Next version of Windows to be ​'fundamenta​lly different'

Future versions of Windows will have to be "fundamentally different" in order to take advantage of multicore processors, according to Ty Carlson of Microsoft. "You're going to see in excess of 8, 16, 64 and beyond processors on your client computer," said Carlson, director of technical strategy at Microsoft, during a panel discussion at the Future in Review conference. Windows Vista, on the other hand, is "designed to run on 1, 2, maybe 4 processors," he said, referring to the fact that quad-core processors are now available from Intel and are on the way from Advanced Micro Devices.

The problem, as has been noted on many occasions, is that loads of PC applications were programmed with serial processing in mind, meaning that the performance of those applications increased as a chip's clock speed increased. That's not how it works anymore. The chip industy has decided that multiple cores are the best way to keep increasing performance, and that means applications now have to be designed with parallel processing in mind.

YearOfTheLinuxDesktop wrote:

Future versions of Windows will have to be "fundamentally different" in order to take advantage of multicore processors, according to Ty Carlson of Microsoft. "You're going to see in excess of 8, 16, 64 and beyond processors on your client computer," said Carlson, director of technical strategy at Microsoft, during a panel discussion at the Future in Review conference. Windows Vista, on the other hand, is "designed to run on 1, 2, maybe 4 processors," he said, referring to the fact that quad-core processors are now available from Intel and are on the way from Advanced Micro Devices.

The problem, as has been noted on many occasions, is that loads of PC applications were programmed with serial processing in mind, meaning that the performance of those applications increased as a chip's clock speed increased. That's not how it works anymore. The chip industy has decided that multiple cores are the best way to keep increasing performance, and that means applications now have to be designed with parallel processing in mind.

I don't understand this and think it was taken out of context.  NT runs on 64 cpus today and could run 32 for some time.  I had a 4 way server NT box ~8 years ago.  I don't know anyone that actually buys those massive machines, but it can work.  NT can support multiple cpus and cores, and thousands of threads - that is not the issue.  The issue I think he is talking about is parallel libraries and apis that *easily allow the dev to actually use those cores and threads.  100 seperate user apps could use 1 core each on a 100 cpu machine (for example) and run fine.  The hard part is getting 1 app to leverage the same 100 cores and use them.  Big difference - that is issue.  Personally, I don't think there will be a majic bullet here.  They can make it a ton easier however with tools and libraries.

bah... i thought title said: Next version of Windows to be Fun, Demented and Different

jamie wrote:

bah... i thought title said: Next version of Windows to be Fun, Demented and Different

I, too, was misled by the title... I was hoping the next version was definitely going to be Singularity.

blowdart wrote:

I'm sure we've heard this before. Oh yes OO file systems. OO desktops. 
Look what happened last time they tried to do different,

It would be good if they finished these things too They just have to find ways to do it. There's nothing fundamentally broken with the concept of WinFS.

staceyw wrote:


I don't understand this and think it was taken out of context.  NT runs on 64 cpus today and could run 32 for some time.  I had a 4 way server NT box ~8 years ago.  I don't know anyone that actually buys those massive machines, but it can work.  NT can support multiple cpus and cores, and thousands of threads - that is not the issue.  The issue I think he is talking about is parallel libraries and apis that *easily allow the dev to actually use those cores and threads.  100 seperate user apps could use 1 core each on a 100 cpu machine (for example) and run fine.  The hard part is getting 1 app to leverage the same 100 cores and use them.  Big difference - that is issue.  Personally, I don't think there will be a majic bullet here.  They can make it a ton easier however with tools and libraries.

I think the solution Microsoft will be looking at is encouraging people to use .NET, and building .net around encouraging parallel computing.

Encouraging .NET is also a good thing also, for allowing Microsoft to make changes to the system without breaking applications, because of the virtual layer.

emet wrote:

I predict the next version of Windows will work exactly like every Windows, expect have a different skin and take a min. of 2048MB RAM to even boot. No one will know why, even Microsoft. (It will be because of the biggest innovation in the new version, WGA 3.0) Meanwhile Ubuntu and Fedora will have gone sentient.

When will fun ever stop?

Or base Singularity on top of Windows, there's no way Singularity can offer much in terms of hardware support as it is.

DigitalDud wrote:

Or base Singularity on top of Windows, there's no way Singularity can offer much in terms of hardware support as it is.

Sure it can.  HW folks just need to create managed drivers, that is the direction anyway to head off a ton of crummy drivers mucking about in the kernel.  IIRC, they working on tech to allow non-managed drivers to somehow play too within some protected bounds.

staceyw wrote:



DigitalDud wrote: Or base Singularity on top of Windows, there's no way Singularity can offer much in terms of hardware support as it is.

Sure it can.  HW folks just need to create managed drivers, that is the direction anyway to head off a ton of crummy drivers mucking about in the kernel.  IIRC, they working on tech to allow non-managed drivers to somehow play too within some protected bounds.

Right just re-write thousands of drivers, support tons of broken ACPI implementations, etc., yeah, we'll get right on that.

Personally I don't see it being that different, It's Windows after all.

Don't mean that in a bad way, I just think Microsoft are scared to do any big changes in Windows, and perhaps rightly so.

Future versions of Windows will be based on the same iterative process that has been used for past versions of Windows.  All of the  groups working on different parts of the system will incrementally work at enhancing their particular piece.  Certainly most new development will be outside the kernel in C# and based on .net.  The major addition to the UI will be the use of speech.  Over time speech will be supported at least for input in parallel with the mouse and the keyboard.  Support will be added to enable easier and more effective use of multiple threads.  The main use of this support within the UI will be to provide the processing power required by Windows to understand spoken words.

stevo_ wrote:

Personally I don't see it being that different, It's Windows after all.

Don't mean that in a bad way, I just think Microsoft are scared to do any big changes in Windows, and perhaps rightly so.

thats the big bad bugaboo.....

to *REALLY* inovate / create a different OS / get rid of some of the baggage of the legacy (DOS, Win16, COM, OLE, and such)
we would need a new OS that's clean from the ground up.

but then you have the "chicken and egg" thing of who goes first....
of how you move all the users apps and files to this new system, of how you make that move a clean process that does not scare them off.

this is also one of the things that the Li**** fans miss when they pronounce that version x.y.z of some new release will take over and such...

10,000,000,000 (or what ever number) of users will tend to stick with what they know, no mater how good "brand X" is *UNLESS* it is just so great that they see a mass move to it by dozens / hundereds of other folks they know / trust.

and for a business that wants to make money that level of risk is....
not gonna happen.

which kinda sucks [C]

but I understand it