Slashdot Mirror
Microsoft Leaks Details of 128-bit Windows 8
Barence writes "Microsoft is planning to make Windows 8 a 128-bit operating system, according to details leaked from the software giant's Research department. The discovery came to light after Microsoft Research employee Robert Morgan carelessly left details of his work on the social-networking site LinkedIn. His page read: 'Working in high-security department for research and development involving strategic planning for medium and long-term projects. Research & Development projects including 128-bit architecture compatibility with the Windows 8 kernel and Windows 9 project plan. Forming relationships with major partners: Intel, AMD, HP and IBM.' It has since been removed."
Who needs 128? I haven't even used all 64 of my current bits yet.
-l
Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
Why, is Google asking that because they're running out of addressable memory space again?
A positive attitude may not solve all your problems, but it will annoy enough people to make it well worth the effort.
Well in *my* OS, the volume goes all the way to 11!
I went to eat some animal crackers and the box said, "Do not eat if seal is broken." I opened the box and sure enough..
has been transferred to another department - the Pit of Despair.
Just because I can hook a shark from a boat, I do no offer to wrestle it in the water.
Sounds fishy to me. Producing a 128 bit operating system before 64 bit is in wide use just seems like a waste of time. Of course this is microsoft so who knows by the time they have that out we may actually have 128bit chips.
We're doing five blades.
With Windows 6.1 being Windows 7, does Windows 8 actually mean Windows 7?
The is no Robert Morgan that works at Microsoft. Not sure who this guy is but if he does work at MS its not his real name.
Well, not anymore, anyway. :-)
Best "String" Ever!
That would make Windows a 128 bit wrapper around a 64 bit implementation of a 32 bit extension for a 16 bit patch to an 8 bit operating system, originally coded for a 4 bit microprocessor, written by a 2 bit company, that can't stand 1 bit of competition.
here.
/...
Yeah, well I'm working on an OS that'll be 129 bits!
http://alternatives.rzero.com/
Windows 7 isn't even officially released and already nonsense is leaking about the next release with promises they can't keep.
FIrst let them release WinFS.
Pretty good is actually pretty bad.
It refers to a 128 bit filesystem ala ZFS, not the whole OS.
Either we're not reading the same article, or I suspect you didn't read it at all. At no point is a filesystem mentioned.
16.8 million terabytes of RAM should be enough for anyone.
Well, that settles it, then! Why on earth would I buy a paltry 64-bit Windows 7 when a much shinier and newer 128-bit Windows 8 is right around the corner? I'd best hold off until then! Thanks, Microsoft!
Demanding constant attention will only lead to attention.
Noooooo! I want to be able to say I have a 23488102 bit OS if that's the size of my bzImage! And once I have 1TB of porn I can call it a 8.79609302*10^12 bit operating system!
Seriously - it's one thing for some IT marketing types not to know that a 128bit OS would need a 128bit processor (which would be a Big Thing, especially if HP were getting back into the market of CPU design and manufacture), but for the submitter and eds to not point it out makes it look a little daft.
Moderation Total: -1 Troll, +3 Goat
That was what I thought at first glance, but no it is definitely referring to the architecture. (hence the forming relationships with vendors).
Core 2 Duo x128 around the corner?
It refers to a 128 bit filesystem ala ZFS, not the whole OS.
Either we're not reading the same article, or I suspect you didn't read it at all. At no point is a filesystem mentioned.
I'm with you, I don't know where he got filesystem from:
The senior researcher's profile said he was: "Working in high security department for research and development involving strategic planning for medium and longterm projects. Research & Development projects including 128-bit architecture compatibility with the Windows 8 kernel and Windows 9 project plan. Forming relationships with major partners: Intel, AMD, HP and IBM."
Clearly says architechture.
I am the lawn!
It refers to a 128 bit filesystem ala ZFS, not the whole OS.
That makes a lot more sense, considering that there are no x86-compatible 128-bit CPUs available or even being publicly discussed. If I'm not mistaken, the 32-bit x86 CPU was around for 20+ years before the 64-bit extensions were added, and several years later we're only beginning to get widespread deployments of a 64-bit Windows. We probably won't see 128-bit Windows widely available until 2025 at least.
This has been discussed on OSNews and it is most likely about the filesystem or FPU and not memory addressing.
http://www.osnews.com/story/22301/128-Bit_Support_in_Windows_8_9_
They haven't even got 64-bit mig. done / smooth for their users, and they're looking at 128-bit already? Is this some kind of a joke? Or have they just realised they've missed the 64 bit boat and they're just going to work seriously on the next step up?
Not necessarily, they could skip a step entirely, just like WinXP was followed by Win 7.
`echo $[0x853204FA81]|tr 0-9 ionbsdeaml`@gmail.com
So, that's right on schedule for Windows 8 then.
'Sensible' is a curse word.
Quoth Balmer, "Let's see hackers find our security holes in this address space!"
None of the linked articles say that the 128 bits is for the filesystem only, but I still believe you're right:
Making the entire os 128-bit would simply waste a _lot_ of memory, for zero real gain. (Rather the opposite: A larger working set always leads to slower code.)
Having 128 bits available for filesystem/storage makes it quite feasible to have globally unique addresses for everything, across huge populations of machines.
This has been done before, afair IBM has used a 128 (or 129!) bit address space for their AS400 platform, where everything is memory mapped.
I.e. there is no visible file system, you just access objects by address (which is really a handle).
I believe Amazon's cloud storage is similar, in that the only way to access a blob of data is via a 128-bit handle.
Terje
"almost all programming can be viewed as an exercise in caching"
- BSOD now in 4D (3 spatial dimensions and time, more precisely from 1988 to 2015)
- That is what requires Security Essentials to have a string sample in memory of every Windows virus/trojan before 2006
- Bill Gates finally agreed that 640k wasnt enough for everyone.
- Codenamed Windows TNG, where no bit has gone before
- You actually will need all that memory to not require swapping (unless you load more than 3 apps)
This seems like major horse poo, there simply is not reasonable general purpose architecture available for 128bit and I even fail to see a requirement for the far future. Sooner we will see shard architectures based on massive parallel multiprocessing like today GPUs but they will be per core a lot simpler than today 64bit systems, not more complex.
"Life is short and in most cases it ends with death." Sir Sinclair
The is no Robert Morgan that works at Microsoft. Not sure who this guy is but if he does work at MS its not his real name.
Well, we don't know who you are, either, so why should your input on this be paid any attention?
That makes a lot more sense, considering that there are no x86-compatible 128-bit CPUs available or even being publicly discussed.
What about non-x86, 128-bit chips? Are there any of those? (I don't know of any, but I probably wouldn't anyway.)
The NT kernel has always run on other architectures... throughout the years, PowerPC, Alpha, MIPS, and Itanium. If there are places out there with 128-bit chips, I could believe that MS might work on targeting them.
Just like that other chap who was always making wild statements about what Microsoft was going to do next.
They let him go too. What was his name again? Will? Billy? ...something.
Genesis 1:32 And God typed
In terms of memory, 64 bits can address 18 exabytes. Even Google isn't going to be using that for a decade or so. Assuming Moore's law continues, it will be about half a century before PCs need that much RAM. Dealing with 128 bit numbers for mathematics is of limited use (if you do want to deal with them, you'll probably have a need for 256 byte and 512 byte numbers as well).
And it's not like there's been much perception of a need for 128 bit CPUs. 64 bit processors have been around since the 1960's with fairly mainstream CPUs sine the early 90s. I don't think this is like RAM. I think there's a limit to how many bits we can use.
Yeah right. Gob like the mersey tunnel.
Microsoft has been aiming towards high-performance computing recently, working with companies like Nvidia. If you are going to have racks and racks of CPU's/GPU's, it would make sense to have everything accessible using a single memory space.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
It refers to a 128 bit filesystem ala ZFS, not the whole OS.
Oh. I thought they pulled a Vista again and the 16 exabytes of RAM provided by 64-bit was not enough for their latest crime against humanity.
However, none of those other architectures were very popular, because they lacked applications. I don't even think I've seen a 128 bit processor. I can't imagine Microsoft would target a market so small, than many don't even know it exists.
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
But I'll tell you how it will end.
The final architecture EVER will be 640-bit. And that WILL be enough for everyone.
Mit der Dummheit kämpfen Götter selbst vergebens
Not only does it say 'architecture', it also says 'architecture compatibility'.
Why is that important? Because it does not mean that Windows 8 will necessarily be 128bit, just capable of being 128bit - for all we know, his entire role is ensuring that the teams code to a set standard which allows ease of porting to 128bit in future.
Clearly says architechture.
Okay, but the question is what does that mean? If it just means 128-bit operations or registers, then that's been around since the original SSE. If it means 128-bit addressing (like it usually does), then who the fuck is making those chips and why? Very few 64-bit chips actually support the full 64-bits of address space (certainly not Intel or AMD), simply because there's no need. You could make every computer on earth part of a huge shared-memory system and have room to spare, not that you'd ever do such a thing. Once systems get far enough apart, shared memory stops making sense as maintaining coherence/consistency becomes too much of an overhead. If you were building a cluster as a shared memory system, and each node had 1 TB of RAM, you could fit ten million nodes in before you started to have address space problems. Even the most wasteful of Stupid Virtual Memory Tricks aren't going to put a lot of pressure on 64-bit addressing any time soon.
I mean I guess I can see the point for the distant future, and hey who the hell knows when Windows 9 is planned for much less will actually arrive, so it can't hurt to make sure it's 'compatible'... I'm just more surprised that any of the partners listed would have 128-bit on even far-reaching roadmaps.
The enemies of Democracy are
long long long?
really long long?
Someone else posted a link to an ArsTechnica article about this. They had more info from the LinkedIn post, which indicated that the work was being done to target the IA-128 instruction set (which is currently only available as a simulator, no actual silicon, *yet*). But, since Intel hasn't abandoned Itanium yet, and they are targetting it at Enterprise and High Performance Computing, I could totally see Intel evolving the Itanium architecture from 64-bits to 128-bits. After all, there are a few servers in the world that handle truly epic amounts of data, and really might be able to use more than 64-bits.
It's probably that they are laying the groundwork now, for release 5 or 10 years down the road.
Really, it is? um, NO. Win 7 64 bit is a joke, well aside from the fact that it works (for which you deserve credit)
Uh. So... what did you want it to do, if not work?
but why is there even a 32 bit version? why in 7-64 is there a program files (x86) folder? that was your kludge for working with 32 bit apps? Really?
It's been that way since XP x64. It separates 32 bit apps from 64 bit apps in the directory tree. What's the problem? You can install programs anywhere you want. Just like always. And there is a 32 bit version because not everyone runs 64 bit processors. Say, for example, the Intel Atom. Or my old original Intel dual core laptop. Oh, but I guess Microsoft should intentionally shoot themselves in the foot and not release a 32 bit version and thus lose that entire market share. Pardon me for asking. By the way, why do Linux distros insist on keeping 32 bit versions?
It also clearly says "compatibility", e.g. no actual 128-bit code yet, but they'll make sure it won't break in the future. Perhaps they learned their lesson from 64-bit XP.
Must I remind that no one will ever need more then 640k of RAM!?!
On what processor would this run? Intel and AMD have shown off no 128-bit CPUs at this point. There is no point in having an OS that doesn't have hardware support to back it up. You can't have a 128-bit OS on a 64-bit CPU.
Now in terms of CPU what they might be talking about is floating point precision. There is some interest in getting quad precision FP. The new IEEE 754 spec has it and there are some research apps that could use it. So that might be another area they are interested in.
However 128-bit memory addressing? No way. Even 64-bit Windows doesn't use all 64-bits at this point. I mean it implements a flat 64-bit virtual address space, but it only is setup to deal with like 2TB of actual RAM. The processors are the same way. The Core 2 is a 64-bit CPU, but only has like 40-bits of actual address lines. Reason is that nobody could actually build a system with 64-bits of RAM right now, so there's no point in putting in hardware that won't be used.
I would do it, but there isn't a -1, Wrong option.
No, that was Windows XP... back at the beginning of this decade. Next question.
64 bit to 128 bit is a bit more then doubled...a lot of bits more. Remember binary, not base-10.
As for why that number...these things are always in creasing in such stages, 2, 4, 8, 16, 32, 64, 128...next would be 256
I do not support "The Man". I also do not support your irrational stupidity
Maybe because he was thinking logically? There are already 128 bit filesystems out there, ala ZFS. Why? Because with huge servers and clustering you can get some insanely huge numbers when it comes to HDD space. But as I'm sure we all know when it comes to a 128 OS, we are usually talking about addressing, ala 16bit, 32bit, 64bit.
As it is now there isn't even any stock machines being built (that I know of) that can support even 1/100th of the amount of RAM that can be addressed in 64bit, let alone any need in at least a decade or two for 128bit addressing. Lets face it-16 exabytes of memory is a whole damned lot in anybodies definition, and I doubt anybody here can think of any reason that we could possibly affordably build a machine that hits that limit in even 20 years. Filesystems on the other hand? Well we already have drives hitting 2TB, so hitting the 64bit limit there, while it will still take awhile, is doable. And let us not forget that MSFT has always been the kings of "me too!" and don't like other groups having features that they themselves don't have, so I have no doubt that when they saw ZFS is 128bit that they said "we need to have 128bit capable servers too!"
So in conclusion it doesn't need to be said, as it is simply common sense. Even at our current rate of RAM size growth it would probably take a good 30 years before we can manufacture RAM sticks that pack enough density that you could hit the 64bit addressing limit without requireing thousands of sticks. With the explosion of hard drive space on the other hand hitting the 64bit limit in large clusters is certainly possible in the near future. Therefor it makes sense that since MSFT does have several server products, and servers are requiring ever increasing amounts of space, that MSFT would have a team working on 128bit file system support for a later version of Winserver. You really don't need to be Colombo to come to that conclusion, as it just makes sense, whereas a full 128bit OS simply does not and will not make sense for most likely several decades. And while we all know MSFT is slow when it comes to releasing OSes, surely they won't be THAT slow when it comes to Win8 and Win9.
ACs don't waste your time replying, your posts are never seen by me.
I'm pretty sure he means that there are twice as many bits. Which is true.
Microsoft Research does a LOT of this type of investigation and research. However, there is a world of difference between researching compatability and 'planning to add'. Whether or not he really works for Microsoft, the claim that he is in R&D makes the claim that Win8 will provide 128 bit support a major stretch. Very misleading headline.
Yeah, I think I remember him. He thought the Internet was a passing fad, claimed he'd single-handedly defeat spam, promised Microsoft was taking security seriously in 2000, all kinds of nonsense like that. Had a really dorky haircut too.
You are in a maze of twisty little passages, all alike.
I'm still confused.
What's the point of having 128 bit compatibility? 128 bit CPUs don't even exist yet. Heck most of us are still just using 32, and haven't even visited the 64 generation yet.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
How's that old joke go, again? "Windows is a 32 bit extension to a 16 bit operating system designed for an 8 bit processor by a company that isn't worth two bit" or something like that?
This 128 bit operating system thing is a joke - or, more accurately, marketing nonsense. "Windows 8" is so far away at this point that it's unlikely they've slated this feature; 128 bit processors are even further away than Windows 8 might be. I say 10 years at the smallest inside.
Now, if we're talking about something other than processors, it might be a possibility (filesystem?). But I doubt it.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
PAE doesn't "hide" memory, really. You can only address 4GB (i.e. a 32-bit address space) of virtual memory at once but that can be *anywhere* across the 36-bit physical address space. As long as no individual app needs more than 4GB of memory you're (mostly) OK. The kernel can alter the mappings as it needs to poke at anywhere interesting in all of physical RAM. It's less efficient than mapping it all in at once but you can manage quite well.
In related news, razor-blade manufacturer Gillette has announced their expansion into the personal computer operating system market. When he was informed of Microsoft's plans for future versions of Windows, CEO James M. Kilts is quoted as saying, "Fuck everything, we're doing 512 bits."
I once read that it took us 30 years to figure out 8 bits then about 5 years to exhaust the abilities of 8 bits so we moved to 16bits which we exhausted in 10 years so we moved to 32bits which took us about 20+ years (this post written on a 32 bit machine which has thus far been far more reliable than the wife's 64 bit machines) to exhaust so following this logic it will take us 40+ years to exhaust 64bits. Does this then mean I can expect Windows 8 in about 2050?
Despite the colossal way we've advanced in computers over the last ten years, I really really think 16EB really should be enough for anyone until at least 2050...! Surely it'd be more beneficial to tack on additional instructions (e.g. AltiVec-style SIMD ops) to current architectures rather than design a whole new one that we don't remotely have the capability to explore...?
Moderation Total: -1 Troll, +3 Goat
Here is a better link for the story: http://arstechnica.com/microsoft/news/2009/10/microsoft-mulling-128-bit-versions-of-windows-8-windows-9.ars/
Let me guess: you've never written any ring 0 code for x86. PAE doesn't hide the memory. It modifies the page table structure slightly (so does 64-bit, by the way, it makes the page tables deeper which makes every TLB fault slower). You have a 32-bit virtual address space and a 36-bit physical address space. No process can see more than 4GB of RAM, but if you have two processes then they can each see a different 4GB of physical RAM. None of my processes currently uses more than 760MB of address space, but I have 3GB of RAM and 3GB of swap used, so with a PAE system and 8GB of RAM each process would be using physical memory and I'd have 2GB for filesystem cache.
Oh, and when people talk about PAE, they also often mean PAE or PSE. PSE just makes pages bigger (up to 4MB), which can be used to address 64GB of RAM without changing the size of the page tables. This is better in some situations, because it involves smaller page tables and fewer TLB faults, but it means that you are swapping 4MB at a time, which can be very slow if you are swapping a lot.
I am TheRaven on Soylent News
Before we see a Windows 8 I'd like to see a real Windows 7. What they call Windows Seven is actually Windows 6.1, a glorified Windows Vista. Type in "ver" in a command prompt to check out. Don't get me wrong, I've be using Seven since the RC and I quite like it. It's a real improvement over XP and Vista. But I still think it's overrated and the name is actually just a marketing thing.
without the hardware to support it
So you think 128 bit processors aren't coming? Where have you been in the past, oh, 40 years?
I personally think it's kind of silly of Microsoft, considering that they have historically offered piss-poor support for 64 bit processors, I don't know why suddenly there would be a rush for 128 bit support, unless it was for Windows 8 Ultimate Gold Pro where you get to pay $700 more so that you can actually run as many windows open as you want at the same time.
Or perhaps they'll need a 128 bit OS to do all that DRM number crunching cryptography stuff without slowing your machine down to a crawl.
Either way, companies DO plan ahead. It's logical to assume that Intel/AMD (if they are still around) will one day provide 128 bit, then 256 bit, then 512 bit processors. The wider your input stream into your CPU, the faster you can perform those operations. Increasing the clock speed has (barring any new technological discoveries) reached physical limits namely temperature and the speed of light- or at least electricity. So the only way to get faster is to a) add more cores and b) process more bits per cycle.
Seven puppies were harmed during the making of this post.
The discovery came to light after FORMER Microsoft Research employee Robert Morgan carelessly left details of his work...
According to Wikipedia
http://en.wikipedia.org/wiki/Glenndale_University
this University isn't even accredited!!
Maybe they mean optimizing windows kernel using SSE2 instructions (xmm registers are 128-bit wide)? It improves performance of many memory transfers, and currently and makes a lot of sense in quite a lot of applications - even the kernel itself.
The original IBM System 38 and its descendants, such as OS/400, OS/500, etc., had a 128-bit address space. In these architectures, the large number of address bits were used to provide an address space that spanned both memory and disks and was used to provide processor-level protection for objects stored there. Using large address spaces to ensure hardware protection of system objects is a good start on a highly secure OS and is probably where this is going.
And Intel is no stranger to hardware object protection, either. The iAPX-432 chipset, although not a commercial success, showed that hardware-level protection of objects is feasible, with more complex access controls than can be provided with reasonable performance than with software implementations of complex access control schemes (note I said complex - one of the reasons that the chip failed commercially is that, besides having a braindead two-chip implementation and instruction lengths that varied at the bit level, it could not support simple protection schemes as quickly as software was able to do). Intel is looking for what to do with the extra transistors that feature shrinks provide - adding better protection at the hardware level might be a win.
That is all.
Not too long ago (15-20 years, maybe?) 64-bit processors would have been unheard of on the desktop. I see 64-bit being stretched as we put more high-definition video into our datasets. And then we'll have the next "ultra high def" format that will stretch it even more. And then you have a small (in terms of units shipped), but very profitable business in supercomputing. Protein folding and subatomic research folks would probably jump at the chance to rerun their simulations with a higher resolution.
Remember, You are unique...just like everyone else.
Thank you, finally a person who seems to know about PAE. I always growl when people speak about 32bit versus 64bit cpus and don't seem to have any clue.
- Raynet --> .
If we just move to 512bit, then we can have a bit for every atom in the universe. that'll solve any address space issues.
But really, there's only 2 reasons that I can think of for "bits" #1. address space. 64bit is plenty for a LONG time #2. computation. Who needs 128bit integers to represent real world objects?
Even the most wasteful of Stupid Virtual Memory Tricks aren't going to put a lot of pressure on 64-bit addressing any time soon.
You heard it here first, folks: 64-bit ought to be enough for anybody.
I only post comments when someone on the internet is wrong.
If it is true, I suppose that Windows the 8th will be published in a far future. I do not know a 128 bit achitecture that is also commercial. The same fact for Windows Vista delay.
Shared memory space among lots of computers, using IP (possibly IPv6) as a protocol.
That's probably what they are referring to if they mean 128 bit address space (not datapath).
the double of 64 bits address space is ... 65 bits of address space ;)
Working in high-security department
Any one else notice the irony of entering such a phrase and following it with proprietary information?
Upgrade options, duh!
They want to help consumers with more choice. Just imagine what the NEXT version of this will look like:
http://mossblog.allthingsd.com/files/2009/08/windows-upgrade-chart.png
There are two types of people in the world: Those who crave closure
You don't need 128 bits for addressing. 2^32 is "only" 4 gigabytes, which was always achievable in theory and actually achieved in practice over a decade ago.
Having a memory — RAM or disk — above 2^64, however, is not achievable in even in theory... 2^64 is only 100 times less, for example, than the estimated number of sand-grains on Earth.
Being able to process as much as 128 bits in one CPU-instruction is nice, and SSE extensions allow that. But neither size_t nor off_t need to exceed 64 bits. Ever... In fact, in the amd64 instruction set, only 48 bits can be used to address memory — the rest are for the CPU instruction, so that both the operation and the operand fit in one 64-bit word. The amd64-architecture is thus "limited" to 256 TB — that's the largest RAM an amd64-machine can have and the largest file and amd64-machine can mmap.
64-bit systems were truly useful, because — by making size_t and off_t the same, they allowed software to be rid of having to segment access to files, which could, potentially, be too large to memory-map in their entirety (many legacy mmap-implementations are still limited to 2- or 4-Gb files). 128-bit systems are not adding that benefit...
(And, of course, most systems — including even the most modern Linux and BSD — still have rather poor mmap-implementations, compared to their highly-optimized read and write calls... But that's another topic...)
In Soviet Washington the swamp drains you.
Specifically current Intel Atom "Diamondville" N- and "Silverthorne" Z-series processors intended for netbooks and UMPCs, other Atoms (current "Diamondville" 230/330, and the upcoming [Q1, 2010] "Pineview" D- (replaces "Diamondville" 230/330) and "Pineview" N- (replaces the current "Diamondville" N) series Atoms are all 64-bit.
The senior researcher's profile said he was: "Working in high security department [emphasis mine] for research and development involving strategic planning for medium and longterm projects. Research & Development projects including 128-bit architecture compatibility with the Windows 8 kernel and Windows 9 project plan. Forming relationships with major partners: Intel, AMD, HP and IBM."
My first reaction was that if you can't fix the security problems in the people, you surely can't expect to fix the security problems in the software. But that might be a little hasty.
My guess is that the actual security gaffe here was little or nothing. He mentioned he worked in this department, and that they have future plans that exceed today's capabilities. Meh. So what. If he had posted the details of what he was doing, then it would have been newsworthy. As it is, this barely notable. Any one of us here could probably guess that MS likely has people looking into the progression beyond 64 bit technology.
It is reasonable to believe that at some point in the next several years the hardware companies he mentions will have some plan to start building 128 bit cpus. My guess is that this guy's job is to make sure that MS has input into the design process where it can, and to provide feedback to the MS dev teams so MS can start planning to include compatibility features relatively early on, to hopefully be the OS of choice when this hardware someday becomes available. I'm guessing that Windows 8 probably won't be seen for a long time. The article mentions 2012, but given MS's rush to push out 7 to stem the bleeding caused by Vista they may rely on it for longer than normal, much like they did with XP after the ME debacle. If I were writing an OS that would likely debut in 4 to 8 years, I would probably want a heads up from the hardware vendors about how to write an OS for their next gen proc. Also, if MS were planning a future move to a fully 128-bit OS, they might start by inserting 128-bit code into a 64-bit OS.
I prefer rogues to imbeciles because they sometimes take a rest.
On the other hand maybe Mr Gates himself needs the floating point precision in order to balance his check book? I once read a story of how the IRS had to rewrite their software and upgrade some hardware in order to do his taxes a few years back. Not that I believe that story completely, but you have got to question why the average person, not running a home brew Cray tower, would need that kind of system. Maybe the LHC?
None of the linked articles say that the 128 bits is for the filesystem only, but I still believe you're right:
Making the entire os 128-bit would simply waste a _lot_ of memory, for zero real gain. (Rather the opposite: A larger working set always leads to slower code.)
Having 128 bits available for filesystem/storage makes it quite feasible to have globally unique addresses for everything, across huge populations of machines.
This has been done before, afair IBM has used a 128 (or 129!) bit address space for their AS400 platform, where everything is memory mapped.
I.e. there is no visible file system, you just access objects by address (which is really a handle).
I believe Amazon's cloud storage is similar, in that the only way to access a blob of data is via a 128-bit handle.
Terje
Since Win8 / Win9 won't be out for 5/10 years...
Why am I getting flashbacks to a discussion that people had back in the 8 bit days?
"Making the entire os 32-bit would simply waste a _lot_ of memory, for zero real gain. (Rather the opposite: A larger working set always leads to slower code.) ... Having 32 bits available for filesystem/storage makes it quite feasible to have globally unique addresses for everything, across huge populations of machines."
I never heard this discussion, but you know it happened. Probably almost verbatim.
I'm still confused.
What's the point of having 128 bit compatibility? 128 bit CPUs don't even exist yet. Heck most of us are still just using 32, and haven't even visited the 64 generation yet.
Maybe because it's easier to include now the ability to extend compatibility to 128-bit processors instead of trying to bolt it on later? Who knows, maybe Microsoft really did learn something from their experience with Windows security.
If you were building a cluster as a shared memory system, and each node had 1 TB of RAM, you could fit ten million nodes in before you started to have address space problems.
Will it then be able to run windows with all features turned on?
Who would win this election: Andrew Weiner vs Andrew Weiner's weiner.
He's right, there's no Robert Morgan here (well, there's a v-, but no Robert Morgan in R&D).
I'm pretty sure that "64" and "128" are in base 10 not binary. I'm also pretty sure that 64*2=128. Now 2^64 is much less than 2^128. But it's also true that 10^64 is much less than 10^128. In short the problem is not the base.
Not too long ago (15-20 years, maybe?) 64-bit processors would have been unheard of on the desktop. I see 64-bit being stretched as we put more high-definition video into our datasets. And then we'll have the next "ultra high def" format that will stretch it even more. And then you have a small (in terms of units shipped), but very profitable business in supercomputing. Protein folding and subatomic research folks would probably jump at the chance to rerun their simulations with a higher resolution.
Just to put this into perspective, the forthcoming IBM Sequoia supercomputer will have 1.6 petabytes of RAM, and only a very small fraction of this can be accessed by a single compute node. The total amount of RAM in this machine is still 4 orders of magnitude smaller than what can be addressed with a single 64-bit pointer.
>>>64 bit to 128 bit is a bit more then doubled...a lot of bits more. Remember binary, not base-10.
Yeah. Exactly double the number of bits.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
...Microsoft is planning to support 5-letter acronym hardware, even though the industry has not yet completed the migration from 3-letter acronyms to 4-letter acronyms. The 5-letter version of the operating system will support 4-letter acronyms, but will not be backwards compatible with shorter acronyms, such as the popular "CPU" and "RAM" used widely in the computer industry.
There's no failure quite as dissatisfying as a complete and total solution to the wrong problem.
The subject is quite stupid. What details? Probably an engineer with a thick accent told the HR manager something about 128 bit (maybe fs, maybe optimizations), and by the time it appeared on the website it went through several revisions each time by people with less and less understand of what they are writing. That's the corporate way of doing things.
In a single address, yes 128 bits is double 64 bits. In terms of arch, this generally refers to the total available address space. 64bit CPUs aren't double 32bit CPUs.
In reality, 128bit arch is 64bit SQUARED.
Well, 16 Exabytes of RAM (ok, just 8 if you use signed relative jumps) ought to be enough for a quite long time. Long enough to develop another OS, that is for sure.
Rethinking email
You're right! 64 bits should be enough for anybody. I think I remember reading that somewhere.
qntm.org
None of the linked articles say that the 128 bits is for the filesystem only, but I still believe you're right:
Making the entire os 128-bit would simply waste a _lot_ of memory, for zero real gain. (Rather the opposite: A larger working set always leads to slower code.)
Right. There's no widely-used 128-bit-native processor architecture either. And there is no reason to have 128-bit address bus either.
I don't think there are 2^128 bytes of DRAM on the planet, even. Lessee... that's 2^98 GiB. Which is almost 10^20 GiB of RAM for every single person on the planet. I think that I personally can account for 10 GiB or so. Maybe 100 GiB if my parents have a secret DRAM trust fund for me that I don't know about. So yeah, 128-bit memory addresses are waaaaay off. I believe current 64-bit processors are currently limited to 40-bit external address buses... that'd be 1 TiB of RAM.
My bicyles
No, it means a 128-bit architecture will still be able to run Windows 8.
That is, the architecture supports a different mode that the Windows 8 kernel includes.
Knowing the history of teh bits, this simply means Windows 8 will be available in both 32-bit and 64-bit versions, and 128-bit processors will be able to run in 32-bit mode, but not 64-bit mode.
So yet again, we will be stuck without 64-bit drivers or optimization, let alone 128-bit drivers or optimization.
32-bits should be more than enough for anybody.
(I HOPE Windows 8 is 64/128, and 128-bit processors are 32/64/128, but I know better than to expect anything sensible).
Somehow, we all managed to survive from 1984 to 1995 by swapping 64k chunks in Expanded Memory. I remember writing assembly to do it, and I personally do not miss that headache. That being said, old ideas die hard, and if we can get some larger page sizes (how about swapping that 4th GB in address space to point at a 5th, 6th, etc?), almost all reasonable applications (by today's standards) could fit in the expanded memory space.
In these architectures, the large number of address bits were used to provide an address space that spanned both memory and disks and was used to provide processor-level protection for objects stored there. Using large address spaces to ensure hardware protection of system objects is a good start on a highly secure OS and is probably where this is going.
But, even 64 bits is enough for that for a long time.
Since you can address over 17 billion terabytes with 64 bits, that means that even with a doubling of storage density every year (which is much faster than things are really happening), that means we have over 20 years before arrays of a couple thousand disks would start to reach the limit.
By then, there will be 128-bit CPUs. So, unless Windows 8 is targeted for 2020, it really doesn't need any 128-bit features.
Exactly.
Windows 8 is scheduled in launch in 2048.
I often refer to them as
Killer
Maybe
Giggle
Terror
Pedo
Once people start using PB-scale storage, I'll need to decide on one for EB.
The last thing the 486, I mean 586, I mean 686, I mean x86, I mean x86-64, I mean IA64, ISA needs is more instructions.
I believe the whole system was designed around 128bit pointers from the very beginning. It used a 48 bit processor between 1988 and 1995 and did a nearly seamless transition to a 64 bit architecture around 1995.
http://en.wikipedia.org/wiki/IBM_System_i
The AS/400 was one of the first general-purpose computer systems to attain a C2 security rating from the NSA (Gould UTX/C2, a UNIX-based system was branded in 1986[4]), and in 1995 was extended to employ a 64-bit processor and operating system.
In 2000 IBM renamed the AS/400 to iSeries, as part of its e-Server branding initiative. The product line was further extended in 2004 with the introduction of the i5 servers, the first to use the IBM POWER5 processor. The architecture of the system allows for future implementation of 128-bit processors when they become available. Existing applications can use the new hardware without modification.
For the uninitiated, v- and a- are Microsoft's way of identifying vendors (Microsoft is their client for whatever reason) and CSGs respectively (a fancy acronym for contractors).
You realize that we are at the end of 2009 right? And that Windows 7 is just now being released? MS might just be considering that in 2020 we might still be running Windows 8. I know that only looking 1Q into the future is hip for businesses, but maybe MS is looking farther ahead than that. We are still living with issues created by PC limitations from the 16-bit era. Probably a bunch from the 8-bit era that I am not aware of. I see no problem with planning ahead so that they can get the transition done sooner rather than later, or just make the transition smoother than previous bit size transitions.
Itanium?
I'd welcome Windows tying its future to the success of that chip...
Well, they don't really need more than 640KB, it's just that ASCII Crysis isn't nearly as entertaining...
A lot of the systems I see these days throw GUIDs (128-bit integers) around like confetti, using them for identifiers on everything.
On 32-bit they suck, since they are 4 times larger than the native integer datatype. On 64-bit they suck less. I'm guessing on 128-bit native processors, they'd suck about half as much again.
There is a market for this, especially for selling Windows to entities utterly obsessed with tracking every object on Earth (basically, big corporate marketing departments and governments). Governments want to put GUIDs in banknotes. Corporations want to put GUIDs in EVERYTHING. Some of them already do.
Wouldn't it be interesting if the equivalent of the entire Internet was accessible using direct memory addressing instead of IP addressing.
Having a memory — RAM or disk — above 2^64, however, is not achievable in even in theory... 2^64 is only 100 times less, for example, than the estimated number of sand-grains on Earth
So? There are more efficient encodings than one byte per sand-grain, you know.
As it turns out, 2^64 is much smaller than Avogadro's Number, the number of molecules in a mole of a chemical compound. If you could find a way to encode information in a 3D hunk of silicon, such that you needed slightly more than 1000 atoms to store each byte, 2^64 bytes of storage would amount to a bit less than one ounce of bulk silicon, occupying less than one cubic inch.
I FULLY expect to see secondary storage approaching this density within the next few decades, and I fully expect that there will be good reasons to support it in a flat address space.
1. Definition of "leak" -- "unauthorized (especially deliberate) disclosure of confidential information".
2a. If this really was a "leak", knowing Microsoft, you can bet money it was not "unauthorized", it is/was just part of the PR/Marketing plan. As a result, it was a calculated, deliberate and authorized disclosure and not a "leak" after all.
2b. If it is a hoax, then of course "leak" makes more sense. And, on top of that, if it is a hoax, I would bet money that Microsoft had a hand in it, just to get the Microsoft and Windows names bandied about for a group ego stroking.
Of course, I might also be a tad bit cynical.
Yeah for now. I remember my first PC that had a 520 mb hard drive and we thought there is now way a home pc should need more HD than that. Now we are 320GB and higher drives.
The original IBM System 38 and its descendants, such as OS/400, OS/500, etc., had a 128-bit address space. In these architectures, the large number of address bits were used to provide an address space that spanned both memory and disks and was used to provide processor-level protection for objects stored there. Using large address spaces to ensure hardware protection of system objects is a good start on a highly secure OS and is probably where this is going.
Sounds like classic IBM overkill, seeing as at the time you could have memory mapped every disk on earth into a single 64-bit address space.
Outside of normal page table protection, what does having extra address bits provide with regard to protecting hardware objects?
The enemies of Democracy are
This has been done before, afair IBM has used a 128 (or 129!) bit address space for their AS400 platform, where everything is memory mapped.
Well, with 129! bit address space, I imagine you could individually address every electron in the universe.
You heard it here first, folks: 64-bit ought to be enough for anybody.
For a while, for a while!
I'm not falling into the Bill Gates trap. Though I will predict this: 64 bits will be enough for everyone for a hell of a lot longer than 640k RAM was enough for everyone! Seeing as it basically never was...
The enemies of Democracy are
Itanium is not unsuccessful for VMS machines (you cannot put VMS on an x86 based chip, 64bit or no), and VMS is used in mainframe and other ultra-high availability applications. The Itanium just didn't pan out for any sort of windows-based operating systems, because windows is so tied to its x86 legacy.
I believe they also have a successor that will be compatible with Itanium as well, I'm not sure though. I mainly only looked at Itanium from the VMS point of view. They certainly have a future their though, their only competitor is the Alpha by HP, and these tend to be very very very expensive applications they are used for.
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
No, IBM never produced an "OS 500". The branding went from OS/400 to i5/OS to today's "IBM i".
No, the system never had a 128-bit address space. The address space of OS400 went from 48-bit to 64-bit when IBM started using 64-bit Power-based processors in those systems.
Yes, the instruction set uses 128-bit pointers, but only the rightmost 64 bits of the pointer are used in the current system.
Yes, The 64-bit address space covers both system memory and disk storage.
This Wikipedia article about IBM System i is a pretty good reference about this kind of stuff.
And no one ever "accidentally" leaves that sort of information on LinkedIn. This isn't like Facebook where you mistakenly think only your boyfriend will see the pictures...
Ephebobyte
Extrabyte
Extremebyte
"When information is power, privacy is freedom" - Jah-Wren Ryel
You could be right, but maybe there is more to it. Nobody thought two digit year dates would be a problem either or saw a need to look forward (maybe some did individually but not as an industry). I can't see why expanding architecture would be a bad thing especially if it could smooth transitions to newer ways of computing.
Don't think of it as a flame, more like an argument that does 3d6 fire damage.
Maybe writable control store is coming back.
It all depends upon how you defined "need". He should have just been honest and said "no one really needs a computer".
with 5-6 billion US dollars you too can be dating a much hotter girl then you could usually get. Make it $40+ billion and she'll have your children.
Money talks and does make a lot of things easier. This is sad but true.
You didn't make it to my last paragraph, did you?
The enemies of Democracy are
Is this an idea so that they could compile code to a RISC like instruction set running on a GPU like architecture with many simple parallel CPU like cores? GPUs have WIDE address buses, and this is logically an extension of NVidia's CUDA.
Thats may be what is going on. Given NVidia's positioning in business wih regard to Intel and ATI/AMD, maybe this is their ace os spades play with MS.
You're probably right. :-)
However, 256 bits is a _lot_ of bits:
The number of sub-atomic particles in the known universe has been estimated at around 1e80 (give or take a few orders of magnitude).
If we convert this number to binary, we get around 260+ bits, let's call it 256 for a nice, round number.
I.e. 256 bits is enough to address every single particle in all observable galaxies.
Terje
"almost all programming can be viewed as an exercise in caching"
He said "in 20 years" and then made the mistake of thinking that is 2020 when he meant 2030.
No, I didn't make that mistake.
I assumed that we can't possibly need 128-bit addressing until 20 years from now (about 2030), and so we don't really even need to start thinking about coding for it (or seeing hardware that thinks about accomodating it) until about 10 years from now.
Intel may support it, but there are few software vendors that do as far as I know. Intel maybe targeting it for HPC, but just try and buy a Mathlab license for this architecture.
Yes 64-bit works smooth as silk for Windows 7 as long as you mean "just Windows 7". Very few user applications are compiled 64-bit. In fact a few old installers don't handle installing 32-bit applications in a 64-bit system correctly. How many utility applications like AV are really 64-bit instead of "64-bit compatible"? Very few. In fact I won't be surprised when it turns out a few more (nasty) surprises pop up with side by side and 32-bit and 64-bit versions.
So far, everything is going swimmingly because very few pieces of software are Win64. Now lets see what happens when things like Photoshop or Word or Firefox go 64-bit and their legacy plug-ins break.
I am still waiting for a version of Windows 8 that I can run on my 4004 CPU with a 4bit storage bus....
Tsukasa: All I really want, is to be left alone...
Unless I'm experiencing some kind of time travel anomaly, I haven't heard of any 128-bit CPUs for micros even being discussed, let alone prototyped. in fact, the only 128-bit CPU I can find any mention of is in IBM's System/370. Anyone have any info about 128-bit microprocessors. And I don't mean vector processing extensions, I mean 128-bit wide general purpose registers, 128-bit wide bus, 128 address lines, etc. a REAL 128-bit CPU. None of this 8088-style crap either: 128-bit general purpose register with 64-bit bus.
cat
GPUs have wide data buses, not address buses. Data buses 128 bits and wider have been on mainstream cpus for a while now, though GPUs still have an advantage there.
The enemies of Democracy are
Yes, I'd assume that is basic knowledge for anyone who'd be posting a reply to this :P
So that's what happened to Bob - he kind of appeared and disappeared early 2000, big round face with a smile, telling people how to use their computer and trying to comfort the technologically challenged...
Will it do Linux?
Microsoft does have a tendency to pre-announce "blue sky" projects that never see the light of day. Remember the promised Cairo OS that never came into being? Remember the promised WinFS that never seems to be released? Remember the promises to get rid of the registry? Remember the original promises of release dates for the OS that would eventually become Vista? I could go on and on and on...
If it looks like vaporware and it sounds like vaporware then it probably is vaporware.
This and no other is the root from which a tyrant springs; when first he appears as a protector - Plato (423 to 327 BC)
Yes, but MSFT is not saying that people will need 2^128 bytes of RAM. All it would mean is that 2^64 won't be enough.
Now 2^64 is still pretty large for RAM. But I am guessing that MSFT is thinking that as more hard-disks become solid state, the difference between RAM and HDD will vanish and that file system might be directly addressed by CPU.
http://slashdot.org/submission/1062723/Cheap-mobile-data-plan?art_pos=2
Knowing the history of teh bits, this simply means Windows 8 will be available in both 32-bit and 64-bit versions, and 128-bit processors will be able to run in 32-bit mode, but not 64-bit mode.
WTF? If the history of the bits indicate anything, they run the previous generation architecture. For example, 64 bit processors support 32bit mode and code but 16bit which can only be run via a CPU emulator software.
This space for rent.
Right, because the first thing you do when hired into a high-security department is tell everyone on the Internet what you're working on. Sounds like this guy was a dunce right off the bat.
This ranks right up there with twittering about how much you hate your boss.
Actually, the correct answer is NT3.5, and that was in the last century. Rock solid. Then they f'd it up.
Is it just my observation, or are there way too many stupid people in the world?
I know it is more fun to speculate on what he MIGHT mean, but maybe this guy was just making up shit to get potential employers' attention... If they ask about any details, he could always hide behind the NDA he signed when working at MS... Sounds like a perfect plan.
I worked on developing synergies between IBM and Google to better facilitate the upcoming merger... When it never happens, whoops? it must have fallen through.
http://www.tomandemily.com
Huh, HAHAHA. Outside contractors HeHEhe. So Captain Morgan must be a contractor for Microsoft. Now I get it.
Robert Morgan doesn't work for research. Check this link: http://www.winfreddekreij.com/index.php?view=article&id=92:re-microsoft-leaks-details-of-128-bit-windows-8
A tadpole is a pollywog
How many times have we been stuck in a jam because people thought "this will never be in use by the time this is a problem"? The Y2k bug, which we were afraid might bring down the entire banking system with the change of a clock, was one major one. Before that was the infamous "640k is enough for anyone" shortsightedness. There has been plenty of it in the x86 architecture that needs to be babied as it has been upgraded from 8-bit to 16-bit to 32-bit to 64-bit.
Y2K ranks up with Swine and Bird flu for being some of the most hyped "disasters", yes Y2K did happen but on only a few things. A far cry from the "planes will be falling out of the sky" garbage the media was spewing out.
The 640K is enough (aside from not being a real quote) does make since though for static machines. Point of sale machines with DOS-Based software don't need gigabytes of RAM and terabytes of HD space. For a time computers were mostly static much more akin to game consoles, there wasn't a wealth of software to do stuff with. If you take an old computer and boot it up, it is remarkably fast even though it might only have a 25 Mhz processor but it was kept in relatively unchanging conditions software-wise so it doesn't get programs not made for it. On the other hand we have Pentium 4s running software designed for use on more modern CPUs and it just fails. Similarly we run programs that are made for lots and lots of RAM and HD space on machines with only a few hundred megs of RAM and a few gigabytes of HD space and they run slow. If it wasn't for the internet and the ability to get new software on old machines via it, chances are there would be fewer "slow computers".
They have had to work around it over and over and over and over again, every time the technology jumps. Now that they are actually thinking "hey, what about the next big jump?" and structuring their systems in such a way that they potentially won't have this problem, and you seem to have a problem with it.
But the thing is, technology -hasn't- jumped past 32 bit. You can still find many, many, many computers with 32-bit only CPUs still being used. And on just about every single 64 bit capable platform it is running a 32 bit OS. It makes no sense to try to go for the "next" thing that is still 2 generations ahead.
Taxation is legalized theft, no more, no less.
Although I didn't read the article (This is Slashdot after all!) I would bet that they are not referring to addressing lines but to data lines. It would allow data to be taken in at larger gulps. I seriously doubt that computers in the foreseeable future would need 128 bit addressing. Hell even 64 bit addressing is unlikely to be exhausted!
My ignorance. I've only looked at the netbook ones. Wasn't aware that there were 64 bit Atom processors, actually. :)
Yeah for now. I remember my first PC that had a 520 mb hard drive
My first computer's storage was a 60 minutes cassette tape recorded at a little less than 1000 bauds. That was 25 years ago. 128 bits (16 bytes) were 1/64th of the total RAM of that machine. The world is getting better quickly.
Ah, yeah, that's totally true, considering we don't have legacy apps from win2000 and win98 any more, it's all good.
Oh wait, we DO have legacy apps from 98 and 2000, in fact we've still got legacy DOS apps (i.e. well past your 10 year prep window)! XP is what, 7 years old now? And yet there a number of custom applications that could not be upgraded to a 64-bit based system.
Maybe, and I'm just sort of throwing this out there, but maybe if we start thinking about 128-bit compatibility NOW, there won't BE any legacy system issues when a 128-bit OS rolls around. Especially if you can avoid a hack job later, I don't see what the problem with thinking ahead is.
If you think there are no issues with our current 64-bit consumer grade hardware, consider the fact that it is nearly impossible to port a pure 64-bit OS to it. Check out OpenVMS, which has been in the 64-bit OS game for a decade now, and you'll learn you need true 64-bit hardware to run it, and Intel and AMD's consumer grade processors don't cut the mustard, and likely never will.
Why NOT try to avoid these problems now, instead of later when it may be too difficult? You know the old saying that an ounce of prevention beats a pound of cure, well it's just as true for software as it is for medicine.
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Mark Shuttleworth announces "Fuck Everything, We're Doing 256 Bits"
The Windows NT architecture has always been multi platform, from day 1. There has never been a time when it wasn't built for at least two platforms, even if one of them was obviously not commercially viable (like Itanic, or Alpha towards the end).
Nothing to do with security, just business: an insistance that the OS not ever be at the mercy of any one processor vendor.
Socialism: a lie told by totalitarians and believed by fools.
When you say accessed, do you mean accessed or accessed quickly? Most machines in this class provide distributed shared memory between nodes, where the nodes use some of the unused bits in pointers to indicate remote pages and fetch (cached copies of) them in response to page faults, implementing a cache coherency protocol and hiding all of the details of this from the running code. If your working set is mainly in your own node's RAM, then this can be quite fast and convenient. If that's the case, then it really does require 60-bit pointers, and one made for the same price in two-three years will probably need more than 64-bit pointers.
I am TheRaven on Soylent News
Yes but there is a BIG difference between 8 bit vs. 32 bit and 64 bit vs. 128 bit address space! We can visualize numbers that are 8 bit (approx. 1,000,000) and 32 bit (approx. 4,000,000,000) numbers. We need analogies like the number or grains of sand on earth or the number of subatomic particles in the universe to visualize how large a 64 bit or 128 bit number is. 128 bit addressing is just wasteful. Don't even get started on estimating the size of a 256 bit number!
(I HOPE Windows 8 is 64/128, and 128-bit processors are 32/64/128, but I know better than to expect anything sensible).
Who cares? 64 bit gives you exabytes of ram and it'll be a while before this is in any way limiting.
"We returned the General to El Salvador, or maybe Guatemala, it's difficult to tell from 10,000 feet"
Look at supercomputing architectures - in the past many systems were based on message passing between individual CPU's or nodes. That requires custom communication software/hardware to handle data transfers. Alternatively, you can memory map the memory space of each node into a 'global' memory map. Then you can write scalable software that works from a single CPU to thousands of CPU's. A 32-bit memory address gives you 4 Gigabytes, but if you want a single memory map for a system with thousands of nodes, then you need another 16-bits. Add some more bits for multicasting and broadcasting, and you get 128 bits.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
There are, however, some advantages to having a sparse address space. You could, for example, use the high bits as an object ID and the low bits as an offset from the object, and then not map anything into the space after the object, giving you automatic (hardware-enforced) bounds checking. Of course, you're effectively then just reinventing segments...
I am TheRaven on Soylent News
That doesn't make sense. A 64-bit address space has only 100x less bytes than there are grains of sand on the planet. We won't be anywhere near exhausting that for many decades.
"Politicians and diapers must be changed often, and for the same reason."
Why? Everything will obviously be virtualized. I run all my old dos apps on an emulator now, and the current generation of Windows (7/2008.2) addresses backwards compatibility through virtualization as well. I would bet that existing Windows XP apps will run on all future Microsoft OSs without any (new) problems or need for porting (simply by running in a virtual XP session).
And WTF is that FUD with "consumer grade processors" not being "true 64-bit hardware"? What on earth are you babbeling about? Some loser can't port OpenVMS to a modern platform and is blaming the platform?
Socialism: a lie told by totalitarians and believed by fools.
To be fair, I was in the opposite position, having only looked previously at the 230/330 thinking about using one of those for an inexpensive Linux server at home; I wasn't aware that the N/Z models weren't 64-bit, too. I started writing GP with "Actually, the Intel Atom processors are 64-bit processors", but then decided to do some checking before I posted it.
Good luck to you Mr Morga. You're about to have yourself one hell of day!
Right, we need to add one bit to architecture sizes every year or two. Therefore, the transition from 32 to 64 bits out to keep us happy for 30-50 years. Not something Windows 8 will need to care about (if MS gets that bad at shipping software, they'll be out of business).
Exponential growth of storage is likely to break before then, however. There is some resolution which exceeds the ability of any human to discriminate. We will eventually get High Enough Def. An individual just won't have any use for more than a few hundred years of playtime of High Enough Def media - sure, you might collect more than you could watch if you spent your entire life doing so, but not a lot more. When hard drives that size become available, demand will quickly fade for more storage.
64 bits of addressible memory may actually be enough for a personal computer. A computer with 128 bits of addressible storage would likely be impractical (the earth has about 2^166 or so atoms).
Socialism: a lie told by totalitarians and believed by fools.
I'm afraid you are wrong. Doubling the number of bits doesn't double the "power" it multiplies it by quite a large number.
For instance a 32 bit time_t allows you to express time stamps from 1970 to about 2037 - 67 years. Changing the time_t to 64 bits doesn't give you another 67 years, it gives you another 4 billion times 67 years. That's much longer than the age of the Universe so far. Why do you think we'll ever need a 128 bit time_t?
How much RAM can you address with 64 bits of addressing? That's 2^64 bytes you can address. Each byte has 8 bits, so that's 2^67 bits. That's about 10^20 bits. How many silicon atoms do we need to implement a bit of memory? I'm going to say 1,000 (probably a massive under estimate) or 10^3, so a 64 bit address space would take 10^23 silicon atoms. A mole of silicon (6x10^23 atoms) has a mass of 28 grammes, so our 64 bit address space would take between 4 and 5 grammes of silicon.
How much silicon would you therefore need for a 128 bit address? 8 grammes? No, 2^64 x 4 grammes or about 10^20 x 4 grammes - 4 x 10^14 tonnes.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe
When you say accessed, do you mean accessed or accessed quickly?
As far as I know, there's no direct way to share memory across nodes. Of course, strictly speaking any node can access any memory location in the computer using a combination of MPI and local memory access, but that's no longer a simple pointer dereference, and isn't related to the size of a pointer.
If you're curious about the details of this computer, IBM's documentation is here: http://www.redbooks.ibm.com/redbooks/pdfs/sg247287.pdf
Do you understand the SIZE of the numbers we're talking about here?
2^128 == 3.4e38
Number of atoms on Earth is ~1.33e50
Suppose we're true bad-asses, and we create a RAM that can store 8 bits PER ATOM. Wow. We would need 3.4e38 atoms to store a 128-bit address space. Considering that as a fraction of Earth's mass, this RAM will have a mass of about 15 trillion kilograms.
I don't know about you but I am not sure where the fuck I would store a stick of RAM that weighs 15 trillion kilograms.
I know it's cute and everything to hearken back to the "Nobody would ever need more than 640k" days but get a brain and run the numbers, ok?
So if we had a memory leak it would bleed transistors.
In order to form an immaculate member of a flock of sheep one must, above all, be a sheep.
Well, kudos for checking more than I did!
What you decide on for the inexpensive Linux server at home? I was recently looking at doing something like that and ended up using my old Dell laptop (E1505), which has an original core duo (*not* 64 bit, either). Most power I've seen it eat is 65w, so it seems to be decent enough in that respect... and I was only looking for something to play music with/store files/run a web server, nothing big that would require a case and 20 hard drives.
After all, there are a few servers in the world that handle truly epic amounts of data, and really might be able to use more than 64-bits.
No, really there aren't. You don't understand how big 2^64 bytes really is.
If I had a Firewire 800 disk with 2^64 bytes capacity, it would take me 5,000 years to fill it if I was writing continuously at maximum speed.
I note that the DDR3 RAM in my laptop can transfer 64x10^9 bytes per second. So using the main memory bus of my laptop, it would only take 10 years.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe
All of those API calls that expected 64 bits will break as their 64 bit variables will be overflowed with larger numbers. Software will need to be modified and recompiled or run in yet another XPM XP Virtual Machine.
So far Windows 7 and the XP Virtual Machine uses SP3 and has limited 3D Graphics abilities, so Games will have a hard time playing old games like Warlords IV on it. Warlords IV and other games don't run on Vista or Windows 7, I am guessing because of all of the API changes and addressing changes.
In the 64 bit operating systems the 16 bit MS-DOS and 16 bit Windows software cannot run, you'd need the 32 bit version or use a virtual machine for that. Windows 7 XPM XP Mode only works with certain VT features and won't run on all systems, so some systems won't have a Virtual machine unless they run VMWare, VirtualBox, or something else like QEMU or BOCHS.
I am guessing 128 bit addressing on a 64 bit processor is like doing bank switching on the old 6502 and 8088 processors to access more RAM. It seems with each new release of Windows, more and more RAM is needed to run it faster, and Windows 8 is going to be a big memory hog that needs 128 Bit addressing to access a huge amount of RAM. Am I right?
Remember, Slashdot does not have a -1 disagree moderation, and no, troll, flamebait, and overrated are not substitutes.
That's enough to address every atom in the observable Universe.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe
Exactly! Everytime it's time for a new Windows version to be developped Microsoft sais: Yeah we promise X and Y and game changing shit! Then it gets delayed. Then again. Then these so called worked on features (which are not worked on and were never planned!) are scrapped. Then we get the previous Windows with a few minor changes and additions.
It happens everytime! And the worst part is; everybody keeps falling for it! There are not even 128bit CPU's!
Come on everybody, wake up! Read this FFS and NEVER, EVER, post shit like this again!: http://www.roughlydrafted.com/RD/Q4.06/4E2A8848-5738-45B1-A659-AD7473899D7D.html
Here be signatures
That sounds like a myth to me.
Can you give me a link to where he said that? I did a quick Google and I found a Wired article on it saying that he never said it.
http://www.wired.com/politics/law/news/1997/01/1484
It also claims that Bill Gates denies ever saying it - "Meanwhile, I keep bumping into that silly quotation attributed to me that says 640K of memory is enough. There's never a citation; the quotation just floats like a rumor, repeated again and again"
I was struck by (not severely, but enough to delay non-essentials) reduced income which has delayed that decision; I'm still leaning toward one of the "nettop" oriented Atoms (by the time I get around to it now, probably one of the new D-series ones, since I understand the 230/330 are goint to be phased out when they are introduced).
That sounds like a good choice: I'd rather keep something existing and repurpose it, but unfortunately my only operational but not otherwise committed computer is a a fairly old laptop that has some issues that make unattractive for use in a situation where it would be on for long periods of time. So I am kind of left with building fresh.
You are a fool.
There doesn't have to be 2^128 mem, there just needs to be a desire for 2^64+1.
Don't let that stop you from talking shit though.
Okay, but the question is what does that mean? If it just means 128-bit operations or registers, then that's been around since the original SSE. If it means 128-bit addressing (like it usually does), then who the fuck is making those chips and why?
Perhaps MS would rather spend some resources making their OS ready NOW and be/stay on top of it, instead of accumulating another 10 years of additional cruft first and THEN try to shoehorn everything in there all at once? The more legacy dependencies they have to deal with, the harder it will be, after all.
Consider Apple -- they had Intel ports of OS X ready for years internally, which would be irrelevant to their (at the time) PPC offerings... But when the market outlook changed and they wanted to switch architectures, it meant that they were prepared and ready for it.
I'm sure Microsoft isn't planning on becoming obsolete any time soon...
The fact that it's been handled like trash and we will be forever tied to 32-bit leads me to believe the next architecture will be 128-bit with support for the 32-bit instruction set, with most 64-bit instructions being replaced by 128-bit ones, and the 64-bit instruction set as a whole being dropped.
And Java and .NET have proved that we don't need a brain dead hardware protection either. NX bit? You can still corrupt the data, and the data is what the instructions are taking as state and input. NX is a stop gap measure that makes your computer marginally more safe.
And now we are going to use a 128 bit address space to prevent attacks? Really? I'd say the disadvantages clearly outweigh the advantages.
Yes, it was a joke.
What's one computer? That's hardly convincing. By your own admission, 5000 servers in a datacenter run by the likes of Google or Microsoft could consume 2^64 Bytes in a year (or a few years, more realistically). See, your mistake is that you are thinking in terms of one computer, but some of the very high end systems in the world are clusters of *thousands* of computers. Granted, other than the virtual filesystem they share on some sort of SAN, which really could arguably need to be 128 bits, I suppose mostly those thousand+ computer clusters are running 64-bit OS's and they each have a relatively small 'chunk' of the data to work with.
But, maybe someone wants to come up with an operating system that shares a single address space among thousands of computers?
Like I said at the end of my post... it makes sense for MS to makes sure Windows 9 is "compatible" with 128-bit addressing. The part that surprises me is that the "partners" listed have 128-bit anywhere on their roadmaps when they aren't even bothering to implement a full 64-bit address space because there's not a machine on earth that's even close to needing it yet.
The enemies of Democracy are
With Windows taking up more resources exponentially with each new version, 128 bit was the natural next step.
I am the unwilling control for my Origin.
It's only one more bit though.
128bits is twice as many bits as 64 bits.
It just also happens to be a bajillion times more addressing space.
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Bzzzt. My lovely wife would be insulted by your comments. Melinda Gates is quite good-looking but my wife is prettier, and I didn't 10 figures in the bank.
You are in a maze of twisty little passages, all alike.
Why not focus on fixing all flaws in existing Windows??
There's a youtube video with Mark Russinovich where the
they ask why Microsoft dont re-write Windows the correct
way with known proper security from the ground up ?
Russinovich replies, "it's too much work"
So there we have it folks, Microsoft CHOOSES ON PURPOSE to
continue on a flawed and insecure and slow bloated systems.
I'm still confused.
What's the point of having 128 bit compatibility? 128 bit CPUs don't even exist yet. Heck most of us are still just using 32, and haven't even visited the 64 generation yet.
Just a hypothesis, but maybe perhaps this would open the door to a migration away from x86 and x86_64 to something entirely new, without it turning into another Itanium. I couldn't in good conscience formulate an opinion on x86 but I have read many people time and time again saying that it was a bunch of shit legacy and bad design decisions lugged around in the name of compatibility. I'm not educated enough in that area to agree or disagree, but if many people voiced that concern so far, maybe there's something there.
I mean, without necessarily ditching backwards compatibility (which *is* being slowly ditched anyways, if you see how many backwards compatibility subsystems are being dropped from each new release of windows. Also, Windows Server 2008 R2 doesn't even offer a 32 bits version. At all). Maybe these new hypothetical processors would enable for dynamic recompilation or just fast enough emulation, like Apple did when they switched from motorola 64k to PowerPC, and how they just did again from PowerPC to Intel with rosetta.
Again, I have no clue what I'm talking about, it's just an hypothesis.
I guess we didn't learn much from the mp/m days and the apple language card, because when memory ran short in the PC, we resisted protected mode and punted with the expanded memory. Sometimes I wonder if the same people were involved. Now using virtual memory, we are going to swap pages in the 4GB space? I am sorry we cannot seem to do things the straightforward elegant way.
If you are going to have a Sagan of memory in your machine, a longer address word is required.
If we go to a new version of IP with a larger source and destination fields (larger than four octets), router memory will be challenged once again. In fact this maybe slowing the adoption of six right now. Just think of the size of the routing tables in a world where every widget has it's own IP address. Just processing routing tables that large may be the application that pushed the clock rate up further for CPUs.
If a spaceship wants to travel out there and count all the stars and planets, it will need s a lot of memory for that. I think Windows 9 is for intergalactic space travel. Lets send Bill on the first ship with the telephone sanitizers.
I just realized that if Microsoft ports Windows into a 128-bit architecture, their errors per amount of ram will go way down. With an unimaginable amount of ram in the machine, they can continue to pile new code on top of old code until the cows come home. They can take their current operating system objects and derive yet another set of objects from them, and not even care about the footprint of software in memory. Of course hard disk speed will have to improve or loading a Windows kernel 32 times larger will take 32 times longer. This could really be a driver for technology.
You are a fool.
Um... ouch?
There doesn't have to be 2^128 mem, there just needs to be a desire for 2^64+1.
Okay, I'll bite.
Doubling the size of all or nearly all 64-bit buses and registers in a processor would vastly increase the transistor count and complexity, depressing performance and increasing cost for little gain ("one extra addressable byte" in the extremely marginal case that you suggest).
More likely, once we reach the need for a 65-80 bit bus, we'll see some kind of bank-switching solution for a few years. This was used in old 8-bit computers with 16-bit address buses that needed more than 64k RAM, the segmented memory of 8086-80286 processors is a variant of this, and Physical Address Extension allowed 32-bit Pentium Pro through Pentium IV processors to address >2^32 memory locations using similar techniques.
It's not a pretty solution, but it does reduce the transistor cost of enlarging the addressable memory while Moore's Law catches up. It's been done time and time again, and I expect we'll see it again once 64-bit address buses are no longer sufficient. Which is still quite a ways off (e.g. AMD64 processors are hardware-limited to 40-bit physical address space with the possibility to extend to 52-bits).
Don't let that stop you from talking shit though.
I wasn't talking shit or trashing the concept. I was thinking about the relative uselessness of a homogeneously 128-bit processor in the near future.
I've seen estimates that address space requirements increase by about 0.8-1 bits per year. So 20 years ago, 20 bits (1 MiB) was about enough for a "typical" desktop user, while today 32 bits (4 GiB) is about enough for the same. By this admittedly rough estimate, it will be at least a couple of decades until typical desktop applications want 64 bits of physical address space. Power users and servers are perhaps 2-4 bits ahead of the curve... I don't think I know of any commercially-available systems with more than 64 GiB of DRAM directly addressable (that'd be 36 bits of address space).
My bicyles
While obviously unnecessery from an addressing point of view it does offer some interesting advantages in terms of security. In particular rather than bothering with expensive hardware memory protection one can simply rely on the extreme sparseness of memory usage to provide process isolation...at least if you do other things right (randomize layouts etc..)
If you liked this thought maybe you would find my blog nice too:
How quickly we forget!
It was C that had that problem, not C++. At least not in the language we call C++ today, which has the std::string class specifically for that purpose.
And for that matter, near and far pointers could be handled in much the same way as C++ smart pointer classes. We didn't have that technology in 8086's heyday, which made near and far pointers a headache. But now we do.
Pet peeve: Profane people propagating perfunctory pedantry.
Well, I have heard about 128 bit computing since I was a wee bairn and 16 bits was something you only had in your wet dreams; it has somehow failed to materialise in the market. It is probably because it is quite hard to imagine that the benefits would justfy the not insignificant complexity; even now, when we talk about "64 bit" architecture, when you read it carefully, it isn't actually quite 64 bits, as in "64 bit instruction and data buses", so I imagine this development is still some way off.
Also, we have seen that 64 bit programs don't normally run much faster than 32 bit ones.
Itanium is not unsuccessful for VMS machines (you cannot put VMS on an x86 based chip, 64bit or no)
Since the Alpha processor has been discontinued, Itanium is pretty much the only choise left for VMS users. HP will still support Alpha machines for several years, but you can't buy them anymore. All VMS installations will gradually migrate to Itanium as the servers are replaced.
I believe they [Windows] also have a successor that will be compatible with Itanium as well, I'm not sure though.
There are Windows Server versions for Itanium, but they're not very popular AFAIK. Itanium only offers a big performance gain over x86/x64 systems, at a substantially higher price.
WWTTD?
You know a lot about PAE/PSE and CPUs/RAM.
But not a lot about the limitations of XP/Vista. Unlike some operating systems(like Linux), XP and Vista actually have caps where they ignore memory beyond a certain point.
You can use 4MB paging, but you're still only going to have 4GB of addressing space. You can enable PAE for XP Pro, but again you'll be stuck with 4GB of addressing space, and max 4GB of RAM.
What's funny is, XP SP0 and Win2k didn't have the same limitations. People successfully ran 4-8GB of RAM w/ PAE before the XP service packs came, but afterwards they got limited to less. And now with Vista they've "improved" System Properties to show you your max RAM rather than max usable RAM, so Vista 32bit will inform you you're using 8GB when only 3.2GB (or less) is available.
Conclusion: Microsoft probably wants you to upgrade to 64bit. They also don't care about lying.
He joined Microsoft in 2002, is he to blame for the Vista debacle?
From Windows IT Pro: "All your rumors are belong to me: Windows 8 to be 128-bit? No. Good God, no. People can be so silly sometimes. Writers at PC World, Ars Technica, Slashdot and many other publications fell for an obviously faked LinkedIn profile from a supposed Microsoft researcher who claimed he was working on a 128-bit kernel for Windows 8. There's just one problem. This guy doesn't exist. No one with his name has ever worked at Microsoft Research. His job title is fake. Microsoft is not working on a 128-bit kernel for Windows 8. And, best of all, the guy's listed university? It's an "online supplier of academic degrees," according to Wikipedia. OK, that's five problems, or four more than those geniuses on the web should have needed to figure out this rumor was fake. Seriously, you guys make me laugh so hard sometimes I could cry. It's just sad." http://windowsitpro.com/Articles/Index.cfm?ArticleID=102939&feed=rss&subj=0
Reliable sources informs us that Duke Nukem forever will be released in both 32, 64 and 128 bit versions.
without the hardware to support it
So you think 128 bit processors aren't coming? Where have you been in the past, oh, 40 years?
No, I don't, not any time soon anyhow. Windows 8 will probably be out sometime in 2011 and that's way too early. Perhaps Windows 9 but even then it's hard to see the need.
You know, there isn't some kind of inevitability to the development of hardware. If the past were any guide to hardware development we'd all be using computers with a single core running at 10 ghz.
It has to solve a real problem and solving that problem needs to benefit someone enough to spend the billions of dollars necessary to develop the hardware.
However if you have any evidence whatsoever (besides "look at the past"), then I'd love to see it.
Actually, x86-64 is doing something like that.
The architecture supports 64-bit addressing, but existing x86-64 CPUs only have a 48-bit virtual address space, and a 40-bit physical address space.
By the time that becomes an issue, of course, the implementation will change, and nobody will notice, because nobody will think to shove more than 1 TiB RAM in an old Core i7 box. (And, for that matter, memory controller limitations mean that you're not getting anywhere near that much in there anyway - my main machine has a chipset that doesn't even support any more than 32-bit addressing, for that matter, despite supporting some EM64T CPUs. That's been true for quite a while, of course - I don't know of a single i386 that can actually run 4 GiB RAM, yet they all (well, all the DXes) can address it.)
wow. a meme on /. thats new.
I reserve the right to have a physical object so I can sell it later, and recover my money.
Not all of us..
Some people used architectures other than x86 which didn't need such nasty hacks.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Not only that but the 10 figures I thought I had suddenly.
You are in a maze of twisty little passages, all alike.
Bollocks. If you want to do it with RAM alone, you need 4 billion machines with 4GB of RAM to reach 64 bits of RAM address space. Yeah, that sounds pretty high, but that's just RAM. If you want to consider disk space, you only need ~16 million terabyte drives to reach the limits of 64 bits of address space. I imagine there's at least that many out there right now. Are these all hooked to one machine? No. But they're probably all hooked to the Internet. If you wanted to globally address all the bytes out there today, you'd need a damn sight more than 64 bits.
And that's today. Storage trends are such that storage doubles fairly regularly, so that we end up burning address bits at a pretty steady rate. In fact, that rate appears to be just slightly faster than a bit per year.
Therefore, I think a 128 bit Windows 8 is ridiculous (and as others have pointed out, most likely a hoax). On that point we agree. At the point 128 bits becomes relevant, the notion of a monolithic OS on a distinct computer may well be obsolete. We'll see though.
I think that we'll see 64 bits as the limit on individual machines for quite awhile--another 20 years at least. In the x86 world, the jump from 16 to 32 took around 7 years whereas the jump from 32 bit to 64 bit took around 15. (8086 came out in 1978, i386 came out in 1985, and x86-64 came out in 2000.) If we burn address bits at a roughly constant rate as these last two jumps did, the jump from 64 to 128 bits will happen around 30 years after x86-64 came out, around 2030. If we hit fundamental technology limits, or simply reach a point of diminishing returns on greater storage (which may well happen unless we just start recording our entire lives in Quad-HD or something), we may get bored before we reach the limits of 64 bits.
Program Intellivision!
Windows is a 32-bit extension to a 16-bit graphical shell for an 8-bit operating system originally coded for a 4-bit microprocessor by a 2-bit company that can't stand 1 bit of competition.
I'm a kernel guy, I work with a lot of kernel guys, and pretty much everybody I talk to just laughs at the idea of a 128-bit "architecture". Here's some back-of-the-envelope math: 2^64 addressable bytes / 4 GHz / 1024 cores = 6 months to even use that much memory. So unless you have a weird need to map the same memory thousands of times, a 64-bit architecture is going to last quite a while.
A witty [sig] proves nothing. --Voltaire
Windows NT has always been cross platform. NT 4 ran on MIPS, Alpha, PPC and i386. No the apps on the other hand, is a problem. but Windows? It is designed from the ground up to be ported to anything. The Itanium didn't pan out for Windows because of slow i386 emulation for APPLICATIONS, not the OS, as well as poor performance in early iterations. Good research though.
In all fairness, life before 1943 completely sucked. I'm not saying that there's a correlation between available address space and the general suckiness of life, but I'm not saying that there isn't a correlation, either.
Outside of normal page table protection, what does having extra address bits provide with regard to protecting hardware objects?
Capability based security at the hardware level.
The larger performance issue is that once PAE or PSE are involved, the kernel and DMA can no longer directly address all of RAM.
Of course the benefits of having a lot more RAM can easily outweigh that drawback in practice.
All the same, 64 bit (46 physical) is much better.
I can only guess he's looking for a marketing deconvolution filter that makes the statement somehow believable at all.
I don't even want to THINK about the mess MS would make of an OS/400 style object system!!! Especially since object systems tend to maintain state rather than shut down and re-init on boot.
The DMA issue is still there with 64-bit unless you have an IOMMU. Most of the cheap x86-64 machines still don't, so 32-bit PCI devices can't DMA anywhere above the 4GB line. Kernels work around this with bounce buffers and relocating userspace buffers below the 4GB line.
It is a problem with the kernel. Often this is worked around by using the same sort of memory split as with 32-bit, so the kernel can see all of its own memory and all of the current process's memory, but not other process's memory, when handling a system call. This limits the userspace code to 2-3GB, which is a bit more cramped. This is a big part of the reason why lots of 64-bit (non-x86) workstations run a 64-bit kernel but mainly 32-bit userspace apps, and get the best of both worlds.
I am TheRaven on Soylent News
Sounds like classic IBM overkill
Well, as the old saw goes:
An elephant is a mouse with an IBM operating system.
Do you understand the SIZE of the numbers we're talking about here?
2^128 == 3.4e38
Number of atoms on Earth is ~1.33e50
Suppose we're true bad-asses, and we create a RAM that can store 8 bits PER ATOM. Wow. We would need 3.4e38 atoms to store a 128-bit address space. Considering that as a fraction of Earth's mass, this RAM will have a mass of about 15 trillion kilograms.
I don't know about you but I am not sure where the fuck I would store a stick of RAM that weighs 15 trillion kilograms.
I know it's cute and everything to hearken back to the "Nobody would ever need more than 640k" days but get a brain and run the numbers, ok?
Except that we have no idea what computing is going to be like in 20 years. And that RAM is not the only thing that uses those address lines. That's why 32 bit memory addressing is such a disaster -- Put 4 gigs in a 32 bit system and you'll actually get about 3.5 gigs out of it, give or take, thanks to addresses being needed elsewhere.
Yes, I can't forsee using 3.4e38 bits of RAM. But then again, show me an iPhone 10, 20 years ago and I would have guessed it was technology from 2100, not 2010.
I guess what I'm saying is we can't predict what the IT teams of 2025-2050 are going to need, so why not engineer it (apparently quite literally) future proof? It's not like it's that much extra engineering.