Alternative distribution to optical disks : SSD, cards, and download*
- Thread starter Cheezdoodles
- Start date
You're human, right? Officially all current human models are full of crap, except when they vacate at which time things get really @#$@#. Welcome to the club
It would not be the first time any of us has remembered something wrong, thanks for clarifying. Especially after another poster said there were 16x BDR so that nicely clarifies the situation. Thanks! :smile:
Anyone know the additional cost of adding additional lasers to Blu Ray? I know the diodes are not cheap and I guess the tech to add, say, 4 lasers for a 4x thoroughput would not be cheap. And the demand may be low in that you are mainly looking at consoles and PCs to push the tech. I am not sure PCs really would push it much as digital purchasing (iTune, Netflix, Zune, Amazon, Steam, etc) is becoming more and more popular so that leaves the consoles to absorb the cost?
And here we are again .
we are continuing to find work arounds for optical media . The ideas here are , well install to the hardrive , add more lasers , put expensive nand in the cnosole to cache too.
Going to flash carts doesn't need all those bandaids.
Also i can't find anything about a 16x bluray drive in any bing or google search
It's the HP AR482AA 16x Bluray drive which you're looking for. The one that HP calls a "16x bluray drive" but which is in fact 6x. So you can stop looking.
With 54MB/s and up on a 400GB disk, we don't need to install on a hard drive. It can stream the data (many game are already managing this very well with only 8MB/s and high seek times). It will have a hard drive anyway since we are slowly transitioning toward DD, so the hard drive comes available for free.
If they want the device to be a universal media center, they have to read all formats, people's DVD and Bluray collections aren't going anywhere, and they better be backward compatible too or they lose a big bunch of users, including me (one of the major Vita issue right now). I can't see any of the three competitors going with a flash based or DD only solution for this generation. It would be suicide.
It's hilarious that "putting expensive flash as a cache" doesn't seem to help you see the contradiction of using expensive flash as a distribution media.
Last edited by a moderator:
In today's news: Seagate Reaches Terabit/Inch Density Milestone with Heat-Assisted Magnetic Recording (HARM). Achievement to eventually pave way for 60TB 3.5 inch drives; short term doubles capacity of hard drives.
The wii u isn't pushing the graphical boundrys . From all leaks its just an xbox 360 + or ps3 +
So there are no 16x bluray drives ?
where we getting 54MB/s ? We are going to go with a 12x bluray drive ? You know they are noise as all heck. I own an 8x drive and its very noisy . One of the things people complained about was the loud dvd drives of the xbox 360 and how quiet the ps3 was
Bluray players are $50 bucks now. I don't think it matters at this point.
Because the cost of the flash in my idea would be absorbed by the user and spread out over a long period of time. Not only that but the box will be smaller and use less power.
This will be amazing. To bad broadband hasn't been rolling out as fast as i'd hope.
And you don't see anything funny on the choice of physical media nintendo made? Surely for company not pushing the graphics flash(or dvd) is more attractive than 25GB blu-ray variant? Also nintendo is no stranger to flash so they surely have already some first hand experience.
And why even bother with anything but 4x blu-ray once you have the flash or hdd to take care of random reads. Faster blu-ray doesn't make sense if it is more noisy, less reliable, uses more power etc. etc. Brute force hardly is the best solution(and neither is the most expensive solution if console manufacturer wants to make a profit)
edit. But I guess the real argument here on eastmens side is because blu-ray is evil and because the wrong companies won hd movie wars. That would explain the twisted logic why flash carts is OK. but flash based SSD is not. Because then there is no need for the evil.
Nintendo hasn't pushed a format boundry in a very long time. While they do use Rom on their handhelds they have used disc since the gamecube. For nintendo BC compatiblity is a huge deal and unless they put a hardrive in they wont be moving to DL games.
As for your second comment. Using the hdd to take care of random reads requires either a long install process or saving alot of space for caching . The problem with that of course is when you change games or when installing multiple games at the same time your going to be faced with storage limitations .
With flash its even worse , your going to need expensive flash that can handle plenty of rewrite cycles
All this is a band aid that
1) Adds cost to the console
2) Increases the size of the console
3)adds noise to the console.
Ok. I'm done with this. I'll come and eat humble pie or come and gloat depending how next gen unfolds. My final speculation is the most graphics/boundaries pushing console is going to be optical+hdd where I wish the hdd would be SSD of reasonable size. I bet there will be option to add additional 2.5"/3.5" hdd for slow mass storage if the included hdd is flash based. There is no reason imho. to have stuff not needing good performance on fast and expensive storage.
As per optical performance it has been shown numerous times on this thread 4x blu-ray is plenty for megatexturing up to 1080p as long as there is additional faster storage media for caching and random reads. Yes, you need to stream more data for 1080p than 720p but the chunks of textures are larger needing less seeking. Also there is absolutely no need to install WHOLE game to the fast media, only the working set/randomly used data. If you haven't seen links to this I suggest re reading the thread instead of posting "I DO NOT BELIEVE". Either way, I'm not going to post the same link to how megatexturing works Fourth or Fifth time(I lost count).
I'm also betting next gen will see digital download only slim model, probably not on launch unless the console manufacturer makes deals with network operators on selected markets(possible, but unlikely).
I don't believe there will be flash cart based distribution because it's stupidly expensive as per anybody who has done the math per Gigabyte and assumes game sizes are going to be growing towards 20GB+, maybe even hitting 50GB storage space permitting. By the time flash hits that 1$/25GB blu-ray is at(including profit for manufacturer+shipping) world is DD only rendering yet again flash distribution null and void. Limiting yourself to 8GB Grand Theft Auto when competitor has 50GB hidef version out would be plain suicide assuming graphics and sound fidelity matter at all.
I do not believe as additional ram as cache insteadd of hdd because the additional ram would get emptied on every bootup making installations impossible. That said all the fast SSD IO Controllers do have reasonable amounts of ram as cache. Depending how fast the drive needs to be there is wiggle room here how ram as cache is defined. The complexity of additional memory is not trivial if it's anything but scratchpad memory + it's single use thingy (instead where hdd enables digital purchases making profit, people are used to buying cheap stuff on their phones/ipads, console manufacturer would be stupid not to tap on that revenue stream).
And this is my final guess and conclusion on this thread unless new arguments come up which have not already been mentioned. This thing is going in spirals where same things just come up again without any better arguments than on previous pages.
edit: Oh, and if those holocubes are feasible next gen I would take one over blu-ray in instant. It would be ridiculously awesome to have something from star trek in your living room.
As per optical performance it has been shown numerous times on this thread 4x blu-ray is plenty for megatexturing up to 1080p as long as there is additional faster storage media for caching and random reads. Yes, you need to stream more data for 1080p than 720p but the chunks of textures are larger needing less seeking. Also there is absolutely no need to install WHOLE game to the fast media, only the working set/randomly used data. If you haven't seen links to this I suggest re reading the thread instead of posting "I DO NOT BELIEVE". Either way, I'm not going to post the same link to how megatexturing works Fourth or Fifth time(I lost count).
I'm also betting next gen will see digital download only slim model, probably not on launch unless the console manufacturer makes deals with network operators on selected markets(possible, but unlikely).
I don't believe there will be flash cart based distribution because it's stupidly expensive as per anybody who has done the math per Gigabyte and assumes game sizes are going to be growing towards 20GB+, maybe even hitting 50GB storage space permitting. By the time flash hits that 1$/25GB blu-ray is at(including profit for manufacturer+shipping) world is DD only rendering yet again flash distribution null and void. Limiting yourself to 8GB Grand Theft Auto when competitor has 50GB hidef version out would be plain suicide assuming graphics and sound fidelity matter at all.
I do not believe as additional ram as cache insteadd of hdd because the additional ram would get emptied on every bootup making installations impossible. That said all the fast SSD IO Controllers do have reasonable amounts of ram as cache. Depending how fast the drive needs to be there is wiggle room here how ram as cache is defined. The complexity of additional memory is not trivial if it's anything but scratchpad memory + it's single use thingy (instead where hdd enables digital purchases making profit, people are used to buying cheap stuff on their phones/ipads, console manufacturer would be stupid not to tap on that revenue stream).
And this is my final guess and conclusion on this thread unless new arguments come up which have not already been mentioned. This thing is going in spirals where same things just come up again without any better arguments than on previous pages.
edit: Oh, and if those holocubes are feasible next gen I would take one over blu-ray in instant. It would be ridiculously awesome to have something from star trek in your living room.
Last edited by a moderator:
If we are going with your suggested size for games then storage would hardly be a problem on even a 250GB drive, an unlikely size for 2013/14 500GB is more likely. As for install time, waiting 15-30 mins for a install on one of those 3 games you buy pr year is a weird problem, just as weird as a several hour download should be ok but not an install.
Of course this is only if the install routine is improved compared to now.
It would be interesting to see how many 360 users that actually end up installing their games to the harddrive, i know i would.
manux said:
You only figured this out on page 59
Btw, I may have missed it, but I don't think it was suggested to go with a large amount of RAM and no HDD but instead that because both optical medias and HDDs have not scaled in transfer speed-to-RAMsize there becomes an issue: Your HDD may be twice as fast, but nominally faster with random reads, but your RAM is going to be 4-16 times larger (2-8GB). Your load times are going to be much, much larger -- even with a HDD. A SSD is one solution. But another is to "bite the loadtime bullet" once during the initial game load and then have enough memory, either through a lot of system memory or through a 2nd "A-RAM" style pool, to allow developers to have a pool of large, fast low latency memory to pull from.
It isn't hard to see why this would be advantageous, just compare a game that is designed for 2GB of memory. On a 2GB system at level transitions you are going to purge and then have to stream back in 2GB of content (or sacrifice some of your 2GB to already begin streaming); on an 8GB system you would face the same initial load screen but you could then begin caching all the currents level content you would be streaming AND the next level. So that not only helps performance as the content being streamed in for the current level is in a much faster memory pool but once that level is completed you have essentially eliminated the load time for the next level. Interestingly, while I believe both Sony and MS consoles will have a HDD of some sort, in a word without a HDD this would be the best solution also for the optical disk problem: get as much of the data you are, will be, and may be using into system memory where bandwidth and latency performance is great. The drop in memory prices over the last 10 years really puts the old dynamics of memory costs on their head.
You only figured this out on page 59
Btw, I may have missed it, but I don't think it was suggested to go with a large amount of RAM and no HDD but instead that because both optical medias and HDDs have not scaled in transfer speed-to-RAMsize there becomes an issue: Your HDD may be twice as fast, but nominally faster with random reads, but your RAM is going to be 4-16 times larger (2-8GB). Your load times are going to be much, much larger -- even with a HDD. A SSD is one solution. But another is to "bite the loadtime bullet" once during the initial game load and then have enough memory, either through a lot of system memory or through a 2nd "A-RAM" style pool, to allow developers to have a pool of large, fast low latency memory to pull from.
It isn't hard to see why this would be advantageous, just compare a game that is designed for 2GB of memory. On a 2GB system at level transitions you are going to purge and then have to stream back in 2GB of content (or sacrifice some of your 2GB to already begin streaming); on an 8GB system you would face the same initial load screen but you could then begin caching all the currents level content you would be streaming AND the next level. So that not only helps performance as the content being streamed in for the current level is in a much faster memory pool but once that level is completed you have essentially eliminated the load time for the next level. Interestingly, while I believe both Sony and MS consoles will have a HDD of some sort, in a word without a HDD this would be the best solution also for the optical disk problem: get as much of the data you are, will be, and may be using into system memory where bandwidth and latency performance is great. The drop in memory prices over the last 10 years really puts the old dynamics of memory costs on their head.
Yeah, I'm stupid troll food
I doubt binary sizes grow significantly faster than optical speed(+this gen hdd perf was horrible so any decent 2.5" hdd or flash would help there). Blu-ray can scale up to 6x from ps3 2x perf (perhaps only on outer rim but that is still fast for quite significant amount of data).
If biggest game sizes are around 14(mass effect3)-20GB(rage?) this gen this 3x perf increase would warrant 42GB-60GB game sizes with no perf degradation. And that's before assuming next gen would have decent hdd. What I hear ps3/xbox360 hdd gives only 20-30MB/s and some claim not even that. Even with mechanical drives it should be possible to hit around 80-100MB/s for 2.5" form factor without going overboard on cooling and spinning speed.
if there was 64GB flash with decent controller we could see 10x+ perf increase on mass storage speed compared to this gen and this is before counting in the benefit in seek time. After all the hdd would be for the random access data...
A moment ago you were talking about 500 GB games. If games become larger, optical becomes more important (HVD is an optical format). Flash is currently offering performance advantages only at the cost of capacity, or price. And to solve those issues you need workarounds, same as everything else.
I also see an implied near-future timeframe in this thread that shouldn't be ignored. Obviously optical will become obsolete some time. The thread openened with thinking about the next-gen boxes or a little beyond, with maybe SKU revisions. We don't need to worry about how BRD will handle games in 2156 - only how it'll handle games in the next boxes versus the alternatives.
And yes, it'll need a good caching system in HDD/SSD. But then HDD/SSD seems a forgone conclusion to me anyway to sell content, so I don't think it's adding negatively to the cost of the machine. And that's the only downside - system cost. Everything else is then a positive with the proper hardware investment. Whereas with cards, you have a lot more issues to worry about in terms of game pricing, manufacturing and distribution, etc. For all your complaints against optical, you aren't presenting a case that's convincing me or others that optical is unworkable or unwieldy next-gen. My personal choice would be SSD or HDD+flash and BRD for distribution. Yes, there'd be installs, but if my choice I'd also have the OS team write some awesome IO systems to minimise this and automate content caching and early-launch systems with background installs. Five minutes of waiting isn't the end of the world. Much as I'd like flash/ROM cards, for being cool and small and allowing a little box and a more consoley experience (although run straight from card will still have significant load times unlike an old-school cart-based console's instant-play experience), their performance advantage is nothing if you have flash in the system, and they cost more and are more difficult to work with (hence why Nintendo haven't used them for Wii U). You could also add a DD only SKU to your optical-based system, and by the end of next gen, optical could possibly be dropped for DD only. BRD seems the most sensible solution for a full-generational advance of conventional gaming to me. I see Nintendo as the evidence for that. Like I say, they know cards better than anyone, and the economies of distributing with them, and the costs of going optical or card, and with their less-demanding next-console where cards would be a better fit than for a full next0gen experience, they have picked optical.
Basically, for you to have any argument at all that could sway me, you'd have to explain Nintendo's possible reasonings. If you can't find a fair, convincing argument for them not picking cards, then the choice for MS and Sony's next boxes has to lie with optical.
Nintendo have to support disks because they require - as an absolutely, completely essential part of the core system - BC with the Wii. That alone rules out Nintendo doing anything different. They're tied to BC more than even Microsoft. As their system has to have an optical drive anyway - and deal with the cost, bulk and engineering issues that go with that - and as they probably don't have the clout to force the used game issue, they'd be insane to not ship their games on disks.
There's also the much smaller issue that Nintendo will be launching and massively mass producing games before 16 GB flash chips are widely available and cheap. And that Nintendo aren't as far along the DD road as Microsoft, meaning a greater proportion of their software will probably be sold in physical form over the lifetime of the system. But the real issue is that Nintendo need optical anyway.
Edit: I've been too busy looking at flash specs and DRM to see that BC for Nintendo would necessarily trump any other reason for taking up 150% of their itty bitty little system with an optical drive.
That's a fair argument, but far from conclusive. If the only purpose in optical in Wuu is BC, and flash is supposed to be better in every other way, why not chuck in a dirt cheap DVD drive instead of a more expensive BRD derivative and have a card port for Wuu games? Wuu hasn't even got an HDD for caching, and likely won't feature an SSD either, so performance data access from optical won't be there. What'll they have? Optical and local flash storage? They consider that more economical than supporting cards? Yep. So why not the same for MS and Sony who will likely be needing more storage than Wuu?
steampoweredgod
Banned
broadband is likely to increase, procedural generation is advancing(even single level advanced design is possible via systems such as angelina ai), and compression can also be used.
Current bluray capacity is pretty good especially if used in creative new ways, content can be infinite and high quality if procedural generation algorithms continues to advance rapidly.
Regards bluray, I've had store bought pristine unscratched commercial dvds fail on me. I prefer my content to be working decades from now, and the information preserved, that is more important than capacity, IMHO.
Millenniata technology or substitute similar tech should be used to allow the discs to remain functional long term. Or else given heavy drm, the discs will become nothing more than useless junk with a pretty label printed on it.
News article
Current bluray capacity is pretty good especially if used in creative new ways, content can be infinite and high quality if procedural generation algorithms continues to advance rapidly.
The ps2 featured an advanced lens that could substitute for multiple lasers(to handle different media), iirc. Though it was never used for multi-simultaneous-reading, iirc. One may hypothesize that combinations of such advance lenses may be able to use a single beam to read to multiple locations at once.
Regards bluray, I've had store bought pristine unscratched commercial dvds fail on me. I prefer my content to be working decades from now, and the information preserved, that is more important than capacity, IMHO.
Millenniata technology or substitute similar tech should be used to allow the discs to remain functional long term. Or else given heavy drm, the discs will become nothing more than useless junk with a pretty label printed on it.
News article
Last edited by a moderator:
Advanced procedural generation != useful and looking anything like artists want to see it. If you want your worlds to look remotely alive and detailed procedural generation won't cut it and/or you'll still need to provide the basic building blocks to make up the world.
There's a thread somewhere about procedural generation, it might be worthwhile tracking that one down.
Thousand-year drive media are irrelevant.
Multi-beam requires multiple beams.
There's no good argument for having a mass of lenses to perform multi-reading. Without some kind of beam splitter and an overpowered laser, I don't know if there's any real way to accompish even a bad job with just lenses.
Response times and reliability would be negatively impacted.
Thousand-year drive media are irrelevant.
Multi-beam requires multiple beams.
There's no good argument for having a mass of lenses to perform multi-reading. Without some kind of beam splitter and an overpowered laser, I don't know if there's any real way to accompish even a bad job with just lenses.
Response times and reliability would be negatively impacted.
Ninjaprime
Regular
Looking like the best and most likely option is just a big ol' traditional HDD. SSD will remain too expensive even into 2014, any sort of alternative faster media is also too expensive or too "5 years away." 2-4 TB drive, maybe 7200 RPM would actually be quite nice.
steampoweredgod
Banned
In some arenas procedural content(evolved), has matched or surpassed human achievement at times. Do not think that the limitations of the past hold for the future.
The thousand year does not matter, the fact that after a decade the disk you bougth might be useless and unreadable does matter, at least to me. At minimum we must guarantee that we're not left with a collection of coaster disks we've to trade back to the corporations for access to the content we paid and bought.
Regards mass of lenses, I don't recall the exact details of the ps2 implementation of clever lens design, but it reduced costs and was practical.
Though it may be true that it is useless for multi-area reading, I've not looked into it. One could imagine maybe fluorescent molecules or q-dots could be used that took in minimal amount of photons and delivered readable information back. Though it might or might not work, I pressume intensity of biological research light may be limited so as not to damage samples. Protein based coated layers have been achieved in labs, and suggested for high capacity future disk storage tech.
Observation of phenomena has been attained at extremely small scales.
Though costs may be prohibiting, don't know.
Last edited by a moderator:
