-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
Next-gen PS5 and next Xbox speculation launch thread |OT5| - It's in RDNA
- Thread starter Mecha Meister
- Start date
- Status
Sure. They started looking for one 'storage architect' and one 'DRAM Memory Lead ' last year. Sony started working on their IO / SSD / cache / filesystem tech 4 years ago (that we know of, they probably started working on it before that, you don't write such exhaustive patent like that in one week). It would be naive to think MS solution to be at the same level as Sony's.
Y'all just now catching up on Anthony Hopkins? Come on, son. Dude's been steady twisting words since these threads started.
anexanhume a very funny update of draco geometry
github.com
Very interesting
GitHub - google/draco: Draco is a library for compressing and decompressing 3D geometric meshes and point clouds. It is intended to improve the storage and transmission of 3D graphics.
Draco is a library for compressing and decompressing 3D geometric meshes and point clouds. It is intended to improve the storage and transmission of 3D graphics. - google/draco
github.com
Very interesting
Talk about hyperbole. If anything, 900p was compared to last gen, not PS1. And you do realize, that the difference between 900p and 1080p is easier to see, than the difference between CB 4K and 4K? Not to mention, that most players had a 1080p display in 2013, while most didn't have 4K display in 2016.
What about last gen, when the 360 had the upper hand, and PS3 games running a little below 720p was a huge deal!?
Colbert's HDD vs SSD vs NVME Speed Comparison: Part 1
I think this is a good moment to repost my SSD testing I did some weeks ago ...
In the below picture you see 3 benchmarks for each type of storage you can use in a console except Intel Optane. The tests were performed on my own PC. While the sizes of the mediums differ it will give you still an idea what would be a possible realistic speed estimation.
I only talk about reading operations!
Testsystem:
Motherboard: MSI B450 Tomahawk | CPU: AMD R7 2700X @ stock | Memory: 32GB DDR4-3000Mhz | GPU: MSI RX Vega 56 Airboost OC watercooled by a Eiswolf GPX Pro 120 | Storage Drives: NVMe = Samsung 970 EVO 500GB (4x PCIe Gen3), SSD = SanDisk Ultra 2TB (SATA3), HDD = Seagate Barracuda 4TB 7200rpm (SATA3)
If someone is not familiar with this kind of benchmark I now do some explaining what the 4 tests are about
I am aware that a real world access pattern would be a mix of those tested access patterns but the differences in speed between PC SSDs is not as high if you leave the roam of best case scenarios. But there is also a high chance that a next gen console game optimized for SSD speeds would actually target nearer to the worst case than to the best case to be able to be on-premise almost 100%.
TL;DR
The jump between HDD and SSD is a huge generational leap. The differences in SSD speeds are not as high as many expect them to be.
In the below picture you see 3 benchmarks for each type of storage you can use in a console except Intel Optane. The tests were performed on my own PC. While the sizes of the mediums differ it will give you still an idea what would be a possible realistic speed estimation.
I only talk about reading operations!
Testsystem:
Motherboard: MSI B450 Tomahawk | CPU: AMD R7 2700X @ stock | Memory: 32GB DDR4-3000Mhz | GPU: MSI RX Vega 56 Airboost OC watercooled by a Eiswolf GPX Pro 120 | Storage Drives: NVMe = Samsung 970 EVO 500GB (4x PCIe Gen3), SSD = SanDisk Ultra 2TB (SATA3), HDD = Seagate Barracuda 4TB 7200rpm (SATA3)
If someone is not familiar with this kind of benchmark I now do some explaining what the 4 tests are about
- The first test is a sequential read of 1GB of data with 32 I/O queues by 1 thread (not a valid use case for streaming data to a game, but initial loading)
- The second test is a random reads of 4KB data objects from a 1GB data file with 8 I/O queues done by 8 threads (not likely in a console because of the 8 threads)
- The third test is random reads of 4KB data objects from a 1GB data file with 32 I/O queues done by 1 thread (a typical console use case)
- The fourth test is random reads of 4KB data objects from a 1GB data file with 1 I/O queue done by 1 thread (a worst case scenario)
- A single non-RAID HDD is 50 times worse than any decent SATA3 SSD in random access data read for everything other than sequential read which is not your typical in game streaming use case. So any SSD will be already a huge jump in streaming-to-game capabilities even not optimized! And that is just the worst case scenario. test pattern 3 looks even better where we see 100 times the performance. I repeat: 100 times the performance on the most common access pattern you will find on any system.
- test pattern 1 is showing a 6.4 times increase in speed in favor for the NVMe. The open question is how often you will see that pattern with your games other than initial loading or copying/moving data from drive a to drive b ...
- In test pattern 2 you see a difference by still 4.35 times but I ask if this a valid use case for consoles because you run 8 dedicated threads while you have to maintain your frame rate. Maybe someone actually have deeper knowledge there can shed some led on it....
- In test pattern 3 the NVMe advantage is reduced to a merely 37% on a pattern which is normally your bread and butter access pattern on a PC and maybe on a console too.
- In test pattern 4 which is the worst case scenario my NVMe is just 22.5% better than a SATA3 SSD.
I am aware that a real world access pattern would be a mix of those tested access patterns but the differences in speed between PC SSDs is not as high if you leave the roam of best case scenarios. But there is also a high chance that a next gen console game optimized for SSD speeds would actually target nearer to the worst case than to the best case to be able to be on-premise almost 100%.
TL;DR
The jump between HDD and SSD is a huge generational leap. The differences in SSD speeds are not as high as many expect them to be.
Last edited:
For fucks sake can you stop with the downplaying of ssd.
We get it, you don't see the benefit 99% of the rest of the world sees with better loading times, start up, performance, reliability etc etc.
Just don't expect anyone else to share your view. The potential of ssd in consoles is plain as day.
Luckily Every next gen console is very VERY likely going to have SSD
I think this is a good moment to repost my SSD testing I did some weeks ago ...
In the below picture you see 3 benchmarks for each type of storage (except Optane) that is in use today. The tests were performed on my own PC. While the sizes of the mediums differ it will give you still an idea what would be a possible realistic speed estimation.
I only talk about reading operations!
NVMe = Samsung 970 EVO 500GB (4x PCIe Gen3)
SSD = SanDisk Ultra 2TB (SATA3)
HDD = Seagate Barracuda 3TB 7200rpm (SATA3)
If someone is not familiar with this kind of benchmark I now do some explaining what the 4 tests are about
Analysis:
- The first test is a sequential read of 1GB of data with 32 I/O queues by 1 thread (not a valid use case for streaming data to a game, but initial loading)
- The second test is a random reads of 4KB data objects from a 1GB data file with 8 I/O queues done by 8 threads (not likely in a console because of the 8 threads)
- The third test is random reads of 4KB data objects from a 1GB data file with 32 I/O queues done by 1 thread (a typical console use case)
- The fourth test is random reads of 4KB data objects from a 1GB data file with 1 I/O queue done by 1 thread (a worst case scenario)
Conclusion:
- A single non-RAID HDD is 50 times worse than any decent SATA3 SSD in random access data read for everything other than sequential read which is not your typical in game streaming use case. So any SSD will be already a huge jump in streaming-to-game capabilities even not optimized! And that is just the worst case scenario. test pattern 3 looks even better where we see 100 times the performance. I repeat: 100 times the performance on the most common access pattern you will find on any system.
- test pattern 1 is showing a 6.4 times increase in speed in favor for the NVMe. The open question is how often you will see that pattern with your games other than initial loading or copying/moving data from drive a to drive b ...
- In test pattern 2 you see a difference by still 4.35 times but I ask if this a valid use case for consoles because you run 8 dedicated threads while you have to maintain your frame rate. Maybe someone actually have deeper knowledge there can shed some led on it....
- In test pattern 3 the NVMe advantage is reduced to a merely 37% on a pattern which is normally your bread and butter access pattern on a PC and maybe on a console too.
- In test pattern 4 which is the worst case scenario my NVMe is just 22.5% better than a SATA3 SSD.
I am aware that a real world access pattern would be a mix of those tested access patterns but the differences in speed between PC SSDs is not as high if you leave the roam of best case scenarios. But there is also a high chance that a next gen console game optimized for SSD speeds would actually target nearer to the worst case than to the best case to be able to be on-premise almost 100%.
TL;DR
The jump between HDD and SSD is a huge generational leap. The differences in SSD speeds are not as high as many expect them to be here (so far).
File are organized in big file in games. I am 100% sure. This is why Sony patent they reduce the allocation table because they can specialize it for reading game file...
EDIT: This is why GT PS3 are so slow to load because they choose a bad strategy for files. The files were too tiny... 100% sure too. ;)
I don't think that Ms is interested to put money in something bigger than an nvme but why would would be that naive?
Yes, I get your meaning. I think the confusion is that you seem to think he was speaking beyond image compression and speaking about the platform as a whole, whereas I thought that the limited scope Of his statement was clear.
anexanhume a very funny update of draco geometry
GitHub - google/draco: Draco is a library for compressing and decompressing 3D geometric meshes and point clouds. It is intended to improve the storage and transmission of 3D graphics.
Draco is a library for compressing and decompressing 3D geometric meshes and point clouds. It is intended to improve the storage and transmission of 3D graphics. - google/draco
Very interesting
LOL. I'm not getting my tinfoil hat out. Did it support FP16 before this?
Even with Big File and an optimized allocation table you will still need to read from certain sectors in a random access pattern. You won't be able to avoid that. The main message of the test was anyway that the gains is speeds will be insane regardless of a further optimized solution.
What I find hard is that, despite the myriad of benefits fast IO brings being explained to him over and over again he keeps ignoring and/or downplaying it. Always "I don't see the big deal with faster loading times" even after it has been explained how vital IO is for most things.
Why is he continually down on it even when it has been explained rationally and logically to him that it could provide huge benefits to almost all aspects of game design and especially open worlds?
That's mostly 2017. Here's 2018 & Q1 2019 https://ubistatic19-a.akamaihd.net/comsite_common/en-US/images/53ubisoft fy19 earnings english final_tcm99-349883_tcm99-196733-32.pdf
PS4: 36% PC 27% X1 20% with Q1 2019 PS4: 34% PC 36% X1 19%
I'd say that a big enough audience to keep in mind when making your games. Even 15% wouldn't be something you ignore.
Almost no AAA games are delayed nowadays. There's Rockstar & some Japanse multiplats. All these Japanese games wouldn't even make in on PC last gen though. With the exception of FF & MH these games wouldn't push graphics & hardware anyway.
Lack of polish is subjective. Most AAA games have very polished ports and have multiple PC specific settings. The exception again is the new Japanese games such as the ones from publishers like Tecmo Koei, which are basically 1:1 console ports.
I'm 70% a PC gamer and I must've missed these exclusives. From Ubisoft there's Anno and uh...what else?
The revenue % numbers above are all from multiplats because the last Anno game launched in Q2 2019, just after the end of their FY, and the one before that was in 2015 I think. So last FY was all multiplats. I mean Q1 2019 where PC was the biggest platform was pretty much all The Division 2.
What helps for revenue is that Ubisoft gets 100% of the money for any sale on their store and PC is mostly digital. Sales would probably look worse for PC than revenue breakdowns but I think in the end revenue is what counts.
Last edited:
Arent more than a few AAA games have the PC version essentially ports made via outsourcing than inhouse?
Not accounting Ubisoft, since they have thousands of devs.
Not accounting Ubisoft, since they have thousands of devs.
That's not what you said. A company doesn't bet in a precise tech feature because they know to have weaker specs of the other counterpart.
That's really fanboystic think a company works in this way.
Of course they do, Sony and ms talked about "the world's most powerful console" when they were, and ms talking about tv/cloud, sony talked about exclusives and game resume etc.
Companies are not going to focus on there inferior aspects in comparison to the competition.
Why would anyone be rattled? It's great for everyone as all the next gen systems will have a SSD.
Popped in for the first time in a few weeks, and damn. This thread is toxic. Thinly veiled bias and fanboyism galore.
Can you confirm or deny that Microsoft haven't been working on their own solution the last few years? No? I thought so. I'm sure Microsoft already have storage architects employed. They do make a lot of hardware, you know.
Sony has revealed just this aspect of ps5 because it's the only which hardly MS can overtake in short time. Reveal the other full specs of the ps5 could give to MS the time to beat their specs in any moment and say "you see? Most powerful hardware ever made". That's quite simple to catch the equation behind such move.
And again I'd like to know what give you such convinction that's impossible for ps5 to offer comparable specs to what will have anaconda.
That's very contorted extrapolate such conclusion just because Sony said to have worked for years to a faster streaming data access for ps5 hardware.
Last edited:
You honestly think that Microsoft only thought about SSD last year?
oh Colbert thanks this is what im trying to say from hours..here the big think is ..the jump from HDD to NVMe. (minimum).....if someone go further...okay .thats it but the big jump is already thereI think this is a good moment to repost my SSD testing I did some weeks ago ...
In the below picture you see 3 benchmarks for each type of storage you can use in a console except Intel Optane. The tests were performed on my own PC. While the sizes of the mediums differ it will give you still an idea what would be a possible realistic speed estimation.
I only talk about reading operations!
NVMe = Samsung 970 EVO 500GB (4x PCIe Gen3)
SSD = SanDisk Ultra 2TB (SATA3)
HDD = Seagate Barracuda 3TB 7200rpm (SATA3)
If someone is not familiar with this kind of benchmark I now do some explaining what the 4 tests are about
Analysis:
- The first test is a sequential read of 1GB of data with 32 I/O queues by 1 thread (not a valid use case for streaming data to a game, but initial loading)
- The second test is a random reads of 4KB data objects from a 1GB data file with 8 I/O queues done by 8 threads (not likely in a console because of the 8 threads)
- The third test is random reads of 4KB data objects from a 1GB data file with 32 I/O queues done by 1 thread (a typical console use case)
- The fourth test is random reads of 4KB data objects from a 1GB data file with 1 I/O queue done by 1 thread (a worst case scenario)
Conclusion:
- A single non-RAID HDD is 50 times worse than any decent SATA3 SSD in random access data read for everything other than sequential read which is not your typical in game streaming use case. So any SSD will be already a huge jump in streaming-to-game capabilities even not optimized! And that is just the worst case scenario. test pattern 3 looks even better where we see 100 times the performance. I repeat: 100 times the performance on the most common access pattern you will find on any system.
- test pattern 1 is showing a 6.4 times increase in speed in favor for the NVMe. The open question is how often you will see that pattern with your games other than initial loading or copying/moving data from drive a to drive b ...
- In test pattern 2 you see a difference by still 4.35 times but I ask if this a valid use case for consoles because you run 8 dedicated threads while you have to maintain your frame rate. Maybe someone actually have deeper knowledge there can shed some led on it....
- In test pattern 3 the NVMe advantage is reduced to a merely 37% on a pattern which is normally your bread and butter access pattern on a PC and maybe on a console too.
- In test pattern 4 which is the worst case scenario my NVMe is just 22.5% better than a SATA3 SSD.
I am aware that a real world access pattern would be a mix of those tested access patterns but the differences in speed between PC SSDs is not as high if you leave the roam of best case scenarios. But there is also a high chance that a next gen console game optimized for SSD speeds would actually target nearer to the worst case than to the best case to be able to be on-premise almost 100%.
TL;DR
The jump between HDD and SSD is a huge generational leap. The differences in SSD speeds are not as high as many expect them to be here (so far).
Phil Spencer was already on record a long time ago saying that emphasis on faster booting and helping developers find optimal loading or solutions were top priority for Scarlett.
The idea is that, if you can stream stuff to a limited pool of RAM much faster, you can have smaller areas - as in play space smaller - pre-loaded, but using the same amount of memory. That means more different assets per area, more textures/models/sounds/etc.
The Drake House in Uncharted 4's Epilogue is a good demonstrator for that - since it's a small play area that you move through quite slowly, the devs were allowed to cram it chock full of assets - that area is, honestly, one of the densest I've ever seen in a game and, the result is that it is much closer to realism than most single environments in games.
certainly companies do not develop technology based on how "weak" they are compared to the competition...but if you ask me that Cerny interview sounded exactly like the other he did about the FP16 (but unlike FP16 this time this customized ssd will be really useful. )
......this is what marketing team are there for. Focus everything on your strengths, minimize or completely hide your weaknesses
I think this is a good moment to repost my SSD testing I did some weeks ago ...
In the below picture you see 3 benchmarks for each type of storage you can use in a console except Intel Optane. The tests were performed on my own PC. While the sizes of the mediums differ it will give you still an idea what would be a possible realistic speed estimation.
I only talk about reading operations!
NVMe = Samsung 970 EVO 500GB (4x PCIe Gen3)
SSD = SanDisk Ultra 2TB (SATA3)
HDD = Seagate Barracuda 3TB 7200rpm (SATA3)
If someone is not familiar with this kind of benchmark I now do some explaining what the 4 tests are about
Analysis:
- The first test is a sequential read of 1GB of data with 32 I/O queues by 1 thread (not a valid use case for streaming data to a game, but initial loading)
- The second test is a random reads of 4KB data objects from a 1GB data file with 8 I/O queues done by 8 threads (not likely in a console because of the 8 threads)
- The third test is random reads of 4KB data objects from a 1GB data file with 32 I/O queues done by 1 thread (a typical console use case)
- The fourth test is random reads of 4KB data objects from a 1GB data file with 1 I/O queue done by 1 thread (a worst case scenario)
Conclusion:
- A single non-RAID HDD is 50 times worse than any decent SATA3 SSD in random access data read for everything other than sequential read which is not your typical in game streaming use case. So any SSD will be already a huge jump in streaming-to-game capabilities even not optimized! And that is just the worst case scenario. test pattern 3 looks even better where we see 100 times the performance. I repeat: 100 times the performance on the most common access pattern you will find on any system.
- test pattern 1 is showing a 6.4 times increase in speed in favor for the NVMe. The open question is how often you will see that pattern with your games other than initial loading or copying/moving data from drive a to drive b ...
- In test pattern 2 you see a difference by still 4.35 times but I ask if this a valid use case for consoles because you run 8 dedicated threads while you have to maintain your frame rate. Maybe someone actually have deeper knowledge there can shed some led on it....
- In test pattern 3 the NVMe advantage is reduced to a merely 37% on a pattern which is normally your bread and butter access pattern on a PC and maybe on a console too.
- In test pattern 4 which is the worst case scenario my NVMe is just 22.5% better than a SATA3 SSD.
I am aware that a real world access pattern would be a mix of those tested access patterns but the differences in speed between PC SSDs is not as high if you leave the roam of best case scenarios. But there is also a high chance that a next gen console game optimized for SSD speeds would actually target nearer to the worst case than to the best case to be able to be on-premise almost 100%.
TL;DR
The jump between HDD and SSD is a huge generational leap. The differences in SSD speeds are not as high as many expect them to be here (so far).
What gets me is that if console use really is similar to the third test pattern, that means we're not looking at a 35-50x improvement, but more like a 200x improvement.
Don't worry, it'll suddenly click for him the very second MS confirms an SSD for the Xbox.
So, not sure if you're aware of this, but MS has teams dedicated to creating new tech and also making hardware. The same team that created the Xbox One X has been using NVMe storage for years with the Surface lineup. This isn't a new kind of tech for them.
I would've thought those sorts of extremely dense areas would take quite a while to make - are there many techniques being worked in to speed that up?
A PCIE4 SSD with some Phison controller is something which can be added to a console design in very short time though.
I think what he meant, it wasn't on MS priority such stuff for their next generation. Doesn't means they won't offer something similar but hardly the hardware will be structured specifically for such feature. From what we know now.
Last edited:
I don't think will be just that. Probably the whole hardware is structured in a way which hardly can't be replied with acceptable cost and in short time with an available hardware. Or they could lied. I don't exclude anything but if it's true what they claimed, probably they talk about it because it's impossible to be beaten in the last minute to another company in such feature.
Last edited:
But there is literally zero basis for such a comment; I'm not sure why anyone is even responding to it as if there is some merit there.
All we know is that MS and Sony started hashing out designs for nextgen consoles some 4-5 years ago and what they've been working on it being revealed now. Any other musings on their reasoning for targeting whatever hardware they have is random made up guesswork.... or "tales from my ass".
Photogrammetry makes authoring this kind of single use, decor asset much faster than it used to be. There's a bunch of really high quality photogrammed asset libraries popping up, too, with everything from rocks, stones and sticks to furniture and furnishings.
Sony doesn't produce nor NAND neither flash controllers and there's only so much you can do with standard parts available from 3rd parties. Basically, the only huge differentiator here I can imagine would be if Xb4 came out without any SSD-type storage while PS5 would have one. Any other option would put them on the same level essentially and as I've said it's relatively easy - from technical POV - to add an SSD to your system.
Solid State Drives are not new tech. They have been around for years, and Microsoft has had use for them not only in the data center but also in Surface devices.
You would think that they know a thing or two about how they work. Do people even reason out some of this stuff before they hit the post button?
What about actually putting the area together? I'm probably overblowing the difficulty but figuring out how to put together something so dense in a way that works, with so many little assets that need to be numerous and messy in a realistic way sounds like such painstaking work.
There are patent which indicates such things. None (for now from what we know) which indicates MS has worked for it in a new hardware.
I'm not, but seeing the Insomniac's Spiderman analysis where they show all the problems they had with HDD and bandwidth... It's obvious that a high speed SSD could solve all those problems.
We don't even know how exactly it works. At least let's see how will work concretely this feature in the hardware architecture and then you can blame them to spin just pr bullshit.
You're saying MS has zero patents on the development of SSD?
There is always a reasoning behind it. It just depends if the reasoning is (almost) bias free or (almost entirely) biased by the platform of your preference. Okay, forgot about the 3rd option: being a troll.
Last edited:
If, and it's a big IF, Sony has some implementation that Microsoft doesn't that will offer a substantially better experience, is it not just a case of Sony opting for revolution while Microsoft opted for evolution. Brute forcing the areas that are obvious, GPU/CPU/RAM like with the Xbox One X, meanwhile the team at Sony chose to change something radically that many people just weren't thinking about..?
Remember that's all IF Sony has some kind of generationally progressive solution the other side doesn't.
I mean us lay people and even seemingly many professional insiders never expected this development. Even basic SSDs being in next gen was thought not likely by many just a few months ago. People were speculating fairly conservative evolutionary developments. And it's not like companies don't drop the ball and create fuck ups. PS3 with the Cell, Xbox One with the huge focus on entertainment and Kinect. Shit happens and maybe the Microsoft team just dropped the ball, or conversely the Sony team have done something pretty cool and unexpected.
I'm not trying to do the console warrior thing, I'm really not, but hearing people discuss these big companies as if they are all operating at absolute balance and harmony with one another and all have the exact same ideas, I mean that's just not reality, and this could be a pretty stark example of that in the "console race".
Will be fun finding out where everything lands that's for sure.
Edit: And just to add these sorts of behind the scenes dramas are very real in this and all industries. Things going on we might not hear about for years about the developments. If there's truth to Microsoft being behind the curve here it will be really interesting to see what they do, if they genuinely make a dramatic "last minute" change or whatever.
Remember that's all IF Sony has some kind of generationally progressive solution the other side doesn't.
I mean us lay people and even seemingly many professional insiders never expected this development. Even basic SSDs being in next gen was thought not likely by many just a few months ago. People were speculating fairly conservative evolutionary developments. And it's not like companies don't drop the ball and create fuck ups. PS3 with the Cell, Xbox One with the huge focus on entertainment and Kinect. Shit happens and maybe the Microsoft team just dropped the ball, or conversely the Sony team have done something pretty cool and unexpected.
I'm not trying to do the console warrior thing, I'm really not, but hearing people discuss these big companies as if they are all operating at absolute balance and harmony with one another and all have the exact same ideas, I mean that's just not reality, and this could be a pretty stark example of that in the "console race".
Will be fun finding out where everything lands that's for sure.
Edit: And just to add these sorts of behind the scenes dramas are very real in this and all industries. Things going on we might not hear about for years about the developments. If there's truth to Microsoft being behind the curve here it will be really interesting to see what they do, if they genuinely make a dramatic "last minute" change or whatever.
Putting the area together definitely takes some time and I wouldn't expect every single area of every game to be of this kind of density, either. I mean, it would be nice, but that isn't very realistic.
That said, there's been some work into procedurally generating environments. Guerrilla used this kind of tech to populate HZD's map with assets, for instance - instead of placing everything by hand, they painted a bunch of feature maps, the system figured out the interactions between such maps and placed different soil textures, rocks, trees, bushes, etc.
For a lot of the less important environments in a game you probably can get away with mostly procedurally generated assemblies - this is one area where AI-assisted tools can speed up development a lot.
- Status