Windows 8 will have 128-bit support

[Trocian]

BRING ON THE MOTHERFUCKING HOLODECKS. Finally, proper 3d porn is within my grasp!

^

[Kaisha]

This somewhat came to mind:
http://imgs.xkcd.com/comics/supported_features.png

At any rate, progress is progress, but it's going to be an extreme niche, even by business standards no? Pretty sure if someone needed 128-bit calculations, they wouldn't be using Windows.

Will most definitely be favoured in the CGI department though. Crunch higher precision values in one go and what not.


I'm still waiting for the next leap in computer technology. Everything we keep getting is just 'slightly' above what we already got. Nothing new that jumps out that gives a significant performance leap over existing technology that can be made affordable for the average consumer.

[Khamsin]

Eh, I dunno. 16 exabytes is more data than there is currently in every computer system on the entire planet. I don't see our storage requirements expanding that much in the next 10 years. 100? Maybe. But not ten. And other than addressable memory space, there isn't much advantage going from a 64 bit to 128 bit OS, in my opinion.

Just look at the advancement of technology in the past few decades and compare.

Technology will advance just as fast with just as significant of leaps in the next 300 years as has been done in the past 300 years. Perhaps faster, as the last 300 years seemed to pick up the pace a bit compared to advancements made from the stone age until then. So the next 300 years should be even faster, if the trend continues exponentially.

And, as technology makes things available, a "need" for it will follow. If we can do something, we'll all of a sudden find ourselves needing to do it to the maximum capability possible.

Kinda like when I upgraded from a computer with a 10 meg hard drive to 100 meg, to 1 gig to 40 gig to 300 gig, at each step I didn't think I could possibly fill up the new hard drive. But I did, within like a month.

[Jodwahh]

And lo, the porn will flow like water unto thine hard-drives, and it will be good.

[Correction]

And, as technology makes things available, a "need" for it will follow. If we can do something, we'll all of a sudden find ourselves needing to do it to the maximum capability possible.

Kinda like when I upgraded from a computer with a 10 meg hard drive to 100 meg, to 1 gig to 40 gig to 300 gig, at each step I didn't think I could possibly fill up the new hard drive. But I did, within like a month.

Then why didn't Itanium take off?

[Yeargdribble]

I remember upgrading from my 386SX with 2 MB RAM and a 60 MB HD to a hoss beast of 233 Mhz and 3 gigs of HD space. I literally remember telling people I couldn't imagine using that much space.

Now I'm sitting with 5.5 TBs on a single PC and just bought another HD. When standards change we will always find new ways to push at the edges.

Imagine if you told someone 200 years ago that doctor would be able to look inside you and see your bones and stuff without cutting you open. It just wouldn't compute for them. We'll find uses when the tech becomes prevalent and it will probably be very different from what we can try to predict to to be.

[Khamsin]

Then why didn't Itanium take off?

Same reason the betamax didn't take off, or the HD DVD. However, the advancement the technology allowed for didn't disappear. Marketing, funding, and other issues play a factor in when and how certain technologies become mainstream, but when it does people always find a way to utilize it fully. Just takes a little time for the logistic kinks to be worked out.

[Akucaen]

Then why didn't Itanium take off?

Because it was underpowered, over-hyped, and was a completely different architecture that was nothing like x86. Oh, not to mention the original hardware translation of x86 code made the top-end Itanium run about as fast as a P2 200mhz. (I don't know if the later software version in the Montecito core made any speed increases to that, they took out the hardware translation.)

The Itanium was an awesome processor in theory and on paper. The implementation was extremely expensive and it couldn't be built anywhere near consumer level. It wasn't as powerful in raw power as the processors Intel hoped to replace it with, and AMD absolutely raped them when they came out of left-field with the K8. (Providing an embarrassing lead over the P4 and ruining the IA-64 dream with a "cheap" way to get to 64-bit computing.)

The reason the Itanium failed so badly is because it largely depends on the compiler to resequence code so that it could run in the most efficient manner possible for the Itanium. EPIC is a very long instruction word architecture. It had a troubled start out of the gate since these awesome compilers didn't exist, just basic stuff. They eventually came, but not after the damage was done.

Bottom line is that it was too expensive and way too radical a change. Itanium based supercomputers are still made and sold, but they hold such a small market that you'd never know they still existed past the original Intel marketing hype.

[Zosi]

Re: Microsoft Leaks Details of 128-bit Windows 8

WinInfo Short Takes: Week of October 12, 2009 -RSS

From the second page:

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 isn't working on a 128-bit kernel for Windows 8. And, best of all, the guy's listed university is 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 that this rumor was fake.

Whole thing is a load of bullshit.

[Rockstaru]

People will always find a way to innovate. Once they know what's possible with the technology available, they'll find a way (and a reason) to make it happen.

[Akucaen]

Same reason the betamax didn't take off, or the HD DVD. However, the advancement the technology allowed for didn't disappear. Marketing, funding, and other issues play a factor in when and how certain technologies become mainstream, but when it does people always find a way to utilize it fully. Just takes a little time for the logistic kinks to be worked out.

Actually you're half right. IA-64 was first and it was definitely worth a look at for the future if Intel could've gotten the costs down and everyone to jump off of x86. You comparison is off in the sense that it's more like IA-64 was HD-DVD and x86-64 as Blu-Ray, except in order to use your HD-DVDs, you had to pray to the gods, align the CD in the tray perfectly or it wouldn't work, use a special sequence to close the tray, then enter 50,000 passwords before you could watch the anti-piracy ads, oh and you couldn't watch DVDs on your HD-DVD player.

And then x86-64, you could just pop in the Blu-Ray and your old DVDs. All x86-32 code was (in terms of the K8) either run 32-bit native or 64-bit compatibility mode. That was the original XP-x64 "flaw" in that the OS would boot the K8 into 64-bit long mode and would run 32-bit code in 64-bit mode. There wasn't a huge speed decrease, but it was still slower than if you were using a K8 in 32-bit windows.

They really shoved a marketing campaign for Itanium and put a huge amount of money behind it. It was the software side of the Itanium (go go EPIC) that caused it to fail, that and the fact that while it *could* run x86 code, it was awful at doing it. It basically ended up boiling down to if you wanted to use an Itanium system, you had to use Itanium software only.

[Zosi]

To add something to the discussion about 128-bit chips... there's more to it than just having 128-bit pointers. For example, current x86-64 chips don't even use 64-bit pointers the vast, vast majority of the time - they use 48-bit pointers so that a 64-bit word can store both a 16-bit opcode and a 48-bit immediate pointer. 48 bits gives 256 terrabytes of memory, which is a "lot", but not a "lot" compared to petabytes or exabytes. There are other benefits, like access to general 128-bit math instructions, but it's all a bit of a moot point since there isn't even a 128-bit instruction set to target with your early research work. There are no native 128-bit CPUs even in the planning phases, as far as I know.

On the subject of Itanium, I'm fairly sure that we could fill an entire thread with examples of how they fucked up the first Itaniums. The second generation was better, but they missed the boat on consumer hardware, which really limited the overall appeal of the architecture. The IA-64 instruction set is actually really decent - much better-designed than x86, but the lack of usable compatibility ensured that it didn't take off.

[Akucaen]

On the subject of Itanium, I'm fairly sure that we could fill an entire thread with examples of how they fucked up the first Itaniums. The second generation was better, but they missed the boat on consumer hardware, which really limited the overall appeal of the architecture. The IA-64 instruction set is actually really decent - much better-designed than x86, but the lack of usable compatibility ensured that it didn't take off.

The tl;dr version of what I said; well said.

[Korietsu]

Eh, I dunno. 16 exabytes is more data than there is currently in every computer system on the entire planet. I don't see our storage requirements expanding that much in the next 10 years. 100? Maybe. But not ten. And other than addressable memory space, there isn't much advantage going from a 64 bit to 128 bit OS, in my opinion.

I think the real reason why 128 bit processing is considered a holy grail amongst EE's and Comp E's is that by that point, we'll have enough room within the actual processor to do efficient parallel processing, as compared to now.

It's amazing that from the original MIPS Pipelined Processor to our Dual Core/ Quad Core i7's, that the technology has advanced in just how data is handled. We've gone from something of a 4 stage pipeline, up to the current 30(?) stage pipeline. Now what's holding us back is actually how we handle lookahead/checking and data branches in pipelines, other than the average, "Oh fuck, we can't fit enough transistors, too much electrical interference".

It's amazing how indepth the physics are behind even the smallest and most insignificant pieces of our computers are. We've even developed new fab technology because we reached a point where we couldn't see what we were making. We're almost to that point again now, with the upcoming sub 30micron designs.

[Shenrien]

all I want is Tits in holograph