Next Generation Hardware Speculation with a Technical Spin [pre E3 2019]

[PSman1700]

Sony gave you exactly what you wanted to hear, but none of the details that actually matter. They gave you the performance of a devkit that is actually a PC and compared it to a completed official devkit.

This isn't mental gymnastics we're performing here, it's called reservation. I think we're allowed to have that for now until the signal is stronger.

this wasn't a technical expose. It was a first look.

Two of the best posts since the 'reveal' imo. They didn't give any specs, not even close really, other then useless 'buzzwords'. Without specs it's really hard to determine how fast things will be. Sony gave 8k, 'RT support', 8 core zen2, and a fast storage solution. I think we knew about 8 core Zen2 and Navi lite. 8k is a HDMI2.1 spec, so it can output 8k. RT support, well, seeing Cryteks demo, the recent DF video on minecraft RT, RTX games running on non-RTX hardware... it's not impossible PS5 is as much RT as its 8k.

Reading some tech forums, people seem to go crazy about TF numbers, way above 14TF even. If navi is going to have RT and other new features, AMD might not have aimed for higher TF numbers like Nvidia with Turing. Turing didn't offer much of a TF boost but it's much faster in even non-RT games anyway thanks to architecture improvements. RT just speeds things up even more.
People also seem to expect Anaconda will be much more powerful, which might not be the case at all.

What if PS5 will be 8 to 10TF navi with some form of RT support, 16GB GDDR6 and 8 core 3.0Ghz zen2, will people be dissapointed?

 

[BRiT]

Some people will always be disappointed. You can never please everyone, even if it does everything and is given away free.

 

[function]

I don't see why this has to be true in any way. I think simply demanding games to be installed, and installed means "on the internal NVMe drive", would work just fine.

Maybe you're right, I just see it as being potentially a negative to consumers in that they can't re-use existing solutions they have come to value. Not that there isn't an upside to the pain, but if a user were to consider total cost of ownership then particularly in 2020 there are downsides.

What I'm predicting is explicitly just a NVMe flash drive, which can easily be user-replaceable.

Fair enough, and that'd fit with what Sony have done in the past (not MS though, unfortunately, but that's another story). Two NVMe ports would be great, as you say.

The total amount of written storage per dollar you get is more if you buy QLC than if you buy a small amount of SLC or MLC. You just also get more capacity, so it's less in terms of full drive writes. It's really hard to spend all the writes on a drive unless you stream a lot of video on it, or use it as a cache

I was under the impression that SLC has 100 times or more the write endurance of QLC, which would mean 64 GB would have several times more write endurance than an entire 1TB SLC drive. Even a small 16 GB SLC cache would have more, though I don't know the prices. Something like the Samsung 970 Evo has an SLC cache for it's TLC, but that might be more about write performance (less important for consoles) than write endurance. The case is less compelling for MLC, thinking about it.

The few megabytes of save data really *don't* pile up, not even if you save every second, and even the "daily suspend of 32GB" is just 0.000007% of the write budget of a 1TB 500-full-writes drive. Assuming even halfway decent write leveling, the only way it is actually possible to spend the full write budget of the drive is to use it as a cache of some sort, or stream video on it.

Could I ask about your figures for ""daily suspend of 32GB" is just 0.000007%"? I get different ones.

960GB x 500 / 32GB = 15,000. Which is the rather higher percentage of 0.0067% - or a factor of about 1000. Granted, this still gives you a huge about of suspends, and suspend file sizes will probably be a lot smaller than 32GB, but it's still a bit higher than my liking.

With suspends, installs, huge patches into the tens of GB, and now potentially 8K video capture to save with the share button (god damn millenials), I think we will be seeing far more total data written than ever before. It's pretty nuts how my friends with game pass and kids churn through downloads - can be several games a week with two or three people a day on the thing.

But just how much of a worst case should you plan for ...

 

[function]

But he showed the problem is about giving devs a guaranteed performance which is sufficiently stable to design a game on this guarantee for streaming, specifically how fast you can move through an open world game. Doing one of the speed run trials in spiderman must not be dependent on the user's chosen drive and it must not create bugs. So the game must limit the speed based on worst case performance of the underlying storage layer.

It cannot be optional, it cannot use external mechanical hdds. This is less about the loading time, and more about real time guarantees. It's why he thinks this is a game changer, it allows new types of game design.

Even fast travel, if you make an online multiplayer shooter big map where you can use teleports points, you can't have different loading times based on users drives speed. It needs to be on the predictable local fast storage for anyone playing the game to avoid competitive advantages.

With the 360 MS would test usb drives to see if they were fast enough for games install, to match a minimum performance level. Perhaps Sony could do the same thing - check a drive after install to make sure it was also up to supporting this "faster than anything on the PC" performance level. Would be a shitter if you'd just dropped £150 on a 2TB NVMe drive from somewhere and it wasn't, but I guess them's the breaks.

 

[DieH@rd]

If they go with a fully proprietary ultra-fast SSD [custom NAND/controller/driver], I expect that they will still offer user expandable storage. Only, this slow memory will be used as "cold storage", with system "retiring" the oldest played games from SSD to slow storage whenever more fast free space is needed. System can even do pre-backup of the oldest played game to make it ready for eventual deletion from SSD. Games will be bootable only when they are placed on ultra fast SSD.

Simple. Useful. Automatic.

 

[borntosoul]

We could get two versions possibly, the vanilla with a 1tb ssd and the extreme with a 2tb ssd + optical drive.

 

[Shifty Geezer]

Business and marketing strategy isn't the subject of this thread.

 

[borntosoul]

Business and marketing strategy isn't the subject of this thread.

Yes, but the previous posts do touch on cost of various options and it does play a large factor in our guessing game. Because basically that’s all we are doing atm.

You and l butt heads on this.

 

[function]

Dah-yum. Intel 660p QLC SSD. Nice and affordable and all, but the 1TB version has a total stated lifetime of 200 TB of writes.

https://www.tomshardware.co.uk/intel-ssd-660p-qlc-nvme,review-34464.html

You probably wouldn't want to be using that drive for caching externally stored games. Do it with a couple of RDR sized games a day and you're technically past the redline in 2.5 years, even if you never suspend, download another game, patch anything, save game videos....

External storage may have to go the way of the Dodo after all. But I think that works be a loss.

 

[MrFox]

Yes, but the previous posts do touch on cost of various options and it does play a large factor in our guessing game. Because basically that’s all we are doing atm.

You and l butt heads on this.

He meant my post. (moved)

My beautiful, well crafted, amazingly written, totally off topic post.

 

[Remij]

So what I'm wondering, and forgive me if this is a bit off topic, is how this super fast storage solutions of these new consoles, could potentially translate to and affect PC gaming?

Could we see a case where PC storage (even SSD) isn't up to the task of streaming in and handling next gen game assets, and thus see storage bandwidth become a bottleneck? Or would it be a case of games finally being coded more efficiently to take advantage of higher bandwidth afforded by hardware already available on PC? (SSDs and NVMe drives)

 

[RDGoodla]

If high speed SSD (read speed > 3GB/s) is still too expensive in 2020, SONY may release a cheapest base model like PS3 20GB, which may contain less than 1TB of SSD (around 750 GB). Since a single pool of super high speed SSD is the best way for insta-loading, for both developers and consumers.

 

[iroboto]

So what I'm wondering, and forgive me if this is a bit off topic, is how this super fast storage solutions of these new consoles, could potentially translate to and affect PC gaming?

Could we see a case where PC storage (even SSD) isn't up to the task of streaming in and handling next gen game assets, and thus see storage bandwidth become a bottleneck? Or would it be a case of games finally being coded more efficiently to take advantage of higher bandwidth afforded by hardware already available on PC? (SSDs and NVMe drives)

Well we see that the alpha kit is able to do it, and I'm not necessarily sure if it has the custom hardware built in. So the assumption I would make is that the hardware is significantly over powerful to use the additional processing power to do these functions. If there are functions being done.

But on the PC space, aren't there settings we have for draw distance (and thus streaming?). You could have like 64 GB of system memory to hold a lot of stuff.

 

[bgroovy]

I still feel very strongly that for what they are proposing to do, a single dumb PCIe 4.0 NVMe controller attached to 1-2TB of cheapest possible flash is not just faster and all around better than any tiered storage solution, it will also be cheaper. Which is why that is what they will do.

Yes, I've been telling people this for months. It's both more economic and more performant.

It doesn't matter what hardware was in that devkit to achieve that. What matters is the practical result of the near instant loading between scenes, which is a result of faster storage and very fast decompression to the RAM.
I don't know if that devkit has a 64 core EPYC or a dedicated Vega 7 to brute force the decompression from mass storage to RAM, or if it has a dedicated FPGA, or an engineering sample of the final SoC that has dedicated hardware for that. And that doesn't really matter that much at this point, IMO.

If you have a guaranteed 3GB/s+ from storage, you may not be compressing data quite as aggressively, either.

 

[temesgen]

He meant my post. (moved)

My beautiful, well crafted, amazingly written, totally off topic post.

It's the Moderator Uncertainty Principal - as the quality of the post increases the less on topic it becomes. You're messing with the first three laws of physics otherwise.

 

[vjPiedPiper]

HMM it's a very interesting set of compramises to look at related to the use of flash based storage in the next console gen.
Consider the following assumptions ( they might not be perfectly accurate, but i'm guessing there are not too far off either)

1. Next gen consoles will need at least 1TB of storage, and probably more.
2. 1TB of pure flash is likely to be too $$$
3. BUT, using a combined solution means flash AND spinning disc, which also increases costs.
4. High performance flash actually requires a decent amount of capacity to get great performance.

I know nothing about the actual costs involved but where is the line,
between 256Gb nvme flash + 2Tb spinning disc vs just using 2TB of flash?
What is the minimum amount of fast flash required? 64Gb probably wont be fast enough? 128Gb provides a pretty good cache for the ram, but it aint gonna hold many games...
Do 128 nvme drives give enough performance?

Obviously the game/platform benefits from an all flash design are huge, but can platform holder get away with only 512Gb storage? or can they really bear the cost of 1TB of flash?
64 or 128Gb, would be enough to enable the spiderman demo, especially with some minor changes to the code / data layout. But it's very different from an "All flash console"

man things are interesting!

 

[HBRU]

what about 64 gb (or 128 gb) of low power DDR used as mass storage... ?!? I mean the same RAM used in Smartphones.... memory powered from some sort of battery when the system is down or even not (just reloaded again as system powers up)... memory that loads the entire game and mirrors automatically the HD (for save & such)... that kind of solution would be much faster than a SSD... also we can have two or even 3 TB HD...

 

[HBRU]

there are probably tons of low performance lpddr chips around that are hard to be marketed into a phone and such due to low frequency they can work....

 

[jlippo]

I expect a some form of cache solution and a drive which can be changed, so even if the drive is changed to bigger/faster they can expect most important data to be loaded fast enough.
Caching can be done intelligently as combination of what they need for each game, use slow drive directly for things like streaming movies and audio which is spoken rarely. (cutscene etc.)

 

[Shifty Geezer]

I don't see any problem with using an external HDD and caching on the SSD, along with progressive loading so you don't need a full copy. It just means playing games from external storage would be a little slower to get going. If the consumers wants to replace their HDD with an external SSD and get that faster, that's their choice, but it's be completely wrong to design the system core storage around having to support arbitrary external drives. Internal storage needs to be the most cost effective solution to get their intended R/W performance.