Support NeoGAF

[RaySoft]

It's amazing how many people skip over this little tid bit

The XSX doesn't need the raw throughput due to efficiently streaming only what texture data it needs (which accounts for the largest part of a games memory footprint). It can stream in, use, then replace texture data mid frame.

While the overall throughput will still tip in Sonys favour, the gap is smaller than what the mere paper specs would have you believe.

So your’e implying that the PS5 loads lot`s of «junk» into memory just to get to what it really needs?

 

[DynamiteCop!]

So your’e implying that the PS5 loads lot`s of «junk» into memory just to get to what it really needs?

No, what's he's implying is that they're possibly approaching the same scenarios in two completely different ways. Sony's device has more raw throughput, but the question that circles back around is how wasteful are they on the firmware and software side when it comes to parsing and streaming relative to Microsoft's solution?

It's not just about throughput, it's about how intelligently that throughput is being managed.

 

[SatansReverence]

So your’e implying that the PS5 loads lot`s of «junk» into memory just to get to what it really needs?

How some of you manage to twist a post into some sort of slight against sony that you then take personally is fucking amazing.

 

[On Demand]

Love how this number keeps growing by the day.

And SFS isn't "just software", there are hardware optimisations required to make it work beyond just an SSD.

You’re being fooled by marketing. Again.

Whatever software and hardware implementation MS is doing the PS5’s SSD already is doing also and already takes into account. Which means by naturally having the faster drive and more custom I/O, it will also be better and faster at those things. There are features of the PS5 SSD that SX doesn’t even have. Among one of them being 6 priority levels.

All this velocity and samplers whatever is just buzzwords marketing to try and make up(they can’t) for having a slower drive.

According to the Velocity trailer it gives them 2.5x Multiplier of SSD performance.

so that puts the XSX SSD above ps5 speeds theoretically??

I'm lost now.

Do you really believe that? Heck no. Whether speed, texture data or whatever else. There’s nothing the SX SSD and I/O does better than the PS5 SSD.

Congratulations. You’re supposed to be lost and making that suggestion.

MS marketing 101.

 

[RaySoft]

Xbox can actually address the SSD directly, as if it was RAM, if I understood correctly. This is a very interesting feature.

That´s the 10GB velocity reservation. Again, don`t get cought up in the marketig buzz.. Even if you mount a solid storage as RAM, you won´t magically inherit it´s speed as well. They only call it «RAM» since the storage is «closer» than ever before and the fact that SSD drives are memory chips masking as a harddisk. (So to take it litterally the SSD is memory chips «mounted» as a disk that again is mounted as RAM? ;-)

 

[Allandor]

It's only applicable for texture data.

Well, which is 99% of the data.
You also forget the "on average". And yes it only applies to "classic" render techniques vs the new. And we don't know what the PS5 gpu can do in that regards. E.g. current gpu can use the texture streaming ms has because most of it is just software. But an important part of it is in the hardware of the new Xbox GPU e.g. which is needed to reduce pop in which would occur with the pure software based feature. This is no problem on desktop GPUs, because they can just preload the higher res texture parts into the main memory and get cached there until needed. This cache can be much much smaller on the new Xbox and only contain small chunks of the texture. I doubt that it i no longer needed.

BTW, you don't need that much bandwidth with when loading small chunks. You need low latency, as low as possible. Even a decompression block adds latency. So I would not be surprised if PS5 and Xbox decompression unit won't be used as much as many expect.

 

[SatansReverence]

You’re being fooled by marketing. Again.

Whatever software and hardware implementation MS is doing the PS5’s SSD already is doing also and already takes into account. Which means by naturally having the faster drive and more custom I/O, it will also be better and faster at those things. There are features of the PS5 SSD that SX doesn’t even have. Among one of them being 6 priority levels.

All this velocity and samplers whatever is just buzzwords marketing to try and make up(they can’t) for having a slower drive.

Yes, Microsoft is just buzz words and marketing while Sony is the gift of the gods flown down the thighs of a virgin presented to the peasants by the one true prophet Cerny.

Do you really believe that? Heck no. Whether speed, texture data or whatever else. There’s nothing the SX SSD and I/O does better than the PS5 SSD.

How quiet is that XSX dev kit you apparently have?

Well, which is 99% of the data.

Probs closer to 35% to 50% actually.

 

[RaySoft]

How some of you manage to twist a post into some sort of slight against sony that you then take personally is fucking amazing.

Let me rephrase then; «Do you think the XSX would load lots of junk into memory just to get to what it really needs?»
Either split your big texture files or just load offsets straight from the drive.
But ofc if you really have to have all your different textures in a single big file, Im sure sampler feedback will come in handy.

 

[Bruno_Arrais85]

is this the new power of the cloud or blast processing?

 

[Bladed Thesis]

So your’e implying that the PS5 loads lot`s of «junk» into memory just to get to what it really needs?

 

[DunDunDunpachi]

This is just a foolish statement, the console has been in development since 2016 and Spencer for literally years has been saying that they're not going to release a console just because a certain period of time has passed or that it's some expected tick on a box that a new console generation needs to kick off. Like the Xbox One X it needs to be because of a profound step forward in gaming, HDR and 4K was a profound step forward, SSD's and high frequency processors are a profound and now affordable step forward.

You're taking it too literally. I'm just making an observation about how SSD / transfer speeds have ended up being this generation's 3D processing GPU, or DVD capacity, or GDDR5, for better or worse. I'm not caught up in the daily console warring and next-gen speculation, but I was watching and discussing both reveals. Microsoft's message was all about raw teraflops, ray tracing, and Dolby audio, with a fast SSD as a footnote. Then Sony announces their system with lower teraflops and an SSD.

But now it seems like the companies are increasingly talking about transfer speeds and SSDs. A part of this is just the PR game, so it's not to be taken too seriously, but I see similarities between both situations, that's why I mentioned it. Microsoft can be self-assured in their teraflop superiority, so why would they need to answer for the SSD side of things unless it actually was going to be a bigger differentiator next gen?

I'm not claiming Microsoft "reacted" to Sony by cramming in an SSD at the last minute. SEGA could've scoffed at Sony's 3D hardware and rightly pointed to their own well-established reputation as both an arcade and a home console maker. But instead they panicked and reacted badly. Microsoft could've kept pushing teraflops while keeping the SSD as a footnote. After all, if PS5's "secret sauce" is just a footnote for the XsX, that's really saying something.

But feeding into Sony's SSD comparison will continue to be a losing battle for Microsoft, just like when Microsoft fed into the format wars with HD-DVD vs Blu Ray.

wasn’t it ms who went first with the reveal? It seems more like Sony up-clocked their gpu to chase Microsoft teraflops.

See above.

Like Sony did by pushing the GPU to crazy speed with variable clock because of the low CU count

See above.

 

[Proelite]

Xsx SSD peak is ~3.7 GB/s. 2.4GB/s is worst case.

 

[pawel86ck]

People were saying XSX is all about bruteforce, but VA achieve the same goal with less. According to Ronald SFS is effective 2.5x I/O and RAM multiplier, because I/O has to load 2.5x less data.

2.4 GB/s SSD + 2:1 compression (4.8 GB/s) + 2.5x SFS gain gives impressive 12GB/s and more if data happen to compress particulary well (above 2:1).

 

[iHaunter]

"We don't care about hardware, we care about games" (Paraphrasing).

-Only talks about hardware and barely any games.

 

[Lort]

People were saying XSX is all about bruteforce, but VA is achieve the same goal with less. According to Ronald SFS is effective 2.5x I/O and RAM multiplier, because I/O has to load 2.5x less data.

2.4 GB/s SSD + 2:1 compression (4.8 GB/s) + 2.5x SFS gain gives impressive 12GB/s and more if data happen to compress particulary well (above 2:1).

exactly and its specifically based on information from actual xbox games... so thats not a peak thats an expected effective SSD throughput of 12GB a sec.

also with that comes more free ram ...about 50% more... awesome stuff

 

[RaySoft]

No, what's he's implying is that they're possibly approaching the same scenarios in two completely different ways. Sony's device has more raw throughput, but the question that circles back around is how wasteful are they on the firmware and software side when it comes to parsing and streaming relative to Microsoft's solution?

It's not just about throughput, it's about how intelligently that throughput is being managed.

Yes, Sony went for a total re-imagination with removing legacy bottlenecks, while MS went for buildIng upon the already standarized PC way. (Patch what you already have)
It`s been well known for some time now that they went with two different solutions to a specified problem, no-one is refuting that.

Ofc it`s about throughput, that`s the base metric, you can`t get any faster than that no matter how «intelligent» your solution is.

 

[DynamiteCop!]

Yes, Sony went for a total re-imagination with removing legacy bottlenecks, while MS went for buildIng upon the already standarized PC way. (Patch what you already have)
It`s been well known for some time now that they went with two different solutions to a specified problem, no-one is refuting that.

Ofc it`s about throughput, that`s the base metric, you can`t get any faster than that no matter how «intelligent» your solution is.

The bold and underlined highlights your total misunderstanding of the situation.

How well can the device sustain its throughput? How good is the compression method? How well is the software handling and communication from the drive to the decompression block? How effective is the decompression? How quickly can it get that information to the GPU or RAM?

It's not just about the peak capability of the hardware, what you're saying is foolish.

 

[RaySoft]

People were saying XSX is all about bruteforce, but VA achieve the same goal with less. According to Ronald SFS is effective 2.5x I/O and RAM multiplier, because I/O has to load 2.5x less data.

2.4 GB/s SSD + 2:1 compression (4.8 GB/s) + 2.5x SFS gain gives impressive 12GB/s and more if data happen to compress particulary well (above 2:1).

No, this is you misinterperating (wich is probably what they want) the given information.
You can`t ever get more throughput than your base raw throughput even how many (TM)`s you slap on it.
That «sampler feedback» claim of 2,5x is probably because if the data is already read from SSD once and already resident in memory, you don`t need to read it back from the SSD again.

 

[RaySoft]

The bold and underlined highlights your total misunderstanding of the situation.

How well can the device sustain its throughput? How good is the compression method? How well is the software handling and communication from the drive to the decompression block? How effective is the decompression? How quickly can it get that information to the GPU or RAM?

It's not just about the peak capability of the hardware, what you're saying is foolish.

I think it´s pretty obvious wich manufacturer that got the «W« in all your points....

 

[DynamiteCop!]

I think it´s pretty obvious wich manufacturer that got the «W« in all your points....

And here I thought you could be objective...

 

[RaySoft]

The bold and underlined highlights your total misunderstanding of the situation.

How well can the device sustain its throughput? How good is the compression method? How well is the software handling and communication from the drive to the decompression block? How effective is the decompression? How quickly can it get that information to the GPU or RAM?

It's not just about the peak capability of the hardware, what you're saying is foolish.

Why is it foolish when I`ve already taken all your points into consideration before posting it?

 

[MasterCornholio]

People were saying XSX is all about bruteforce, but VA achieve the same goal with less.

Actually Sony achieved more with their I/O customizations than Microsoft did. Hence why the PS5s I/O performs better than the XSXs I/O.

Not saying the XSX doesn't have other advantages (GPU CUs for example) but the I/O is the PS5s.

How much better in practice is the real question.

 

[Ascend]

No, this is you misinterperating (wich is probably what they want) the given information.
You can`t ever get more throughput than your base raw throughput even how many (TM)`s you slap on it.
That «sampler feedback» claim of 2,5x is probably because if the data is already read from SSD once and already resident in memory, you don`t need to read it back from the SSD again.

Nah... Things are only loaded as they are needed.

So your’e implying that the PS5 loads lot`s of «junk» into memory just to get to what it really needs?

It's not accurate to call it junk. It's a safety net to have things in RAM that might be used, rather than things that will be used. That's the main difference... The latest SFS explanation gives a good explanation;

Game textures are optimized at differing levels of detail and resolution, called mipmaps, and can be used during rendering based on how close or far away an object is from the player. As an object moves closer to the player, the resolution of the texture must increase to provide the crisp detail and visuals that gamers expect. However, these larger mipmaps require a significant amount of memory compared to the lower resolution mips that can be used if the object is further away in the scene. Today, developers must load an entire mip level in memory even in cases where they may only sample a very small portion of the overall texture. Through specialized hardware added to the Xbox One X, we were able to analyze texture memory usage by the GPU and we discovered that the GPU often accesses less than 1/3 of the texture data required to be loaded in memory. A single scene often includes thousands of different textures resulting in a significant loss in effective memory and I/O bandwidth utilization due to inefficient usage. With this insight, we were able to create and add new capabilities to the Xbox Series X GPU which enables it to only load the sub portions of a mip level into memory, on demand, just in time for when the GPU requires the data. This innovation results in approximately 2.5x the effective I/O throughput and memory usage above and beyond the raw hardware capabilities on average. SFS provides an effective multiplier on available system memory and I/O bandwidth, resulting in significantly more memory and I/O throughput available to make your game richer and more immersive.

Xbox Velocity Architecture: A Closer Look at the Next-Gen Tech Driving Gaming Innovation Forward on Xbox Series X - Xbox Wire

When we set out to design the Xbox Series X, we aspired to build our most powerful console ever powered by next generation innovation and delivering consistent, sustained performance never before seen in a console with no compromises. To achieve this goal, we knew we needed to analyze each...

news.xbox.com

 

[Hawking Radiation]

Dial back the hostilities towards each other.

These consoles can't come out soon enough.

Its always the case when E eastwood333 is allowed to troll thread after thread that this happens.

 

[jimbojim]

That was the speed of the decompression block.

"Our second component is a high-speed hardware decompression block that can deliver over 6GB/s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads decompression work from the CPU and is matched to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new compression [system] called BCPack that is tailored to the GPU textures that typically comprise the vast majority of a game's package size."

Inside Xbox Series X: the full specs

This is it. After months of teaser trailers, blog posts and even the occasional leak, we can finally reveal firm, hard …

www.eurogamer.net

And that was Jason said too. 2.4x2.5 is theoretical 6.x GB/s. Like Goossen stated before.
No need for further spin. XSS SSD numbers are what they are

2.4/4.8/6 and that's all folks. It really pales in comparison to PS5 SSD. Move on

 

[Ascend]

And that was Jason said too. 2.4x2.5 is theoretical 6.x GB/s. Like Goossen stated before.
No need for further spin. XSS SSD numbers are what they are

That would be uncompressed i.e. raw data. SFS works independently of decompression, meaning, they stack.



That means it's effectively 12GB/s. Or in other words, if you would be doing things without compression and without SFS, you would need 12GB/s of raw throughput to achieve the same results.

 

[Allandor]

Probs closer to 35% to 50% actually.

No way.
What should be so big?
Sound ? no
Geometric data? No, it won't be small but still really small compared to texture data.
Most loaded data of games are textures. Even for baked lighting texture formats are used to store where light should be how intensive.

 

[Batiman]

Next Thursday.

What time is this thing gonna be on?

 

[Starcheif]

If Sony fans all believed the Xbox had a faster SSD we would all be talking about the games....so lets just talk about the games. Most gamers do not care about SSD etc.

 

[pawel86ck]

Nah... Things are only loaded as they are needed.



It's not accurate to call it junk. It's a safety net to have things in RAM that might be used, rather than things that will be used. That's the main difference... The latest SFS explanation gives a good explanation;

Game textures are optimized at differing levels of detail and resolution, called mipmaps, and can be used during rendering based on how close or far away an object is from the player. As an object moves closer to the player, the resolution of the texture must increase to provide the crisp detail and visuals that gamers expect. However, these larger mipmaps require a significant amount of memory compared to the lower resolution mips that can be used if the object is further away in the scene. Today, developers must load an entire mip level in memory even in cases where they may only sample a very small portion of the overall texture. Through specialized hardware added to the Xbox One X, we were able to analyze texture memory usage by the GPU and we discovered that the GPU often accesses less than 1/3 of the texture data required to be loaded in memory. A single scene often includes thousands of different textures resulting in a significant loss in effective memory and I/O bandwidth utilization due to inefficient usage. With this insight, we were able to create and add new capabilities to the Xbox Series X GPU which enables it to only load the sub portions of a mip level into memory, on demand, just in time for when the GPU requires the data. This innovation results in approximately 2.5x the effective I/O throughput and memory usage above and beyond the raw hardware capabilities on average. SFS provides an effective multiplier on available system memory and I/O bandwidth, resulting in significantly more memory and I/O throughput available to make your game richer and more immersive.

Xbox Velocity Architecture: A Closer Look at the Next-Gen Tech Driving Gaming Innovation Forward on Xbox Series X - Xbox Wire

When we set out to design the Xbox Series X, we aspired to build our most powerful console ever powered by next generation innovation and delivering consistent, sustained performance never before seen in a console with no compromises. To achieve this goal, we knew we needed to analyze each...

news.xbox.com

Man, Jason Ronald has explained everything, yet some guys are still doubting and laughing. I wonder if these guys are questioning Cerny the same way.

 

[RaySoft]

And here I thought you could be objective...

I am objective. Even though, I´m not advocating the XSX in this scenario, doesnt mean I can`t do it in others... (look up my post history if you will) Fact is fact even how much we try to bend it.

 

[too much eslim]

This thread is still chasing the wrong rabbit for the most part. While theoretical bandwidth maximums are interesting, latency is the king parameter. And so far XSX does not seem to have much dedicated silicon for that while the PS5 has its two I/O co-processors plus cache scrubbers with the sole purpose to decrease latency as in asking for textures from the SSD and have them available for the GPU.

Looking forward to the MS hardware disclosure in August but my assumption is that they have a fairly straight forward PC-like solution and limited dedicated hardware here. The net result will be a very large difference between the two platforms in terms of actual performance to deliver assets to the GPU.

They specifically talked about latency. MS has talked about latency in every area of console design. Including controllers.

 

[jimbojim]

That would be uncompressed i.e. raw data. SFS works independently of decompression, meaning, they stack.



That means it's effectively 12GB/s. Or in other words, if you would be doing things without compression and without SFS, you would need 12GB/s of raw throughput to achieve the same results.

System architect Goossen explicitly said what is max. theoretical number for XSX SSD. It's 6 GB/s decompressed and that's it.

 

[Ascend]

System architect Goossen explicitly said what is max. theoretical number for XSX SSD. It's 6 GB/s decompressed and that's it.

2.5x is the raw hardware speed and it doesn't compound with the decompression speeds

Why are you ignoring what was already said? Not only said, but replied directly to you... Quoting myself;

That was the speed of the decompression block.

"Our second component is a high-speed hardware decompression block that can deliver over 6GB/s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads decompression work from the CPU and is matched to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new compression [system] called BCPack that is tailored to the GPU textures that typically comprise the vast majority of a game's package size."

Inside Xbox Series X: the full specs

This is it. After months of teaser trailers, blog posts and even the occasional leak, we can finally reveal firm, hard …

www.eurogamer.net

Obviously the decompression block will only be decompressing what will be loaded. If you are loading a third of what you would normally load, obviously it is effectively a 3x bandwidth increase. Microsoft has chosen 2.5x, which is fine. There are always some losses and no numbers are truly constant. And we have James Stanard, which is a Graphics Optimization R&D and Engine Architect at Microsoft, confirming that they stack. That you wish to deny this is your problem. It is still the truth that decompression and SFS stack.

 

[Thugnificient]

Since we're talking about the architecture around the storage as a whole (i.e. "Velocity architecture",) I was referring to also the PS5 architecture around the storage as a whole. The decompression tech is probably effectively the same. But Sony includes other custom components as well to offload CPU/GPU and that's what I was referring to as Microsoft hasn't mentioned yet tackling the same bottlenecks that Sony has.

The decompression technology in the PS5 is better as Kraken is significantly superior to zlib, yeah.

I’ve never seen so much spin based on software.

SX ain’t making up for PS5’s 3x faster SSD.

It’s funny seeing XB fans make fun of PS5 fans about raw teraflops of the SX. That PS5 can’t match it with whatever it’s doing. Yet here y’all are going gaga over buzzwords and software implementations thinking it’s a match for what PS5 SSD is doing, and even more laughable, thinking it’s better.

PS5’s SSD is 4x faster.

 

[RaySoft]

That would be uncompressed i.e. raw data. SFS works independently of decompression, meaning, they stack.



That means it's effectively 12GB/s. Or in other words, if you would be doing things without compression and without SFS, you would need 12GB/s of raw throughput to achieve the same results.

Ehh.. NO
Read what he said.
He corrected the original tweets question regarding SFS.
What he said was stacking was regarding raw readspeeds and compression.
SFS (if reading from SSD) wont magically extend SXS bandwith limitations. Compression will, but not anything else.

 

[Ascend]

Ehh.. NO
Read what he said.
He corrected his tweet regarding SFS.
What he said was stacking was regards to raw readspeeds with compression.

Link it.

 

[Panajev2001a]

It's amazing how many people skip over this little tid bit

The XSX doesn't need the raw throughput due to efficiently streaming only what texture data it needs (which accounts for the largest part of a games memory footprint). It can stream in, use, then replace texture data mid frame.

While the overall throughput will still tip in Sonys favour, the gap is smaller than what the mere paper specs would have you believe.

That would be uncompressed i.e. raw data. SFS works independently of decompression, meaning, they stack.



That means it's effectively 12GB/s. Or in other words, if you would be doing things without compression and without SFS, you would need 12GB/s of raw throughput to achieve the same results.

Quoting SFS again comes without any baseline. It stacks and so does virtual texturing with or without the use of PRT (so you can do similar math with the other console... the bigger difference is having BCPack over Kraken + Oodle Texture where the latter has a bit better efficiency to match only if you add the BC7Prep step and spend some async compute shader cost at runtime to unpack on the GPU).

He is saying that one technique helps the texture to stream data in and out as needed and people are puzzled that compression stacks up? Of course it does .

 

[On Demand]

SSD Speed and I/O management is just that. You know what you're getting. Unlike teraflops and GPU CPU speed which is more theoretical. There's no way around the faster SSD of PS5. No amount of software can make up for raw speed, 12 channel lanes, 6 priority levels. There's other I/O features the PS5 has that SX does not.

Like this engine developers said, with SSD speeds you know that's what it is. It's more tangible. Teraflops mean nothing until your game is up and running




So supposedly a 3x faster more customized SSD is not that much better against a significantly slower one. And the slower one can do the same things and is equal to the 3x faster SSD because of software.

Yet a 18% gpu difference in the SX is this huge world changing spec that the PS5 gpu is a lot weaker compared to it and won't be able to keep up. All based off theoretical limits no less.

Yeah.

 

[jimbojim]

Why are you ignoring what was already said? Not only said, but replied directly to you... Quoting myself;



Obviously the decompression block will only be decompressing what will be loaded. If you are loading a third of what you would normally load, obviously it is effectively a 3x bandwidth increase. Microsoft has chosen 2.5x, which is fine. There are always some losses and no numbers are truly constant. And we have James Stanard, which is a Graphics Optimization R&D and Engine Architect at Microsoft, confirming that they stack. That you wish to deny this is your problem. It is still the truth that decompression and SFS stack.

I'm not ignoring myself. Just stating what SYSTEM ARCHITECT said. 6 is MAX. No more no less. Accept that XSX SSD is 2.4 raw/4.8 compressed. Nor SFS method ( which is an old tech ) nor BCPack would mitigate or close the gap in any way compared to PS5 SSD

 

[RaySoft]

Link it.

I dont have a link, I just read his answer...

 

[On Demand]

The decompression technology in the PS5 is better as Kraken is significantly superior to zlib but that's because zlib isn't needed for what the SX does.

PS5’s SSD is 4x faster.

Kraken 22GB right?

Yeah then 4x faster.

 

[Thugnificient]

Kraken 22GB right?

Yeah then 4x faster.

Oh you’re right. 22GB vs 2.4GB so almost 10x!!!

 

[RaySoft]

I'm not ignoring myself. Just stating what SYSTEM ARCHITECT said. 6 is MAX. No more no less. Accept that XSX SSD is 2.4 raw/4.8 compressed.

Yes, the XSX decompression block has a max throughput of 6(ish)GB/s
That`s the theoretical maximum bandwith the XSX can load data into memory. Nothing else (compression ratio, SFS Etc.) can change that.

 

[MasterCornholio]

Oh you’re right. 22GB vs 2.4GB so almost 10x!!!

Whoah hold on a second there.

That 22GB/s figure is the theoretical maximum for the SSD with compressed data. The average is actually 9GB/s with compressed data. Most of the time you will get that 9GB figure but very rarely will you have the 22GB/s.

As such it really isn't fair to compare theoretical maximums with each other. It's best to stick with the spec sheets because those are the most likely values that you will get.

 

[RaySoft]

Oh you’re right. 22GB vs 2.4GB so almost 10x!!!

22GB/s vs 6.5GB/s would be the right comparison in that regard. Theoretical max throughputs of both decompression engines (they wont reach those numbers though except maybe once in a blue moon)

 

[On Demand]

Oh you’re right. 22GB vs 2.4GB so almost 10x!!!

SX 2.4gb raw PS5 5.5gb raw
SX 4.8gb compressed PS5 9gb compressed
SX 6gb compression software PS5 22gb Kraken.

 

[Thugnificient]

Whoah hold on a second there.

That 22GB/s figure is the theoretical maximum for the SSD with compressed data. The average is actually 9GB/s with compressed data. Most of the time you will get that 9GB figure but very rarely will you have the 22GB/s.

As such it really isn't fair to compare theoretical maximums with each other. It's best to stick with the spec sheets because those are the most likely values that you will get.

22GB/s vs 6.5GB/s would be the right comparison in that regard. Theoretical max throughputs of both decompression engines (they wont reach those numbers though except maybe once in a blue moon)

SX 2.4gb raw PS5 5.5gb raw
SX 4.8gb compressed PS5 9gb compressed
SX 6gb compression software PS5 22gb Kraken.

Wooooosh

 

[iJudged]

Pretty amazing if I say so my self. Can't wait.

 

[MasterCornholio]

Wooooosh

Were you trolling?