microsoft probably didn't just make up Sampler Feedback Streaming purely as marketing drivel to say that "SSDs make streaming better", and there's probably actually some substance to this featureset.
• Microsoft intended to claim that SFS gives them a 2X-3X improvement in efficiency.

This is what Microsoft said. Are you pretending they didn't say this, or are you saying that Microsoft is lying?

MS have perfectly clarified this here already:

"textures have ballooned in size to match 4K displays, efficiency in memory utilisation has got progressively worse - something Microsoft was able to confirm by building in special monitoring hardware into Xbox One X's Scorpio Engine SoC. "From this, we found a game typically accessed at best only one-half to one-third of their allocated pages over long windows of time," says Goossen. "So if a game never had to load pages that are ultimately never actually used, that means a 2-3x multiplier on the effective amount of physical memory, and a 2-3x multiplier on our effective IO performance."

You're just failing to understand what the comparison is.

I can only explain your issue with the subject with an analogy and your theatrics. Hopefully that would help.

Once upon a time before trains were standard:
MS: "by measuring the time it took on a horse and carriage we realised it takes 10 hours to get to London ."
"if we use Anglia rail trains we could get there in 5 hours."

You: Anglia trains must be providing the 2x speed improvment! Could you even get that on other trains.

Me: what makes this train 2x faster than National rail trains?

You: the fact that Anglia trains stop more frequently, Why don't you believe what they claim?!

Me: No I believe their claim but I think you've misinterpreted what is offering the speed improvement.

You: no this is custom MS made stuff and we don't know if Sony has it. I'm an optimist. I want to believe Anglia trains!


Look the simple answer is that the paradigm in next gen game development has changed mainly due to SSD. Any new tech or custom hardware may provide additional benefits but you're misinterpreting where most of the 2x is coming from.

Example from road to PS5 slides

"Lots of inactive RAM/Most RAM in active use. "
The paradigm shift is not some breakthrough from MS in SFS. Something else brought about this paradigm shift. Guess what?

 

[Thirty7ven]

This always happens. There's a distinct double-standard applied to any claims by Sony and MS when it comes to the next-gen consoles. All claims from Sony (specifically Cerny) are to be assumed true until proven otherwise. All claims from Xbox engineers are to be assumed false until proven otherwise.

Absolute bullshit. Overall that is. Really need a short memory to claim that.

Can't speak much on the management side of things; personally I wouldn't of mind a small gameplay teaser for Halo: Infinite early next month ahead of the July event to show off some genuine next-gen gameplay for a game actually in development, visually as good (at least) with UE5 demo and showing off the investments in SlipSpace Engine and a taste of what it can do. Because personally I think waiting until July to show off more gameplay is probably going to hurt MS especially if Sony's June event turns out as good as it's sounding it'll be (got my fingers crossed for Tomba 3. Hey, I can dream alright!).

Really this is your expectation? Putting a lot of pressure on 343i there. Sounds like a very unrealistic expectation but maybe we will be surprised. They have shown the game before, but yeah.

JareBear: Remastered Yeah I saw Redtech Gaming's vid (also watching the BRAP podcast they were on; pretty good stuff and a bit edgier) and I hope they can confirm some of the other PS5/XSX rumors they have heard from their sources too, particularly regarding APU customizations on the CPU/GPU front.

I still don't think XvA will close the SSD I/O performance gap between XSX and PS5, but I do think that delta will be a lot smaller than some think, provided devs actually target XvA's feature set. Both systems are going to have very good SSD I/O solutions with their own methods that work best for their respective platform holders and their market interests in mind.

Well yeah, both have good I/O solutions, good GPUs, good CPUs, good Audio processors, good everything.

 

[JareBear: Remastered]

It's kinda become a wolf-in-sheeps'-clothing thing at this point; I have some evolving concepts on the systems in a multitude of areas, but probably gonna need some time to gather them all into an organized state. Would help if we had more concrete information on the more unknown aspects of these systems but hey, beggars can't be choosers xD.

"Concrete information" seems to be a rare commodity right now but that will hopefully change within the next two months

 

[Trueblakjedi]

This is all very interesting but doesn't relate to the SSD bandwidth. This would actually hammer the SSD more or at best the same because you have things not even resident in memory. This would be a compute overhead saving not an SSD bandwidth saving.

Unless... the memory addresses are independent NVME locations.

We won't have this discussion if you actually bother to read the documentation.

Terminology
Use of sampler feedback with streaming is sometimes abbreviated as SFS. It is also sometimes called sparse feedback textures, or SFT, or PRT+, which stands for “partially resident textures”.

Plus... don't forget the "+"

What does the plus refer to?

Its PRT plus something(s) which do not currently exist in PRT.

 

[Trueblakjedi]

Maybe a dumb question, but is she implying that only Xbox Series X can transfer data from the SSD directly to the GPU?

How she came to this conclusion without have access to the PS5 Dev kit?

I thought UE5 was streaming 8k textures directly from the SSD.

SSD to VRAM

Supposedly the XSX can do that plus direct to GPU skipping over VRAM.

 

[Three]

Unless... the memory addresses are independent NVME locations.

What do you mean by this? Why would this matter with a given read/write speed/number of channels?

 

[Trueblakjedi]

MS have perfectly clarified this here already:



You're just failing to understand what the comparison is.

I can only explain your issue with the subject with an analogy and your theatrics. Hopefully that would help.

Once upon a time before trains were standard:
MS: "by measuring the time it took on a horse and carriage we realised it takes 10 hours to get to London ."
"if we use Anglia rail trains we could get there in 5 hours."

You: Anglia trains must be providing the 2x speed improvment! Could you even get that on other trains.

Me: what makes this train 2x faster than National rail trains?

You: the fact that Anglia trains stop more frequently, Why don't you believe what they claim?!

Me: No I believe their claim but I think you've misinterpreted what is offering the speed improvement.

You: no this is custom MS made stuff and we don't know if Sony has it. I'm an optimist. I want to believe Anglia trains!


Look the simple answer is that the paradigm in next gen game development has changed mainly due to SSD. Any new tech or custom hardware may provide additional benefits but you're misinterpreting where most of the 2x is coming from.

Example from road to PS5 slides

"Lots of inactive RAM/Most RAM in active use. "
The paradigm shift is not some breakthrough from MS in SFS. Something else brought about this paradigm shift. Guess what?

I don't know why this has been built up as even a conversation of baselines.

The baseline is the texture map.

If the texture map is 10MB then the XSX will sample and request only the page that is necessary. Let's say that page is 4MB. So instead of compressing 10MB its only compressing a 4MB page.

Now apply BCPACK to that page and the stream requirement is only 2MB. Thats an 80% reduction in stream requirements and now you can fill the other 8 MB request with either an 8 Mb texture 2 4 MB or 4 2MB etc...

The multiplier is axiomatic not referential.

 

[Three]

I don't know why this has been built up as even a conversation of baselines.

The baseline is the texture map.

If the texture map is 10MB then the XSX will sample and request only the page that is necessary. Let's say that page is 4MB. So instead of compressing 10MB its only compressing a 4MB page.

Because people are making it a conversation of baselines. When I stated that their baseline isn't correct strangley you were liking those comments that disagreed with me:

Xbox Velocity Architecture - 100 GB is instantly accessible by the developer through a custom hardware decompression block

That was how it worked in gens past. Its easier to develop for the most limited platform and port that code to others. For SSD matters, multiplats can just use downgraded textures where necessary if Xsex should have trouble keeping up. Personally I don’t think it will. Couldn't it make...

www.neogaf.com

 

[Dory16]

MS have perfectly clarified this here already:



You're just failing to understand what the comparison is.

I can only explain your issue with the subject with an analogy and your theatrics. Hopefully that would help.

Once upon a time before trains were standard:
MS: "by measuring the time it took on a horse and carriage we realised it takes 10 hours to get to London ."
"if we use Anglia rail trains we could get there in 5 hours."

You: Anglia trains must be providing the 2x speed improvment! Could you even get that on other trains.

Me: what makes this train 2x faster than National rail trains?

You: the fact that Anglia trains stop more frequently, Why don't you believe what they claim?!

Me: No I believe their claim but I think you've misinterpreted what is offering the speed improvement.

You: no this is custom MS made stuff and we don't know if Sony has it. I'm an optimist. I want to believe Anglia trains!


Look the simple answer is that the paradigm in next gen game development has changed mainly due to SSD. Any new tech or custom hardware may provide additional benefits but you're misinterpreting where most of the 2x is coming from.

Example from road to PS5 slides

"Lots of inactive RAM/Most RAM in active use. "
The paradigm shift is not some breakthrough from MS in SFS. Something else brought about this paradigm shift. Guess what?

They very clearly say in their slides that they are able to predict the parts of textures that will be needed and load only that. saving memory space and SSD bandwith. I think everybody understands that because quite frankly they made it very understandable.
If we also all understand and accept that 100GB will be instantly accessible to the developers until an actual XSX developer with experience on the dev kits directly disputes it, I'm sure everybody will get along.
Thanks for your maturity.

 

[Trueblakjedi]

Because people are making it a conversation of baselines. When I stated that their baseline isn't correct strangley you were liking those comments that disagreed with me:

Xbox Velocity Architecture - 100 GB is instantly accessible by the developer through a custom hardware decompression block

That was how it worked in gens past. Its easier to develop for the most limited platform and port that code to others. For SSD matters, multiplats can just use downgraded textures where necessary if Xsex should have trouble keeping up. Personally I don’t think it will. Couldn't it make...

www.neogaf.com

The deniers are making it question of baselines or benchmarks against past/current/future systems.

The proponents aren't saying what you are saying. From the perspective of the average reader, the multiplier applies to the 2.4 GB/s Raw number. MS saysy that they can move in that 2.4GB/s rate, what it would take another system (regardless of system) 7+ GBs or 3 seconds @ 2.4GB.

That's a win for the XSX in the face of the massive advantage that PS5 has in terms of raw and compressed speeds.

 

[OptimistPrime]

MS have perfectly clarified this here already:

...

You're just failing to understand what the comparison is.

This Eurogamer article where Microsoft supposedly "clarified" the multiplier claim was published the same day as the "Series X Technical Glossary" that I was quoting.

A few sentences after your selection in that article is this:

Microsoft considers these aspects of the Velocity Architecture to be a genuine game-changer, adding a multiplier to how physical memory is utilised.

It takes a lot of leaning over backwards to interpret this the way you seem to want to. Let me give you some options:

• Maybe "these aspects of Velocity Architecture" REALLY means "the new paradigm brought to us by SSDs in general"
• How about "a genuine game-changer" simply means "a small improvement over SSDs alone"
• Or maybe the easiest approach for you would be to say that "adding a multiplier" really just means a multiplier of 1.05.

 

[Trueblakjedi]

What do you mean by this? Why would this matter with a given read/write speed/number of channels?

I'm NOT saying that the following is what XVA is.

I am saying that there are designs that will allow MS to do what they predict and not be restricted in the way you see to be predicting.

 

[Trueblakjedi]

The paradigm shift is not some breakthrough from MS in SFS. Something else brought about this paradigm shift. Guess what?

PCIE4.0?

In a console? Because if you go into the actual paradigm thats the enabling factor. Put as many lanes as you like at 2.x GB each and you get this benefit no matter what you attach....
I'm not even being facetious.

 

[Panajev2001a]

Here is what you're claiming as far as I can tell:

Texture streaming systems are perfect at knowing precisely which MIP level or tile to request from disk. No meaningful improvement in accuracy is possible.

This is the key to the argument. There is a huge difference between what is still a massive 10-20% improvements over PRT because you are implementing an accurate page fetching algorithm and an 100-200% improvement or more.

Microsoft intended to claim that SFS gives them a 2X-3X improvement in efficiency.

What is the basis of this? What is the basis to claim such an incredibly large efficiency improvement over what games relying on custom virtual texturing solutions with or without PRT?

We stopped arguing about this kind of Xbox One dGPU arguments a few pages back and the discussion was getting again interesting with investigation over patents and other published info... then boom! again with this secret sauce upgrade which closes the SSD gap (because this is the point it seems like you want to get across, not sure why if this gap were as insignificant as people used to say we need to invent such trump cards)

 

[Panajev2001a]

This Eurogamer article where Microsoft supposedly "clarified" the multiplier claim was published the same day as the "Series X Technical Glossary" that I was quoting.

A few sentences after your selection in that article is this:



It takes a lot of leaning over backwards to interpret this the way you seem to want to. Let me give you some options:

• Maybe "these aspects of Velocity Architecture" REALLY means "the new paradigm brought to us by SSDs in general"
• How about "a genuine game-changer" simply means "a small improvement over SSDs alone"
• Or maybe the easiest approach for you would be to say that "adding a multiplier" really just means a multiplier of 1.05.

XVA goes way beyond what is currently available on PC today, the custom I/O controller and BCPack + lzma decompressor just are not available on PC. You need more RAM and CPU resources to match what XSX can do. Nobody is disputing that.

The memory multiplier you quoted there referred to using the SSD as virtual memory as an extension of the RAM pool available to games (RAM becomes a cache of sort / a scratch pad / work ram while the SSD becomes what RAM used to be...). Isn’t that good? Are people that unhappy XSX is just better than PS5 in some categories but does not DESTROY it as you were expected?

 

[Panajev2001a]

MS saysy that they can move in that 2.4GB/s rate, what it would take another system (regardless of system) 7+ GBs

That could be implied, but using the same math and baselines you do know you need to admit PS5 going way above 9 GB/s or equivalent system bandwidth right?

The numbers for both are only 2.4 GB/s and 5.5 GB/s which is the speed each system can transfer data with from SSD over. I can see you factoring average BCPack and Kraken compression to calculate what the equivalent bandwidth on another system lacking that kind of HW compression (and not using software to decompress the data).

If we also all understand and accept that 100GB will be instantly accessible to the developers until an actual XSX developer with experience on the dev kits directly disputes it, I'm sure everybody will get along.

I get that, nobody wants to be unpleasant, but this is not an echo chamber or a hype group or cult . XVA allows fast access to the SSD and, based in their comments, about 100 GB of it can be accessed as if it was virtual memory transparently to developers (which is great).

It is a very powerful system regardless, no need to make up unsupported scenarios to make it seem even more powerful and expect people to just accept it as is for fear of (peaceful one would hope) “conflict”.

 

[OptimistPrime]

What is the basis of this? What is the basis to claim such an incredibly large efficiency improvement over what games relying on custom virtual texturing solutions with or without PRT?

I‘m not claiming the improvement is actual. I’m claiming that this what Microsoft said. You can tell me that it’s a settled matter all day but it doesn’t change the fact that Microsoft claims SFS improves bandwidth by 2X-3X.

 

[Panajev2001a]

I‘m not claiming the improvement is actual. I’m claiming that this what Microsoft said. You can tell me that it’s a settled matter all day but it doesn’t change the fact that Microsoft claims SFS improves bandwidth by 2X-3X.

What is the basis of your claim that MS claimed a 2-3x improvement over games implementing virtual texturing already (with or without use of PRT/tiled resources)?

The article gave a good number and provided no baseline for the relative speed up. Why would one have to assume the baseline is so much higher?

 

[Three]

The deniers are making it question of baselines or benchmarks against past/current/future systems.

The proponents aren't saying what you are saying. From the perspective of the average reader, the multiplier applies to the 2.4 GB/s Raw number. MS saysy that they can move in that 2.4GB/s rate, what it would take another system (regardless of system) 7+ GBs or 3 seconds @ 2.4GB.

That's a win for the XSX in the face of the massive advantage that PS5 has in terms of raw and compressed speeds.

Based on what quote and what evidence are the proponents basing this on? Because they have provided nothing.

This Eurogamer article where Microsoft supposedly "clarified" the multiplier claim was published the same day as the "Series X Technical Glossary" that I was quoting.

A few sentences after your selection in that article is this:



It takes a lot of leaning over backwards to interpret this the way you seem to want to. Let me give you some options:

• Maybe "these aspects of Velocity Architecture" REALLY means "the new paradigm brought to us by SSDs in general"
• How about "a genuine game-changer" simply means "a small improvement over SSDs alone"
• Or maybe the easiest approach for you would be to say that "adding a multiplier" really just means a multiplier of 1.05.

So much information I can't process. I've already given up. what is physical memory? /s
How would this contradict my interpretation. It doesn't. You're just replying with nothing substantial and arguing for arguments sake. Yes an SSD helps with having to load less into RAM just like the slides I showed you.

I'm NOT saying that the following is what XVA is.

I am saying that there are designs that will allow MS to do what they predict and not be restricted in the way you see to be predicting.

This may very well be the case but
1) what has this got to do with the storage speed on the right?
2) This isn't about SFS its some actual fast NVMe that MS have never mentioned ever in any form. Something you're just saying they may have based on nothing again.

PCIE4.0?

In a console? Because if you go into the actual paradigm thats the enabling factor. Put as many lanes as you like at 2.x GB each and you get this benefit no matter what you attach....
I'm not even being facetious.

Yes the storage speed spec.

The issue here is that people are saying there is something else (SFS and GPU hardware) that makes the SSD speed 2x. So what's this you're mentioning? How is it related?

People are just throwing everything against the wall and hoping something will stick at this point.

 

[OptimistPrime]

What is the basis of your claim that MS claimed a 2-3x improvement over games implementing virtual texturing already (with or without use of PRT/tiled resources)?

The article gave a good number and provided no baseline for the relative speed up.
Why would one have to assume the baseline is so much higher?

You know what, you're right. I just checked Microsoft's website and they made some edits to their Sampler Feedback Streaming description:

Sampler Feedback Streaming (SFS) – A component of the Xbox Velocity Architecture, SFS is a feature of the Xbox Series X hardware that just a new name for Partial Resident Textures which allows games to load into memory, with fine granularity, only the portions of textures that the GPU needs for a scene, as it needs it. This enables Partial Resident Textures have always enabled far better memory utilization for textures, which is important given that every 4K texture consumes 8MB of memory. Because it Partial Resident Textures avoids the wastage of loading into memory the portions of textures that are never needed, it Partial Resident Textures is an effective 2x or 3x (or higher) multiplier on both amount of physical memory and SSD performance. This is nothing new so we should probably stop talking about it.

 

[Thirty7ven]

That could be implied, but using the same math and baselines you do know you need to admit PS5 going way above 9 GB/s or equivalent system bandwidth right?

The numbers for both are only 2.4 GB/s and 5.5 GB/s which is the speed each system can transfer data with from SSD over. I can see you factoring average BCPack and Kraken compression to calculate what the equivalent bandwidth on another system lacking that kind of HW compression (and not using software to decompress the data).



I get that, nobody wants to be unpleasant, but this is not an echo chamber or a hype group or cult . XVA allows fast access to the SSD and, based in their comments, about 100 GB of it can be accessed as if it was virtual memory transparently to developers (which is great).

It is a very powerful system regardless, no need to make up unsupported scenarios to make it seem even more powerful and expect people to just accept it as is for fear of (peaceful one would hope) “conflict”.

I think it’s time to stop being bothered by those who only have an interest in this as means to further their agenda.

Devs with access have spoken, there’s no other way around it.

 

[Panajev2001a]

You know what, you're right. I just checked Microsoft's website and they made some edits to their Sampler Feedback Streaming description:

Ok, there should have been an easier way to say “no basis beside personal wishful thinking” I think. Still, funny reading .

 

[thicc_girls_are_teh_best]

Absolute bullshit. Overall that is. Really need a short memory to claim that.

Unfortunately it's not. I've been discussing these systems around the board since at least September, and I've seen how the tone has shifted as more data and evidence on both systems had came out.

I saw the shift in a lot of Sony fans becoming more aggressive towards perceived good news regarding XSX up to Road to PS5, a big part of that fueled by the impression Sony was being relatively quiet, coupled with some events like show cancellations adding into the mix. I think the Github leak is what caused the sharp turn among some of the more staunched PS fans because prior to that speculation seemed more even-handed.

And yes there were some Xbox fans who took to using Github and other related stuff in antagonistic ways to take digs at Sony and PS5, which caused some of those reactions, but then you had some Sony fans using vague insider quotes to try justify completely disowning Github and testing data from the active conversation, and if you brought them up in any way, you were a hater.

So now that Road to PS5 happened, a lot of PS fans who were obsessed with winning the "TF battle" have simply shifted that energy to the SSDs. Some Xbox fans have also taken to focusing on the SSD and I/O as the new prevailing narrative and, yes, some do it in a way to try downplaying Sony and PS5 in that area. But it's like I said way back, there are way more PS fans than Xbox fans both on these forums and in general, so common sense would say there are more PS fans who are the more staunched diehard types that fall into that label of fanboys or console warriors than Xbox ones. And I've seen that happen in various threads for over half a year, multiple times.

However, one thing I notice when it comes to PS fans discussing the SSD I/O is that they will, without fail, always refer back to a quote Cerny said at Road to PS5. And I get it; the man's a genius. I have mad respect for him and he knows his stuff. But when I see a lot of people continuously quoting him so as to downplay any concepts related to XSX's SSD I/O implementations, they do so out of the idea that Sony's approach is the ONLY approach that can work. That's where their problems start.

You have literally an entire history of electronics to see that Sony's approach is not the only one that is feasible, even if their approaches are very good. Hell, not all of their approaches have even been the best approach for a particular technological problem a sector of the market has faced, if we're being honest. MS's also had some duds, but in the context of PS5/XSX some of the people valiantly arguing on behalf of the former perceive Sony and Cerny as infallible, perfect, and the only way something can be done. None of those things are true.

There's also an undercurrent implied where, if anybody actually questions (with respectable critical critique) certain claims by Sony and/or Mark Cerny when it comes to the SSD I/O on their end, then they must not be credible or worth listening to because they aren't Mark Cerny, or they don't work for Cerny. By that notion, none of us should be listening to one another at all because 99% of us don't fall into those two categories! There's examples of that in this very thread!!

If you want other examples, just look at the instances where we were expected to assume features MS directly confirmed for XSX such as VRS, VRR etc. to also be present on PS5, even though Sony had not officially stated those features themselves (and still haven't). In good faith yes, we can assume those features are present on PS5, bt the expectation there in terms of taking the assumption if it's an overall benefit to PS5, is not generally provided to XSX.

An example of this is with assumed L3 cache sizes on the XSX GPU; logic would assume that MS might've increased this beyond the standard amount to account for more CUs to feed, but the very moment this was suggested I saw many posters shoot it down instantly. The take from them being that MS wouldn't make customizations to their GPU, even those that would overall benefit the system's design in crucial areas, and it was mainly a lot of PS diehards that shot down that speculation. You can probably guess as to why, considering when these were brought up and what so many people were focusing on before Road to PS5.

Really this is your expectation? Putting a lot of pressure on 343i there. Sounds like a very unrealistic expectation but maybe we will be surprised. They have shown the game before, but yeah.

Not expectation. Just something I figure they could do if they wanted, to keep momentum going and building. Especially considering they want to target Infinite for release later this year, by now they should have the game in a state where slicing a demo is relatively easy, since they'd just be playing a very small segment of a mission.

I don't see that as being unrealistic, As for the visual part, dunno. For sake of next-gen wouldn't you want in-game graphics to match the UE5 demo? Why not a planned launch title? It would only be a good signal for upward mobility in regards graphical capabilities getting deeper into the gen.

The Halo:Infinite footage at E3 208 will have been two years old by this point, that is enough time for further visual polish. I also strongly doubt that was running on final XSX hardware, even final devkit hardware. For all we know that was an in-game cinematic from the XBO or X version, in all honesty.

"Concrete information" seems to be a rare commodity right now but that will hopefully change within the next two months

Yeah. The wait is killing me. Not literally, of course, but figuratively.

Well, maybe not even that. It's a bit annoying as now it feels longer than it actually is.

 

[OptimistPrime]

So much information I can't process. I've already given up. what is physical memory? How would this contradict my interpretation. It doesn't. You're just replying with nothing substantial and arguing for arguments sake. Yes an SSD helps with having to load less into RAM just like the slides I showed you.

Seems like a convenient breaking point for you.

 

[JareBear: Remastered]

Unfortunately it's not. I've been discussing these systems around the board since at least September, and I've seen how the tone has shifted as more data and evidence on both systems had came out.

I saw the shift in a lot of Sony fans becoming more aggressive towards perceived good news regarding XSX up to Road to PS5, a big part of that fueled by the impression Sony was being relatively quiet, coupled with some events like show cancellations adding into the mix. I think the Github leak is what caused the sharp turn among some of the more staunched PS fans because prior to that speculation seemed more even-handed.

And yes there were some Xbox fans who took to using Github and other related stuff in antagonistic ways to take digs at Sony and PS5, which caused some of those reactions, but then you had some Sony fans using vague insider quotes to try justify completely disowning Github and testing data from the active conversation, and if you brought them up in any way, you were a hater.

So now that Road to PS5 happened, a lot of PS fans who were obsessed with winning the "TF battle" have simply shifted that energy to the SSDs. Some Xbox fans have also taken to focusing on the SSD and I/O as the new prevailing narrative and, yes, some do it in a way to try downplaying Sony and PS5 in that area. But it's like I said way back, there are way more PS fans than Xbox fans both on these forums and in general, so common sense would say there are more PS fans who are the more staunched diehard types that fall into that label of fanboys or console warriors than Xbox ones. And I've seen that happen in various threads for over half a year, multiple times.

However, one thing I notice when it comes to PS fans discussing the SSD I/O is that they will, without fail, always refer back to a quote Cerny said at Road to PS5. And I get it; the man's a genius. I have mad respect for him and he knows his stuff. But when I see a lot of people continuously quoting him so as to downplay any concepts related to XSX's SSD I/O implementations, they do so out of the idea that Sony's approach is the ONLY approach that can work. That's where their problems start.

You have literally an entire history of electronics to see that Sony's approach is not the only one that is feasible, even if their approaches are very good. Hell, not all of their approaches have even been the best approach for a particular technological problem a sector of the market has faced, if we're being honest. MS's also had some duds, but in the context of PS5/XSX some of the people valiantly arguing on behalf of the former perceive Sony and Cerny as infallible, perfect, and the only way something can be done. None of those things are true.

There's also an undercurrent implied where, if anybody actually questions (with respectable critical critique) certain claims by Sony and/or Mark Cerny when it comes to the SSD I/O on their end, then they must not be credible or worth listening to because they aren't Mark Cerny, or they don't work for Cerny. By that notion, none of us should be listening to one another at all because 99% of us don't fall into those two categories! There's examples of that in this very thread!!

If you want other examples, just look at the instances where we were expected to assume features MS directly confirmed for XSX such as VRS, VRR etc. to also be present on PS5, even though Sony had not officially stated those features themselves (and still haven't). In good faith yes, we can assume those features are present on PS5, bt the expectation there in terms of taking the assumption if it's an overall benefit to PS5, is not generally provided to XSX.

An example of this is with assumed L3 cache sizes on the XSX GPU; logic would assume that MS might've increased this beyond the standard amount to account for more CUs to feed, but the very moment this was suggested I saw many posters shoot it down instantly. The take from them being that MS wouldn't make customizations to their GPU, even those that would overall benefit the system's design in crucial areas, and it was mainly a lot of PS diehards that shot down that speculation. You can probably guess as to why, considering when these were brought up and what so many people were focusing on before Road to PS5.



Not expectation. Just something I figure they could do if they wanted, to keep momentum going and building. Especially considering they want to target Infinite for release later this year, by now they should have the game in a state where slicing a demo is relatively easy, since they'd just be playing a very small segment of a mission.

I don't see that as being unrealistic, As for the visual part, dunno. For sake of next-gen wouldn't you want in-game graphics to match the UE5 demo? Why not a planned launch title? It would only be a good signal for upward mobility in regards graphical capabilities getting deeper into the gen.

The Halo:Infinite footage at E3 208 will have been two years old by this point, that is enough time for further visual polish. I also strongly doubt that was running on final XSX hardware, even final devkit hardware. For all we know that was an in-game cinematic from the XBO or X version, in all honesty.



Yeah. The wait is killing me. Not literally, of course, but figuratively.

Well, maybe not even that. It's a bit annoying as now it feels longer than it actually is.

That Hot Chips conference presentation seems like our best bet to get the Cold Hard Facts, or whatever

 

[Exodia]

Easy. Originally, TF were the greatest thing ever and the SSD was only for cutting loading times. This was MS's original pitch. Then the PS5 came along with a blitz fast SSD that, and I'm paraphrashing, can change the way game worlds are built (the Xbox SSD can also do this). Then suddenly, since MS understood that their custom IO and SSD doesn't hold a candle to Sony's, they started pushing the narrative that every single new tech SW related would fix and decuplicate their system's output and capabilities.

When was the first time Sony officially mentioned this? That their SSD will change the way game worlds are made?

 

[Dory16]

That could be implied, but using the same math and baselines you do know you need to admit PS5 going way above 9 GB/s or equivalent system bandwidth right?

The numbers for both are only 2.4 GB/s and 5.5 GB/s which is the speed each system can transfer data with from SSD over. I can see you factoring average BCPack and Kraken compression to calculate what the equivalent bandwidth on another system lacking that kind of HW compression (and not using software to decompress the data).



I get that, nobody wants to be unpleasant, but this is not an echo chamber or a hype group or cult . XVA allows fast access to the SSD and, based in their comments, about 100 GB of it can be accessed as if it was virtual memory transparently to developers (which is great).

It is a very powerful system regardless, no need to make up unsupported scenarios to make it seem even more powerful and expect people to just accept it as is for fear of (peaceful one would hope) “conflict”.

Calling a manufacturer's claims made up would normally require proof of the opposite or at least to have some experience with the product (short of being consistent and equally doubtful of all manufacturers) but not in the world that you advocate for, we've noted.

 

[Ascend]

just look at the instances where we were expected to assume features MS directly confirmed for XSX such as VRS, VRR etc. to also be present on PS5, even though Sony had not officially stated those features themselves (and still haven't). In good faith yes, we can assume those features are present on PS5, bt the expectation there in terms of taking the assumption if it's an overall benefit to PS5, is not generally provided to XSX.

QFT.

The whole nature of the thread was discussing the possibilities to try and discern what is really going on with the XSX. We got quite far, but now we're running in circles again, because all the circles we've been through a gazillion times in the PS5/XSX speculation thread have been carried over to here. The thread is painted multiple times as some sort of cult that wants to hype up the XSX. And then the derailing starts with ridiculous claims like "we know all this already", "this has already been done", "there is nothing special here". It's basically the bullying of everyone with a different perspective to accept the point of view of the masses. And because they are the majority, it works and the mods are powerless to stop it, which is exactly why I will not be discussing anything else until we have official information. And then we wait and watch the next inevitable new wave of excuses for dismissal or bragging pop up. It's not as if official information is taken seriously now, so, why would it in the future?

I would say, let the games speak for themselves as the final judge, but we already know, even that will end up with the same rhetoric. Either we'll have excess bragging, or we'll have downplaying. It's quite obvious people are not interested in the actual possibilities of the technology, in learning and in sharing their own knowledge. They are only interested in being right themselves and making sure everyone else looks wrong, and they team up with everyone that agrees with them. And I'm done playing that game. If I end up being wrong regarding what I said in this thread, I'll gladly admit it. If I end up being right, I doubt the same courtesy will be provided in return.

Lastly, I wouldn't be surprised if this post gets flagged for derailing the thread, while all the others actually derailing are left untouched, and honestly, at this point I don't care.

 

[Panajev2001a]

Calling a manufacturer's claims made up would normally require proof of the opposite or at least to have some experience with the product (short of being consistent and equally doubtful of all manufacturers) but not in the world that you advocate for, we've noted.

I am not calling the manufacturer’s claim to be made up, but your interpretation of it. They claimed SFS allows 2-3 SSD I/O bandwidth and memory storage reduction improvements, but that is relative measurement. Relative to... what? What is the baseline?

MS left it at that, some people in this thread want to believe the baseline is very high and thus the 2-3x bandwidth improvement would shoot the 2.4 GB/s SSD into 8+ GB/s territory or so. MS is not claiming this... you are (well I assume you are one of the believers of such position).

Taking the peak decompression rate of the BCPack unit out of context as the number is similar to what you expect to have is numerology not much more at this point.

 

[Panajev2001a]

It's quite obvious people are not interested in the actual possibilities of the technology

No, it is not obvious. Just because claiming a sky high number / performance improvement scenario without much evidence is not taken as gospel does not mean not being interested in the tech. The two are not mutually exclusive And I am not sure what is the pro of painting it as such.

 

[Trueblakjedi]

That could be implied, but using the same math and baselines you do know you need to admit PS5 going way above 9 GB/s or equivalent system bandwidth right?

The numbers for both are only 2.4 GB/s and 5.5 GB/s which is the speed each system can transfer data with from SSD over. I can see you factoring average BCPack and Kraken compression to calculate what the equivalent bandwidth on another system lacking that kind of HW compression (and not using software to decompress the data).



I get that, nobody wants to be unpleasant, but this is not an echo chamber or a hype group or cult . XVA allows fast access to the SSD and, based in their comments, about 100 GB of it can be accessed as if it was virtual memory transparently to developers (which is great).

It is a very powerful system regardless, no need to make up unsupported scenarios to make it seem even more powerful and expect people to just accept it as is for fear of (peaceful one would hope) “conflict”.

I think that the initial perspective on how we analyze claims should not be from the perspective of "x claims a boost so apply that same boost to another system."

The boosts stated are only for the system being discussed. The other system its own boosts achieved differently so XSX boost tech doesn't apply to PS5.

Once we stop with that initial frame of reference we can evaluate xsx on its own stated merits.

Stanard says that the boosts are stacked. So you apply the benefits of SFS then apply bcpack and BC7 LZ compression for what your gains to throughput mean.

Simple clear and seemingly obvious.

2-3 gains means either a reduction of almost 75% in texture size or and equivalent throughput rate of 2.5 times the stated rate or 2.4 * 2.5 =6gb/s

Halving that again is the equivalent of 12...

The same exact way the PS5 can mathematically achieve its compressed throughput gains.

Everyone is making this harder than it really is.

 

[Panajev2001a]

I think that the initial perspective on how we analyze claims should not be from the perspective of "x claims a boost so apply that same boost to another system."

The boosts stated are only for the system being discussed. The other system its own boosts achieved differently so XSX boost tech doesn't apply to PS5.

Once we stop with that initial frame of reference we can evaluate xsx on its own stated merits.

Stanard says that the boosts are stacked. So you apply the benefits of SFS then apply bcpack and BC7 LZ compression for what your gains to throughput mean.

Simple clear and seemingly obvious.

2-3 gains means either a reduction of almost 75% in texture size or and equivalent throughput rate of 2.5 times the stated rate or 2.4 * 2.5 =6gb/s

Halving that again is the equivalent of 12...

The same exact way the PS5 can mathematically achieve its compressed throughput gains.

Everyone is making this harder than it really is.

I get it, the boosts are stacked... just that the boost due to SFS over tiled resources/PRT or standard virtual texturing (software managed) is not 2-3x and I look forward to proof it is. That is all, MS never claimed this or they would have been more specific in defining the baseline.

I understand what they are implying and what SFS can do, but thanks for taking the time to repeat yourself anyways (no sarcasm).

 

[Trueblakjedi]

Based on what quote and what evidence are the proponents basing this on? Because they have provided nothing.


So much information I can't process. I've already given up. what is physical memory? How would this contradict my interpretation. It doesn't. You're just replying with nothing substantial and arguing for arguments sake. Yes an SSD helps with having to load less into RAM just like the slides I showed you.


This may very well be the case but
1) what has this got to do with the storage speed on the right?
2) This isn't about SFS its some actual fast NVMe that MS have never mentioned ever in any form. Something you're just saying they may have based on nothing again.


Yes the storage speed spec.

The issue here is that people are saying there is something else (SFS and GPU hardware) that makes the SSD speed 2x. So what's this you're mentioning? How is it related?

People are just throwing everything against the wall and hoping something will stick at this point.

No. Everyone including MS says and "effective 2x or 3x (or higher) multiplier on both amount of physical memory and SSD performance."

Effective is key word here.

Now at 4 lanes we know that lane utilization is the key to the speed of the spec. You only need one lane to get to 2.4GB/s... what are they using the other 3 lanes for?

 

[Trueblakjedi]

I get it, the boosts are stacked... just that the boost due to SFS over tiled resources/PRT or standard virtual texturing (software managed) is not 2-3x and I look forward to proof it is.

This is crux of your argument. You believe Cernys claims without reservation. Yet you deny MS' direct claim on their publicly available page and directly from the architects of the solution itself (Goossen and Stanard).

There's nothing left for you to post my friend. You literally just don't believe it and thus its not true and everyone else is a fool for believing it.

Don't repeat yourself again. We heard you the first and hopefully final time on this topic.

 

[Panajev2001a]

This is crux of your argument. You believe Cernys claims without reservation. Yet you deny MS' direct claim on their publicly available page and directly from the architects of the solution itself (Goossen and Stanard).

This is the crux of your misunderstanding (unintentional or not). I am taking Cerny and Goossen/Stanard at their word just like you are doing to all three of them.

They are not claiming what you are trying to make them claim, so your appeal to authority does not hold in this case: I am taking them at their word and if they gave more context we could understand what their word means instead of building up fantasies. 2-3x is a relative improvement claim: if you think the baseline includes a game that already implements virtual texturing (purely in software/shaders or using PRT/tiled resources), what is the evidence in what those engineers have been saying?

I am also not calling people fool, so I would like words not being put in my mouth, thanks .

 

[Dory16]

I am not calling the manufacturer’s claim to be made up, but your interpretation of it. They claimed SFS allows 2-3 SSD I/O bandwidth and memory storage reduction improvements, but that is relative measurement. Relative to... what? What is the baseline?

MS left it at that, some people in this thread want to believe the baseline is very high and thus the 2-3x bandwidth improvement would shoot the 2.4 GB/s SSD into 8+ GB/s territory or so. MS is not claiming this... you are (well I assume you are one of the believers of such position).

Taking the peak decompression rate of the BCPack unit out of context as the number is similar to what you expect to have is numerology not much more at this point.

You are correct, you didn't see a baseline. That's because those slides were not a technical specification of the component. Just a description of what it can achieve.
Cerny did not provide a baseline for the clocks rates in the PS5. We know the max speeds but nobody knows how low they can fall. They're "variable" after all. Yet I don't see too many people scream smoke screen maneuver and winge for a baseline, at least not you.
All we have so far is the claims from the platform holders. I will never hold against hem to make claims. Just to make claims that could be disproved. I actually believe that this claim from MS (100GB instantely accessible) is too specific to be made up, as any indie developer with access to a dev kit would be in a position to disprove it the following day and affect the platform's reputation.
Personally I'm done watching the continuous grandstanding in speculation threads for the sake of ego and platform honor. There are more important things happening in the world.