![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
Windows 9, 128-bit ?
- Thread starter JoseJones
- Start date
Like a mindless drone you throw the same link into my face. Thanks, sherlock, I noticed it the first time. It's equally pointless as your post.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
IA-64 is normally associated with Itanium CPUs
http://en.wikipedia.org/wiki/Itanium
IA-128 I doubt will ever exist. And if it did exist in the future you surely will not be running a windows desktop OS on it.
http://en.wikipedia.org/wiki/Itanium
IA-128 I doubt will ever exist. And if it did exist in the future you surely will not be running a windows desktop OS on it.
I'm not in charge of Windows 9, so I can't say with exact certainty...but my best guess is a very strong "no".
Well, considering the address length in Windows right now is only 48-bits and that already provides a 256 TB address space, I would say from a memory perspective, no, we haven't. It's also worth noting that managing things like paging for such a large address space (like 128-bit) would have very large overhead in management. So even if Windows moved to a 128-bit platform, I can't envision addressing moving past what is possible with 64-bit, since there's more overhead and no real benefit.
32-bit primitive data types are still the norm for most application developers, despite the fact that on most 64-bit datapaths have wide enough registers and ALUs to work with 64-bit data types directly. Most software is only using 64-bit primitives where they're needed (usually when you're working with values that can't be represented with only 32 bits), since they take up more space in memory/cache.
So from my perspective, no, we're nowhere near saturating the potential of 64-bit.
Last edited:
I wish people would stop skipping their junior high math classes and remember that we're dealing with a problem of exponentially here. The trend might work at 'close to zero' values, but it's not going to continue to apply, because the numbers are growing faster and faster and faster with every bit we add on, and because we're receiving diminishing returns on the value of having large numbers. Address spaces are still 48-bit because even if system memory doubles every single year from here out (and we know that's not going to happen because that's absolutely absurd), it's going to take 10 years for us to exceed the capabilities of a 48-bit address space and another 16 years after that to exceed the capabilities of a 64-bit address space. If we'd like to follow your lead, and apply historical trends unsafely to the future, we didn't move to 64-bit in the commodity computer market until we were already shipping some computers with 4GB of RAM in them. So by your own logic, we probably won't reach 128-bit until computers have 16,777,216 TB of RAM anyways.
You have to remember here that the jump from 64-bit to 128-bit is going to push the limits to 18,446,744,073,709,551,616 times their current amount, and yes, that's a real number. You also have to remember that we're reaching the point where such large numbers aren't useful. When computers were 16-bit, any integer number larger than 65,535 had to be operated on in twice as many steps, because you needed more than 16-bits to represent it. Lots of real world numbers exceed this point fairly easily (such as salaries, for example), so doing extra work for anything which could be more than 65,535 was a major performance limitation. There are definitely situations where you would want numbers to be able to be larger than 32-bit (4,294,967,296), so 64-bit data types come up in computers somewhat regularly too, so it was nice to be able to work with them in a single step within a single register. Numbers larger than 64-bit are not very common, so for consumers a 128-bit processor isn't really relevant at all. Maybe there are some physicists out there that are really giddy about being able to work with super high precision numbers in a single register, but Steve, who watches Netflix on his iPad isn't going to get much out of that.
In actuality, a 128-bit data path just lets you do things like stick 4 32-bit numbers together and do the same operation on all of them at once. Single-Instruction-Multiple-Data is cool and all, but we already have things like GPUs for that.
No, I think larger than 4 GB address spaces had something to do with that too. Certainly not the only reason, but definitely a significant factor. In 2 years, they're most likely going to want more than 4 GB of ram on their devices, so it's a good idea to make the switch now. Desktops aren't going to have more than 16,777,216 TB of RAM anywhere in the short term future.
bigdogchris
Fully [H]
- Joined
- Feb 19, 2008
- Messages
- 19,091
Wow, what a long winded and asinine rant.
I think 128 bit CPU's will come well before the memory limitation is hit, like I pointed out with some companies updating to 64 bit CPUs.
I think 128 bit CPU's will come well before the memory limitation is hit, like I pointed out with some companies updating to 64 bit CPUs.
They did that largely for marketing reasons. Anyone remember that Qualcomm guy that got fired for calling it out?
http://www.cultofmac.com/251349/qualcomm-exec-soft-fired-after-calling-64-bit-a7-chip-a-gimmick/
Now everyone is doing 64-bit chips only because, you guessed it, Apple is doing it.
You're allowed to think anything...but since there's no benefit tangible benefit to be had, feel free to remember that there's no good reason to think that.
To shift the discussion just a bit, why would Microsoft want to start doing development work on x128? I don't really know, but I'll offer two possible, but very speculative answers:
1. Their own experience in going from 32-bit to 64-bit was so painful that they are "future proofing" Windows as much as possible. Of course, they have to work with a simulator, and who knows if the "128 bit instructions" are the right ones that will actually be implemented? Of course, that's just a "simple matter of programming" the compilers.
2. Apple demonstrated that by developing (secretly of course) a version of MacOS that could run on x86 or x64, they had a very, very smooth transition from PowerPC to Intel architecture CPUs. At the time, I was surprised by just how quickly that transition was completed.
SuperSubZero
2[H]4U
- Joined
- Nov 21, 2000
- Messages
- 3,780
There's another guy who posts very obvious troll posts with the intent of causing flame wars.
Maybe that account got banned or something. (please tell me it was)
Well the link proves I'm not making it up but, why do you need to be such a fucking dick about it? If you don't like the thread feel free to ignore it as nobody is twisting your arm forcing you to post here.
Microsoft Working on 128-bit Windows
Last edited:
SuperSubZero
2[H]4U
- Joined
- Nov 21, 2000
- Messages
- 3,780
Because that link directly references IA-128, an ITANIUM platform, which at this point is unlikely to ever exist, and also, you never followed up.
http://www.techhive.com/article/193426/article.html
So no, there won't be a 128-bit Windows or consumer platform anytime soon.
Bolding again: ITANIUM
Instead of being jerks to someone who is obviously not as knowledgeable on the subject, why not instead try to educate the person and not be a jerk while doing it? Being mean or hateful to someone asking a question may cause them to refrain from asking questions later in life and limiting their potential to learn something or worse, put someone else at risk.
TDLR: Let's try and be more mature.
TDLR: Let's try and be more mature.
defaultluser
[H]F Junkie
- Joined
- Jan 14, 2006
- Messages
- 14,398
You won't see it in your lifetime because it's unfathomable that consumer desktops will need that much ram, combined with the fact that Moore's law is rapidly approaching a brick wall. So you have two limits burning the candle at both ends: the need is disappearing, and the technology to continue filling that need is also disappearing.
Web is not driving memory needs anymore, video will hit the wall at 4k consumer / 8k industry (I'll be amazed if 4k ITSELF gets wide acceptance in consumerland). All the other aspects of the industry that are video-driven (gaming, 3D modeling, simulation, etc) will also eventually come to a halt.
Games cannot drive innovation forever because we are hitting a finite limit on what we can do with a game engine and still make a profit. Once your effort to make the fabulous 3D effects surpasses what people will pay, you can't go any further. And right now, games are as monetized as they can ever be without pissing off the customers to the point of abandonment:
1. Day-one DLC
2. Buggy releases to get free testing from userbase. This also technically means they're further in the hole financially because they can never recover from the money/schedule slip.
3. Pay to win, even in games that are not MMOs.
4. In-game ads a plenty!
Every month we hear about some game studio closing shop. Some are just plain mismanaged, but for a large number they just hit the wall. Won't be long until only the heavyweights are left, and they survive only because they know the difficulties - they survive off Madden and COD because doing anything more demanding is hard.
Outside of some massive sea-change, consumers will be on 64-bit for the next hundred years. Servers may need it, but nobody else will, and that won't happen for another 20 years.
Web is not driving memory needs anymore, video will hit the wall at 4k consumer / 8k industry (I'll be amazed if 4k ITSELF gets wide acceptance in consumerland). All the other aspects of the industry that are video-driven (gaming, 3D modeling, simulation, etc) will also eventually come to a halt.
Games cannot drive innovation forever because we are hitting a finite limit on what we can do with a game engine and still make a profit. Once your effort to make the fabulous 3D effects surpasses what people will pay, you can't go any further. And right now, games are as monetized as they can ever be without pissing off the customers to the point of abandonment:
1. Day-one DLC
2. Buggy releases to get free testing from userbase. This also technically means they're further in the hole financially because they can never recover from the money/schedule slip.
3. Pay to win, even in games that are not MMOs.
4. In-game ads a plenty!
Every month we hear about some game studio closing shop. Some are just plain mismanaged, but for a large number they just hit the wall. Won't be long until only the heavyweights are left, and they survive only because they know the difficulties - they survive off Madden and COD because doing anything more demanding is hard.
Outside of some massive sea-change, consumers will be on 64-bit for the next hundred years. Servers may need it, but nobody else will, and that won't happen for another 20 years.
Last edited:
I don't know squat-jack about games, so I can't comment there.
I do know that last month my wife and I were in a Sony store and saw one of their 4K TVs. Drool.
More drool. Makes the normal HD set look not so great, especially at a larger screen size. Only "problem" is the price. Let me also comment in a slightly different way here. When I got my Nikon D3, I was thrilled with the12 Mpixel, 14-bit color depth sensor. And I can get amazing images with that camera. Images so good that my older Nikon lenses "should" be replaced. But the big, big weakness is projection. Normal 1024x768 projection looks crappy side-by-side with Kodachrome slides. Now that more and more digital cameras, even phones, have 10 Mpixel or more sensors, it's only a matter of time before someone says, "Hey let's use a 4K TV to project these images."
I don't know enough about semi-conductor technologies, but I do know that there is big money in that industry, including big money for R&D. If Silicon or CMOS stops working, then another material will be used. Sure, it might mean that power supplies would need to be updated, but that's a minor detail.
It's always dangerous to make assumptions like that - such has been proven very wrong before. Even the early 1990s, it was inconceivable that everybody would be playing full-length movies on their home computers and TVs with qualities higher than the movie theaters had at the time.
But to start using data on an exabyte scale instead of a terabyte scale, it seems to me that a whole new form of media would need to be invented. Film and video and 3D video were not new forms of media even when computers took up an entire floor. Whatever that new form of media would be, I'm not seeing any evidence of it, and can't imagine what it would be.
defaultluser
[H]F Junkie
- Joined
- Jan 14, 2006
- Messages
- 14,398
I don't know squat-jack about games, so I can't comment there.
I do know that last month my wife and I were in a Sony store and saw one of their 4K TVs. Drool.
Let me also comment in a slightly different way here. When I got my Nikon D3, I was thrilled with the12 Mpixel, 14-bit color depth sensor. And I can get amazing images with that camera. Images so good that my older Nikon lenses "should" be replaced. But the big, big weakness is projection. Normal 1024x768 projection looks crappy side-by-side with Kodachrome slides. Now that more and more digital cameras, even phones, have 10 Mpixel or more sensors, it's only a matter of time before someone says, "Hey let's use a 4K TV to project these images."
The problem you have is, you visualize your needs as the needs of the entire industry, but you need to realize that you are a corner case. Very few people have the money to spend thousands of dollars apiece on audio-visual equipment AND care enough about the quality difference to make the purchase.
People point to costs coming down thanks to the industry producing screens en mass, but there is a price floor on all things, and we have hit it with everything 1080p and below. And the silicon/crystal on screens, much like analog silicon, cannot just take advantage of a die shrink to reduce costs, because those are driver transistors, not logic. Cost reductions require development of entirely new display technologies, like IGZO (in development for a decade). These are no-longer cheap and easy to develop because the technology has reached end of the line.
TV prices are as low as they're going to get, and adding more resolution or screen area will raise the cost of the panel. If people can't tell the difference, and wonder what exactly they can watch on these new displays, they're going to buy something bigger and cheaper.
Last edited:
These are not genuine questions. These are questions that 1) demonstrate a serious lack of _any_ understanding and reading comprehension 2) are only asked to ignite controversy, i.e. troll questions.
I refuse to treat such bullshit with decency when it's totally not deserved. This whole thread should be deleted.
There's big money in the semi-conductor industry, sure. There isn't, however, big money in developing useless technology. We want devices that consume less power and generate less heat. We want devices that can be manufactured with economical yields. And we're already at a point where we can't keep shrinking dies, because we're reaching a level of physical scale where Newtonian physics is no longer a good enough approximation of how things behave, and we're now having to deal with quantum mechanics more and more, so upping the transistor requirement of every device on the data path better have a good reason behind it....and 128-bit computing just isn't a very good reason.
The last thing we need is to make it take twice as many transistors to make registers, encoders, decoders, logic and arithmetic units that get the same amount of work done. The last thing we need is another set of extensions to the instruction set. We don't need devices that are more complicated to design, are more expensive to manufacture, consume more energy and produce more heat, all for the sake of nothing....After all, having 16,777,216 TB of addressable system memory or having yet more SIMD instructions mean nothing to us.
Unknown-One
[H]F Junkie
- Joined
- Mar 5, 2005
- Messages
- 8,909
Modern 64bit Intel and AMD processors include a total of eight 128bit registers, exposed through SSE.
So yes, modern x86 chips are, indeed, already capable of running 128bit code. It's possible Microsoft is simply looking at taking advantage of more of SSE's feature-set in future versions of Windows. The OS would still be largely 32bit/64bit, but with 128bit used when appropriate to increase performance or efficiency.
Yes. From my understanding, the 16 64-bit registers are paired together and treated as 8 128-bit registers using the SSE instructions.
If you read carefully, that is NOT what I said. What I said was that as more and more cameras have 10+ Mpixels .... That's not me as an "early adopter" or "corner case." That's saying that while in this case I was an early adopter, 3-4 years later, that same technology goes mainstream. There is a lot of research into consumer market behavior to back up this point. See http://www.amazon.com/Crossing-Chasm-Marketing-High-Tech-Mainstream/dp/0060517123
So 4-5 years ago, trying to sell mainstream consumers on 4K TV would have been a complete mistake. In 2014, we can sell that technology to "early adopters." By 2015-2016, probably "early majority." I'm using terms discuss in the book above.
As an Amazon Associate, HardForum may earn from qualifying purchases.
defaultluser
[H]F Junkie
- Joined
- Jan 14, 2006
- Messages
- 14,398
You can whip out all the fancy marketing terms that you want, it still does not change the fact that the market is in a completely different stage from that which you are visualizing.
This"growing demand" system you visualize does not exist for flat panel displays because it's already come and gone, just like the digital camera market you've highlighted. Both televisions/monitors and digital cameras have hit their price minimums, and every improvement is incremental.
And yes, I just called 4k monitors "incremental." We have already had double the resolution of 1080p for a decade now, with very little progress. Today resolutions higher than 1080p are more common than they were in 2004, but they've grown as much as they possibly can due to reaching absolute minimum monitor costs. These screens are still hundreds of dollars more expensive than 1080/1650/720p devices, and they're not going to fall any further.
This research assumes there is demand for your product, which there is not. The majority of consumers will not be able to visually see the difference between 4k and 2k, given that they typically buy TV sets under 50" in size, and tend to sit at least 8 feet away from the set. Monitors will sell better, but no better than 27" 1440p displays do already.
Unless you can show a brand-new use case and content system that will change things, there is no way you can expect this demand curve to reverse itself. And let me tell you personally man: you are the ONLY person I know who would consider upgrading their television in order to improve your picture slideshow experience. Most people I know don't even know they can view pictures on their television, and don't care.
Last edited:
As an Amazon Associate, HardForum may earn from qualifying purchases.
Right now the idea is a solution looking for a problem of course. The move to 64-bit was very much necessary and well-timed and we just don't have that scenario for 128-bit yet. There are still a lot of inefficiencies to squeeze out of what we currently have.
B00nie
[H]F Junkie
- Joined
- Nov 1, 2012
- Messages
- 9,324
The main issue against 4k tech is that there are no content providers. Once you get tired watching the odd 4k demo video you have nothing else than your home photos to use it with. It quickly eats the appeal from the devices.
Same exact problem is with 3D - there are practically no 3D content providers so why would people start buying 3D tv:s. No point buying a 4k tv if all you have is a regular BD to watch movies from.
As an Amazon Associate, HardForum may earn from qualifying purchases.
techtechie
Gawd
- Joined
- Oct 7, 2013
- Messages
- 571
i know this quote is from the first page but what i had to say was nagging me throught the comments.
its a pattern. double the 'bit' value double the time so wait about 20 more years and maybe just maybe 128bit will be where 64bit currently is.
I think it is likely we will see 128-bit registers and possibly 128-bit data buses eventually, 128-bit address buses is unlikely, though it may be similar to the 64 bit implementations where 128-bit address bus is designed in to the architecture but practical implementations only use some number of low order bits (similar to how AMD64 used only the lower 48-bits of the address bus).
The issue will be pushed by necessity. There are lots of applications for 64-bit computing (and there were even back in the 32-bit days), then there was also the memory addressing issue. By contrast there are not a lot of things that require computing with 128-bit precision. This may slow down the push for 128-bit general purpose CPUs.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
I think it will be quite a bit longer than that to put the equivalent of 268,435,456 64 GB dimms into the typical 1 to 4 slots a PC has.
I think you guys that are so quick to dismiss a 128 bit OS in the near future are overlooking something vitally important here.
Take a look.
/s
Take a look.
/s
You sick, twisted person, you.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
That article is complete BS. They mention an IA-128 processor for windows 9. IA = Itanium. Itanium is a server only CPU that has not been updated in years. Also Microsoft cancelled development for Itainum around 4 years ago. On top of that why would desktop OS be developed for a server only processor and why does this need a 128 bit address space? How many people will be able to afford the $500 billion dollars of memory? And how will they put 268,435,456 slots on a motherboard?
Last edited:
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
lol
You didn't catch the part where he was joking, by chance?
By this definition we already have 128b operating systems.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,933
Nope, Not at all.