Storage Matters: Why Xbox and Playstation SSDs Usher in a New Era of Gaming
anandtech.com
anandtech.com
No, it is not going to usher a new area of gaming, just make game have less 'loading' screens, and perhaps larger open worlds.... but it is no near the innovation and changes we experiences in the 90s and early 2000's....
Let me say in a more popular language: History has demonstrated that gamers don't give a f@ck on how fast is the SSD.... PS1 won over N64 even though it had huge loading times, compared to near instantaneous N64's loading screens....
I meant, personally, i think it is great, as PS4 loading time are bad, but just fast loading time it is not going to make me buy a game.
Anyway, just a sad PR piece trying to get people exited for something that looks a small (but good) evolutionary improvement over the current PS4 Pro/Xbox X.
Now if we had custom fast/ray-casting technology, than it would have been exiting. But it looks like if enabled, a game will either have to drop resolutions and cap frame-rates to barely playable 30fps.
But, hey your SSD will be fast.....
They only want the console for the games, man, nothing else.
Starting point on SSD: https://youtu.be/ph8LyNIT9sg?t=310 for about 20 minutes
Until all (or most) PCs are equiped with high performance NVMe SSDs, those kind of features won't be possible other than on consoles.
Also, the PS5's architecture is optimized end-to-end for faster loading times, it's more than just faster storage.
No one is going to design gameplay for a special hardware constraint unless the gameplay can degrade to lowest common denominator. Which of course, makes needing the special hardware optional.
There are few exceptions to this rule. Some platforms pay for exclusivity, effectively covering lost revenue from other platform streams. And Nintendo alone makes a profit on hardware, so they can produce platform exclusives to drive addtl revenue from hardware sales.
Special SSD pipelines, while PC gamers are still using 7200rpm HDDs, are about as appetizing to game devs as waggle controls or Kinect sensor games.
The new consoles include these SSDs not to make something possible now, but to remain relevant in ten years time when PCs may have caught up.
This is the game industry's equivalent of supporting IE 11.
I’m impressed with the tech but it seems like the end goal was to keep console manufacturing costs down. Now it’s being sold as a gameplay-enabling feature and reason to upgrade. The upgrade only looks impressive because the PS4 by now is so old.
Relying on cheap SSD storage instead of expensive RAM, and relieving CPU effort via the storage streaming chip is a cool trick. But that tech alone enables absolutely zero gameplay experiences.
It hasn’t been proven to us whether or not a typical gaming PC’s increased memory just overcomes the need for this tech. If I have a PC with 32GB of RAM and my GPU has 8GB of its own RAM I’m not convinced that a PS5 with 16GB of shared RAM will do anything that the PC setup can’t.
Desktop computers eclipsed the performance of current consoles gen consoles so long ago that I am still suspect: my prediction is that a decent mid-range gaming computer is completely capable of playing any PS5 game.
Finally the gamer gets low/no loading which is nice to have.
But also it becomes much easier for Game Developers.
I'm still looking forward to it, after all, it is an huge improvement to current gen, independently of how long it took and how old the ps4 is.
And i'm not 100% sure if this doesn't affect PC Gaming. After all Direct Storage will hopefully fix small SSD Issues you also have on PC right now.
Takes a little imagination, but when you look at the ingenuity of something like F-Zero on SNES (2.5d graphics), it’s pretty clear that your typical tech really can used in amazing ways.
Obviously its graphical detail is nowhere what was shown here, but the precedent exists, and the only thing that would be different is the geomotry and texture res.
We still are nowhere near photo realistic, cinema type game play (and yes I realize that such animation has to be scaled down a bit because it can disturb people if it's too real, but we're not even at that point yet).
Real time ray tracing was going to be nVidias whole push into a new generation of 3D graphics, but that turned out to be not all that impressive (and with huge performance costs).
Nothing special here, no need to hype
edit:
You know what really makes a difference?
You playing an VR open world game where every single artificial intelligence NPC does thing in every possible way leading to different gameplay and outcome. And you as a character can just grab anything you want and throw at any monster that you just CREATED in that game itself
Or you can control that monster that you created in your own way, time travelling to another open world game saying hello to your friend playing in his house back and forth
Together with some AI NPC friends you made in that game, you live in that dimensional space forever as you want even after your human body dies, your conscious stays in electronic form
Referencing https://en.wikipedia.org/wiki/Sword_Art_Online:_Alicization
It means you have a lot more RAM, e.g. thus, PS5 has 825GB of RAM. RAM in the TB, soon.
Would a couple of orders of magnitude make a difference? If software is designed with terabytes of RAM in mind?
To get specific about what this enables, I think we will see many more indie games with great looking graphics. The combination of high res asset scans, automatic resolution scaling, automatic texture compression, generally less tight performance budgets that don't need teams to do optimization work (next gen consoles), and a financial model around tools to take advantage of all of this (Unreal + Quixel as the leader here) should make this next generation of games pretty awesome.
But it's overused, giving each game world less weight. The technological imperative, preoccupied with whether they could, etc. In contrast, artificial constraints add depth to gameplay, e.g. limited movement speed/duration ("sprint").
All that said, keeping processors fed with data is a central problem of CS. The innovations here are not just the SSD itself, but elimination of bottlenecks in the architecture (e.g. direct placement in GPU RAM).
Typically, storage is 1000x slower than RAM. On PS5, it's 50x. That has got to be a revolution in algorithmic space-time tradeoffs... which has got to be reflected in gameplay, somehow, somewhen.
Conversely when Nintendo makes new hardware, it’s the same deal - they’re the only ones financing games that use the balance board or gesture controls or labo or whatever. They just put up a lot more money and have internal studios with more autonomy. SSDs are a Sony problem not a game design/engineering problem.
Having said that, the original Prey looks pretty rough these days. And I don't see how a single portal is "the same thing" at all.
In fact, I think the only thing that I can think of to compare to is the ending of Portal 2 where you shoot a portal to the moon, but even that was just a single portal to a fairly low detail environment.
Are you referring to Prey's usage of that portal? Or something else.
FYI, the new Prey game released a few years ago is amazing. It has basically nothing to do with the original because they apparently just wanted to use their rights to the name on a flagship title. But it's amazing if you're into System Shock/BioShock type games.
amazing time to be alive for computing hardware power and great to see the consoles letting us fully leverage it
Mod chips.
storage space for CD quality audio, storage space for full-motion video, on-board full motion video decoders, cheaper media costs, easier media to deal with on development side, and most importantly a game library that catered more towards 'adultish' fans.
I remember reading back when these consoles were current that the biggest draw for most consumers was the fact that full-motion-video was a first class citizen on the PS1, and it was a huge technical wow for people. FMV was difficult on the N64 due to size and codec constraints, so it was rarely done -- and when it was if was usually subpar compared to the PS1 version of the same game. Resident Evil franchise was notorious for this phenomenon.
There was also a supply issue early on in the n64's lifecycle, I seem to recall it being just shy of Tickle-Me-Elmo hard to find.
The ps1 was also easier to pirate, double edged sword though it be.
On the PS5, a Blu-Ray type drive is optional. Are you going to have 100 GB downloads from the network? That's going to take a while. Even if you have a gigabit network connection, it's going to take 15 minutes or so to download a large game. Few people have that much bandwidth. Red Dead Redemption 2 needs 150GB of disk space now, so games are already that big.
Games will have to have some starting level that keeps the user busy while the downloader pulls in content. But after that, performance should be impressive.
It's also going to be a driver of 1440 and 4k gaming, because there will be more games that look good enough to make high res gaming worth it.
I'm sure the cost of speeding up loading isn't worth it after you've written it with one set of expectations in mind.
I just hope an unintended side effect here is that load times for big studio PC games speed up too.
This will change as the new lowest common denominator will eventually become the new consoles with their faster NVMe drives
These new SSD's are more of an upgrade then we've had in any console over multiple generations. Xbox One, Xbox 360, PS4 were all pretty standard "mini-PC" consoles. PS3 had a different architecture that didn't really pay off until much later in the console's life-cycle and still didn't have a huge impact, hence the PS4 dropping Cell for the PS4.
Lets say you only have 8gb of memory. Most importantly: You know you'll have 8gb (not more, not less). You build your game around having 8gb; textures are a specific size and quality, levels are designed around this, etc.
With a hard drive (or disk), games don't just store what is "on-screen" in memory; they have to store everything which could reasonably be expected to appear on-screen within the next, lets say, 10 seconds. This is because disks are slow; if the player violently flips the camera around, they can't wait 30 frames to page in the textures for that wall behind them; it has to be stored, even if the player isn't using it, and may never use it.
But, with SSDs, this changes entirely: new textures can be paged in within just a couple frames. Now, that paging can happen more on-demand.
So, wait: This means that 8gb of memory, which previously contained N% textures just used for what is being rendered, and M% textures which weren't being used, can now contain even more textures for what's just on-screen. Those textures can now be bigger! Higher detail. Bigger open worlds. The like.
Second point: Series X and PS5 do have ray tracing (not just ray casting, which you seem to suggest they don't have?) The Ratchet & Clank video we saw during the PS5 press conference seemed to showcase this, as it had multiple scenes showcasing reflections that didn't seem possible with just screen-space reflections. These consoles can absolutely, 100%, do super-1080p rendering at 60fps with ray tracing. Maybe not 4k60, we'll see, but 1440p60 w/ ray tracing is clearly attainable, at launch.
Third point, and this is critical, possibly the most critical: These consoles are all built with the same underlying technology that powered the previous generation; they just turned the dial to 12. This has, in the history of Playstation at least, never happened before. PS2 was very different from PS1. PS3 was a clusterfuck of technology that took developers years to get their heads around. But PS4 to PS5? Same platform tech. Developers know these platforms; the ramp up will be startlingly quick, and they'll be able to re-use and improve many of the very powerful game engines already there on PS4, like Decima, to make use of the new tech instead of starting closer to scratch like they did with PS4, PS3, and PS2.
I really don't think you could be more wrong than your assertion that this generation isn't going to be big. I'll go one-further: The media is underhyping the impact these technological advancements are going to have, and this generation will be the biggest single-generation leap in technical performance since the Nintendo 64. Only one AAA cutting edge title that I'm aware of was capable of consistently hitting 4K60 on a previous generation console (Gears 5 on X1X), and really, most games looked more like Control or Fallen Order (barely hitting 1080p30, no impressive technical achievements on consoles). We're now entering an era where practically every game will be 4K60, with substantially increased graphical fidelity, with ray tracing, with 3D audio, and it wasn't enabled with some crazy technology like IBM Cell; its the same shit that everyone already knows.
It's clear there's no "big technological breakthrough". SSDs are great but have been around for a while already, we simply had to wait for the technology to mature enough.
What we are seeing is that we can't just keep slapping better CPUs and GPUs to the machines, because the improvements nowaday are not that impressive anymore, so at least for the PS5 they have switched and started focusing on the architecture, removing bottlenecks, removing major pain points, and using additional hardware to handle tasks like compression and freeing resources for other purposes.
Some people might say these improvements happen because they were lazy before and never focused properly on dev needs and bottlenecks and the specific needs of game development. I partially agree, but the SSDs really don't just improve disk speed, they allow a completely different set of architectural optimizations.
Putting an SSD in a PS4 is not interesting. You can go do it today. Hardcore gamers and streamers will do it, because it does reduce load times by maybe 20% or so. But, until the entire pipeline, from SSD to hardware decompress to system memory to platform libraries to everything, is aware of that SSD and built around it, we wouldn't get the 10x benefits. This generation is where that's happening.
When I think about the PS3 to PS4 generation jump, I didn't get that excited. Sure, games looked better, but they're still 30fps by-and-large, they're still using rendering tricks all over to speed things up instead of truly awesome new tech like ray tracing, we're still waiting minutes for many games to load. The PS4 to PS5 jump (and similar jump on X1, don't want to leave out those players) will be far more transformative, in ways that actually matter to players and developers alike. It sucks that some people are still so jaded as to think "I have an SSD in my computer, its nice but its not huge"; they're missing the bigger pictures, and that bigger pictures is what people like Mark Cerny and Phil Spencer spend their entire lives thinking about.
Either this is going to be using the system RAM as a cache for the SSD data, (and then it will have the same problem, plus the developers will have to make sure that scenes can stay completely in the system RAM they have available, so this will be the limitation, not the size of the SSD), or they are getting the GPU to pull data directly from the SSD, and SSDs are very slow compared to RAM.
Well then, its good that the PS5 doesn't have system memory; like the PS4 before it, it uses a unified pool of GDDR memory for everything.
> or they are getting the GPU to pull data directly from the SSD, and SSDs are very slow compared to RAM.
Yes, this is what they're doing.
SSDs are slow compared to memory. But, that's not the important metric. The important metric is, are they fast enough to meet a specific rendering target, given each game's specific asset quality. But, actually, even that's not accurate: Games will change their asset quality to match the speed of the SSD such that they're capable of doing this, not the other way around. That's the special sauce of consoles; developers optimize specifically for the hardware available, because everything has the same hardware.
On PC, game developers say "alright, we'll bin 8 different asset qualities the player can select from based on their PC, hope they pick right, ship it." On consoles, developers say "hmm, if the player turns around in this segment, the framerate drops from 30fps to 20fps; lets dig in there and pre-load some of those assets." You cannot, fundamentally, at any level, compare PC gaming to console gaming, because the optimizations go all the way back to the overworked engineers at the development companies making sure the experience is great for the hardware console gamers have available. On PC, developers say "gamers can theoretically see the whole world, but we'll layer in some adjustable fog in case they can't handle it." On consoles, developers say "we know the system cant handle it, so lets block some sightlines with a mountain range and a few buildings."
The PS5 SSD can load data at upward of 9GB/s. This means the SSD is capable of loading 150MB of data within one frame at 60fps. If you develop games, you'd know how much data that is (its a ton). Hell; you can drop ten whole frames, 60fps to 50fps, most gamers short of DigitalFoundry wouldn't even notice, and you can literally replace 10% of the entire GPU memory. Even at 60fps, the SSD is fast enough to become a pretty dang suitable hot-cache for GPU memory.
At the end of the day PS5 will still precache streaming geometry/textures like PS4 does now (COD https://www.youtube.com/watch?v=mvdTtl27TpM), like PS3 did (GTA4), like PS2 did (GTA3), like PS1 did (Crash Bandicoot). Sadly there is no magic breakthrough coming with next gen consoles, and this SSD thing was the best they could muster to build the hype :(
Which brings me back to that 150MB in 1 frame number; that's not saying that nothing is being predicatively cached; its saying that some N frames are being predicatively cached, but if the developers make an intentional or unintentional cache miss, recovery from that miss is far, far faster.
Imagine you're walking down a grimy hallway; the player opens a door, and walks out into a sun-drenched paradise. 98% of players will just walk down the hallway and open the door, so the developers have a problem: They should definitely be predicatively loading the assets for the paradise outside, but should they also load the assets for the hallway behind the player? Most players won't turn around. If they throw out some of those assets behind the player, that could make more room for the paradise that's coming. Could have more particles, more bloom, higher res textures, all good things. Today, the answer to that question is usually Yes; if those hallway assets get de-loaded, the frame drops would be too huge if a player decides to turn around. Or, more commonly, its Yes (But): Maybe we could build a bend in the hallway, or put a second door behind the player, so if they turn around the immediate assets we'd need to load would be less, and we'll have more time to load what's further behind them if we need to.
With an SSD, that answer may be No; if they can load at 8 GB/s, they know the maximum turn speed of the player, the math is more likely to work out that they can actually just throw away those assets, or never build that bend in the hallway. If they end up needing them, maybe the framerate will drop from 60fps to 45fps for a couple seconds; that's still faster than nearly every game from current gen.
Technical limitations like IO bandwidth and latency have guided the development of games for generations, all the way to the very artistic design we take for granted as being the "vision" of some creative leader on the development team. You're playing Halo; you drop down a ledge; you can't climb back up the ledge to go backward. It feels real, we don't question it, but really that ledge is only there because an engineer needed a checkpoint to make sure the player couldn't go backward, so they could de-load some assets. Simple as that. Its not driven by artistry (though, it could be); its driven by technology and its limitations (learning about these "checkpoint ledges" is a "I can't un-see that" moment for many people; checkpoints are everywhere in many boundary-pushing games. Even the hyper-recent FF7R uses them all over, often like "duck underneath this pipe" or "squeeze through these boxes"; they usually let you go back, but it takes several seconds in a confined area, which gives the game more time to load the next area. Same concept).
With an SSD, maybe a creative designer says "lets put a ledge there, we want to keep the player moving forward". But it'll be an artist making that call, not an engineer.
This is also why this generation is significant even for PC gamers! What we're talking about isn't at a technical level; its not faster load times, its not higher framerate, its fundamental changes to the game design itself which were necessary for games running on consoles. It didn't matter that your $3,000 PC had a 970 Pro NVME SSD; Assassins Creed Odyssey was designed to run on a PS4, and it doesn't. But, now that everyone has an SSD, even PC games will benefit because the game will be designed under the assumption that its alright to load multi-gigabytes per second of assets from the cold storage. This "SSD-by-default" will get built into game engines which power games across every platform. And sure, we'll see the biggest benefits on the PS5 and with its exclusives, as its SSD and hardware decompress pipeline is far-and-beyond more powerful than anything PC gamers or the Series X has, but everyone will benefit none-the-less.
Here is another example, 6 year old Far Cry 4 with custom implementation https://www.youtube.com/watch?v=SVPMhGteeuE
Because its not solved. Games have only gotten bigger; assets are bigger than ever. Having a hard drive in every console actually helped a ton, because the hard drive/memory bandwidth is larger than the disk/memory one, but it didn't help enough.
There's no way around the fundamental bandwidth at the hardware: If a game needs a gigabyte of assets for the next major scene transition, it will probably need to use tricks to make sure that scene transition takes longer than, say, 10 seconds (50-100MB/s is a typical read rate). Checkpoints; stall sections which slow the player down; limiting the top speed of the character or (especially) vehicles; slow opening doors (think Dark Souls/Bloodborne); vestibules between two doors with an adjustable length based on expected load time... They use technical tricks on the hard drive, copying commonly used assets all around the platter to reduce seek times. I've even heard it said by a developer that "we'd be fine in this transition if the game gets installed to the inner edge of the platter, but on the outer edge the load will take too long"; that's how close modern games run against hardware limitations.
The software in play hasn't changed. Duh. We're not talking about the software, we're talking about the SSD.
In the past you could choose between capacity or speed now you get both
Now they can do both.
I mean, it's nice, but it's not really revolutionary.
It's a welcome incremental improvement, at best.
But the PR machine worked really well.
Massive textures, massive look up tables, massive geo, etc. A lot of techniques can use pre-baked data in interesting ways.
Another benefit is reduce CPU involvement. Textures are fundamentally stored on disk, and GPU needs to access that data for rendering. Before we would have the CPU use OS's filesystem storage stack, read a file block, decode the texture in CPU memory, copy the texture to GPU memory using PCI-e, and ask the GPU to render with that data. This method allows the GPU directly read from a SSD, bypassing all of the steps in between. As SSD becomes faster and faster, OS's filesystem is becoming the bottleneck.
Ehh... unless you have more specific knowledge that’s still under NDA, the reasonable assumption is that the CPU still mmap()es the file into the GPU’s address space, and then the CPU pages in data from the SSD as the GPU generates page misses. Which is technically possible on PCs today, but isn’t done because you can’t assume fast SSDs and you definitely can’t assume shared memory. (actually I’m pretty sure discrete GPUs can generate page misses for mapped CPU memory, but I’m not certain which graphics APIs really let that happen)
2 MB page size would reduce this to "just" 11k page faults per second.
I mean I'm not complaining about TLB issues which large pages do better at for obvious reasons, but thrashing the TLB is likely to be a lot cheaper than reading too much from the disk.
edit: and there have been SPARC CPUs with memory controllers that can do 256GB pages
You seem to know quite a bit about this - just wondering - how does this work? The GPU can generate PCIe bus traffic to do RAM reads or writes, but how does it cause a page fault in the CPU? Is this some kind of IOMMU? Is there any place I can read more about this?
No offence but I just cannot understand this viewpoint. A game is about fun, about gameplay. Pretty graphics do little for that (IMO) once the novelty's worn off, and can easily start to get in the way. A game's tech is mostly decoration (I said mostly, sometimes it can help). The real breakthrough in games would be in their depth of interaction, and that would take some kind of comprehensive model of humans and human experience, and more. We're not getting that soon.
Vision is one of our most important senses and discounting it and saying well it shouldn't matter because gameplay is the most importing is shortsighted. Graphics are a huge part of gameplay.
Sometimes people running lower specs might even play better than someone running higher specs, if that means the higher spec has more grass textures that better hide character models or something like that. Those sort of improvements are improvements for the sake of vanity, and are generally panned by a game's core players.
For me, this next Gen basically just solved Load Times, SSD Load Priority, Raytracing/Light, Detailed Geometry with Light etc.
It feels very obvious, don't get me wrong, but it is still not available yet. It is new.
I personally like the independed IO chip very much. I have and had seen enough latency issues on current Operating Systems due to full busses due to super fast NVMs.
I was used to SSD Speed, then i added an NVM and holy shit that shit is so much faster.
Good time to be alive.
People have even resorted to using external 256GB SATA3 SSDs in USB3.0 enclosures, which are faster than the internal 5400 rpm HDD in an Xbox One S.
I do think we will still see a lot of people connecting USB3.0/USB3.1 external spinning drives to their new-generation consoles. 1TB of SSD space is enough for maybe 12 or 14 games, maximum. "Serious" console gaming people buy many more games than that...
Fallout 76 was something like a 45gb install and then a 50gb day 1 "patch"...
I'm pretty happy to not have to keep a large amount of game boxes and worry about scratching discs too.
EDIT: To your specific example, FO76 released on Nov 14, PS4 preload with base game + day1 patch started on Nov 11 according to online sources I've found.
I'm not completely sure but I also think that Just Cause 3 required a patch at day one to work on PS4 and IIRC it was considerably big against a weak 4 Mbit DSL...
The real question is: How many games will support that configuration?
A game written for a slower HDD can always be run on an SSD, either with no improvement or an improved experience. But a game specifically designed for an SSD may be unplayable or even unstable on a system with an HDD.
We aren't used to this yet, because they're so rare, as games are designed for the lowest common denominator (i.e. HDDs). But sooner or later an SSD will just be a minimum requirement, and if you lack it that open world game might lock up if you move too fast.
Heck some games might just do a speed test and refuse to even run.
With PS5 there should be a number of single player games where greatly-increased loading times won't have a negative impact on anyone else, unlike multiplayer games where everyone should load into a map at the same time. I have in mind the sequel to Horizon Zero Dawn, and the upcoming Assassin's Creed.
Archiving games you aren’t currently playing because a USB hard drive is most certainly the faster option for anyone without a gigabit internet connection? Sure. But there’s no chance in hell they will let you run games off one.
Given the drive speed it might be nice if games could auto-archive into a USB-attached device. The XBox requires custom SSD hardware for expansion (those weird non-USB drives?) and the PS5 will require swapping out the SSD entirely to expand usable space.
... but sequentially copying a compressed archive to a fast USB3.1 device might not be so bad for games not currently in use.
The hard drive in the Xbox One is a 2.5" 9.5mm Samsung Spinpoint M8 drive [2]. Seagate purchased Samsung's HDD business in 2011 [3].
[1] https://techreport.com/news/24439/seagate-to-stop-making-720...
[2] https://www.reddit.com/r/xboxone/comments/1r5gc6/xbox_one_us...
[3] https://www.seagate.com/gb/en/news/news-archive/seagate-comp...
Just a couple of days ago they had a 40GB update, which ballooned to 85GB for a good chunk of people due to an issue with the updater.
I doubt this very much. For one, we already know the storage expansion options for the Xbox series x… which is an external drive equally fast to the internal one.
Ps5 is similar deal except there aren’t any NVMe drives that are compare right now. Maybe there are and I’m unaware, but regardless you can still attach a fast NVMe drive. Ps5 will be approving these drives, to ensure they are fast enough for the system.
Basically no, nobody will be using any external spinning drives for their games. They will be NVMe drives.
The limiting factor for SSD endurance is volume of writes, which can be very carefully controlled in a walled-garden console OS+app design. Ordinary consumers are already very unlikely to run into write-endurance issues with something like a samsung nvme m2.2280 512GB-1TB SSD in a desktop PC.
As the article says that kind of IO becomes a real strain on valuable general purpose CPU time. So I wonder if we'll actually start to see regular PCs begin offloading again, reversing the generalize everything a bit with more support specialized silicon so that they can actually feed the beast. I could see that having real implications for a lot of future software and OS development too, particularly for OSS. If things go that way I hope some good open standards can get out ahead of it.
----
Reference:
1. https://youtu.be/4ehDRCE1Z38?t=480
2. https://linustechtips.com/main/topic/1205112-i%E2%80%99ve-di...
SSD is compressed (by "Kraken"), giving 8-9GB/s typical.
You could compress RAM too, but not worth the latency penalty.
Will we see something similar with these SSDs? Are devs actually prepared to take advantage of the special decompression hardware? Or will it be the stuff of dreams and tech demos for the first few years?
But more importantly, transparent decompression has been implemented many times on many systems; it's not a pipe dream. Developers don't have to do much of anything special to use the decompression hardware. They just have to use the platform's standard APIs for reading data from disk, and the OS or standard library takes care of routing it through the hardware offloads. From the application programmer's perspective, the decompression just shifts the performance characteristics of the storage a bit—higher latency, higher throughput, stronger preference for certain block sizes.
The bigger question is whether developers will be able to make good use of high bandwidth storage, regardless of how that storage is provided. Replacing everything in RAM on a ~2 second timescale is pretty wild. Lots of game engines have systems to coordinate loading assets on the fly, and the scalability limits of those existing solutions are going to be tested.
Hard to say, but another big point in Cernys talk was developer productivity and why they don't try to use "revolutionary" hardware like the Cell architecture anymore: Onboarding time for new teams is just too long. We don't know yet if it is true, but he talked about an onboarding time of 1 month for a team of developers new to PS5 (but with game dev experience). The last time a team could be productive this fast was PS1. If that's true I think we have a decent chance of devs using all the tech provided from the start. More so for PS5 only titles than for multi-platform titles.
I feel like gaming companies really don't emphasize load times unless they have to and their energy goes into pretty graphics or other things.
In theory couldn't they have ultra fast load times now if they just lowered the amount they needed to load to start? / Manage the amount of content they need to load to start?
I wonder if they'll just focus on pushing any speed benefits into....load times for even more assets, pretty graphics or whatever they choose.
In the meantime console announcement promises/ hype have a terrible history as far as accuracy goes...
Additionally the huge bandwidth means that the consoles can fill their entire ram in between 3ish and 6ish seconds, which should reduce max loading times by a lot.
Previously, a PC or a console might have 4-8 cores (or hyperthreads, or cell processors, or whatever). Typically, you'd want to be doing sequential I/O from just one of those processor cores, because more than likely be a mechanical disk. At most you could do asynchronous or threaded I/O so that the CPU core could also do other computation while waiting for textures to stream in (or whatever), but that's about it.
More than one concurrent I/O at a time (even from other programs!) would cause the disk head to seek, killing your performance. You'd get stuttering in the game or texture popping as the I/O fell behind. You could budget on maybe 20-30 MB/s if you're lucky, or less than 5 MB/s random I/O if unlucky.
It made no difference if you had a PC with an SSD, as zero games were written for it.
So now, with SSDs in consoles as standard you would automatically assume, that we can finally utilise all the CPU threads and throw 8 cores worth of I/O at the SSD. Maybe increase the queue depth a bit more, have some asynchronous and overlapped I/Os or something.
But that's crazy difficult. The models and textures being streamed in are actually used on the GPU, not the CPU. Each texture is typically mip-mapped, building up a "pyramid" of lower and lower resolution versions to prevent moire when viewed at a distance. The "currently used set" of each texture is a complicated angled 'slice' through this sort-of-3D texture 'volume'! This is used asynchronously by the GPU, and ideally you'd somehow want to track this, figuring out which parts are used by the GPU, which aren't, from which mip-map layer, etc.. and then predict ahead the next set of 32x32 tiles or whatever you'd want to load. In parallel. Across 8 CPU cores. While the GPU is doing its own thing.
Just no. That's not happening. The synchronisation alone would be a nightmare. It's crazy talk. No human can write code like this, and certainly not optimally and efficiently.
So the brilliant thing about the GPU on the PS5 is that the GPU can make direct I/O calls. No CPU involvement. Whatever it "needs", as and when it decides, it can simply directly fetch.
It's not going to do this with 8 cores, or 16 threads. No, it's going to do it with thousands of cores. The I/O queue depths that used to be "exactly 1" for literally 100% of all previous games are not going to be 8. They're going to be tens of thousands! This is massive. This isn't "a couple of random I/Os done a little bit better", it's four or five orders of magnitude.
Saved you some time
Well, yet. As history shows, many big game shops care very little whether their game is unplayable on current systems and push for a more expensive setup instead. On the other hand, this is partly the reason I can now do ML at home doing things that were far out of my range a few years ago.
I’d give up this ultra fast SSD in a heartbeat if it meant there wouldn’t be any pay to win or DLC to buy for a half completed but shipped anyway game.