PDA

View Full Version : Porting DirectX 10 onto Windows XP?


Pages : [1] 2

Digital_Trans
10-26-07, 12:43 AM
I heard around on the internet that someone was trying to port DirectX 10 onto Windows XP. What ever became of this? Could it ever happen successfully one day?

nekrosoft13
10-26-07, 12:53 AM
not gonna happen

einstein_314
10-26-07, 01:31 AM
not gonna happen
+1

LycosV
10-26-07, 01:40 AM
Never gonna happen.

JOPS
10-26-07, 01:48 AM
The Wine project will probably do it (they have just started) but they're still focusing on DX9 so it would probably take years to finish DX10 support. The GLX->WGL switch is done so the underlying infrastructure is mostly in place to support DX on Windows platforms too.

The main purpose of DX10 support in Wine is to run DX10 games on Linux but if it's fully ported to the Windows architecture XP would benefit from it too. The question is if XP or even Vista is still around by then? But some people like to play older games. ;)

Dragunov
10-26-07, 05:07 AM
If it's gonna be done, it's all with emulation, so it's gonna slow down everything again, so what benefit remains if dx10 normally is faster for ur games if u gonna slow it down again cause of an emulation layer???

jolle
10-26-07, 05:36 AM
I guess they would have to bring over the WDDM to XP first since alot of things in dX10 rely on that.
And that doesnt sound like a easy thing to do.

Shamrock
10-26-07, 06:33 AM
http://alkyproject.blogspot.com/
http://www.fallingleafsystems.com/

called the Alky Project...no it's not an emulator. He recompiles the files.

Dragunov
10-26-07, 08:14 AM
Well DX10 is always gonna run faster on Vista than XP, it's especially made for Vista....

I guess MS is gonna stop the development one day of the Alky project

ephmrl
10-26-07, 08:15 AM
http://alkyproject.blogspot.com/
http://www.fallingleafsystems.com/

called the Alky Project...no it's not an emulator. He recompiles the files.

Looks like another reverse engineering project (like WINE) + OGL wrapper. Unless they have an insider hijacking the latest dx10 source from MS, doesn't seem likely they're recompiling stuff.

Unless you want to have more headaches than trying to get SLI profiles to work on a crossfire setup, you'd probably be better off just buying Vista :).

EDIT: Oh, and donating probably isn't money well spent it seems since the devs working on this have a history of empty promises: http://linux-and-free-software.blogspot.com/2006/11/project-alky-is-dead.html

JOPS
10-26-07, 01:27 PM
If it's gonna be done, it's all with emulation, so it's gonna slow down everything again, so what benefit remains if dx10 normally is faster for ur games if u gonna slow it down again cause of an emulation layer???

"WINE = Wine Is Not an Emulator"

Wine is a reimplementation of the Win32 API's so the design itself doesn't cause any slowdowns. There are a few extra layers for translating 3D calls to OpenGL but they don't add much overhead and the biggest bottleneck is probably in the OpenGL drivers anyway. When they get better and more functionality is added to the Wine codebase the overall performance will get much closer to running the applications on Windows. Many CPU and memory related operations are actually faster on Linux + Wine than on Windows.

six_storm
10-26-07, 04:29 PM
DX10 in XP would be sweet. At least we know the performance would be better COUGH Crysis COUGH ;)

Rakeesh
10-26-07, 05:08 PM
The Wine project will probably do it (they have just started) but they're still focusing on DX9 so it would probably take years to finish DX10 support.

AFAIK DX9 support is pretty infantile, and will be for some time, nevermind dx10.

JOPS
10-26-07, 06:18 PM
AFAIK DX9 support is pretty infantile, and will be for some time, nevermind dx10.

Yepp, AFAIK the DX7 part is the only feature complete version in Wine but a lot of DX9 games already works.

Redeemed
10-27-07, 01:57 AM
DX10 in XP would be sweet. At least we know the performance would be better COUGH Crysis COUGH ;)

Um, no, if anything it be comparable at best. Realistically, it'll be slower.

DX10 does tie in with the new driver model found only in Vista that is not available in XP. As such, there is going to be some degree of rather "unnecessary" overhead necessary (if you get what I mean) to get it working properly with the older driver model.

So, keep waiting. Maybe by the time the next version of Windows and DX is released they'll have this finalised. ;)

JOPS
10-27-07, 04:11 AM
Um, no, if anything it be comparable at best. Realistically, it'll be slower.

DX10 does tie in with the new driver model found only in Vista that is not available in XP. As such, there is going to be some degree of rather "unnecessary" overhead necessary (if you get what I mean) to get it working properly with the older driver model.

DX10 is just an API so at least in theory nothing prevents a reimplementation to be optimized for XP and it's driver model, GNU/Linux or Mac OS.

In practice there is some overhead in Wine with the D3D->WGL translation but it's not as bad as it sounds as long as you make it right and have good OpenGL support in the drivers.

Remember that we're not talking about porting Microsoft's own implemetation of DX10 but a version that is written from scratch (without source or full documentation from Microsoft) for non-Vista systems.

BTW: I think you're a little bit too optimistic about the timeframe for a final release (The Wine project is already 14 years old and hasn't reached 1.0 yet). ;)

Redeemed
10-27-07, 05:32 AM
DX10 is just an API so at least in theory nothing prevents a reimplementation to be optimized for XP and it's driver model, GNU/Linux or Mac OS.

In practice there is some overhead in Wine with the D3D->WGL translation but it's not as bad as it sounds as long as you make it right and have good OpenGL support in the drivers.

Remember that we're not talking about porting Microsoft's own implemetation of DX10 but a version that is written from scratch (without source or full documentation from Microsoft) for non-Vista systems.

BTW: I think you're a little bit too optimistic about the timeframe for a final release (The Wine project is already 14 years old and hasn't reached 1.0 yet). ;)

So why all the fuss about it then? Why do people get there panties in a bunch screaming about DX10 making it to XP when odds are this project might never see the light of day? :p

My point: I just wish they'd let this dead horse R.I.P. ;)

LycosV
10-27-07, 09:42 AM
You're actually both right: In theory you could write a completely new interface for Direct X in an OS other than Vista that would work, but without the source it's like building a car engine from scratch, in the dark. In other words, it's a dead horse. :)

Besides, if anyone ever completed one in a reasonable ammount of time Microsoft would be on them like a hoarde of locusts. The only way you could do it fast would be to have access to the source code. And if somebody posted something made with stolen Microsoft source code on the net, it would be the fastest ticket to white-collar prison you've ever seen.

ViN86
10-27-07, 10:51 PM
itll never happen. only way it could be possible is to write it out of the view of the public eye and then release it through torrents.

JOPS
10-28-07, 08:47 AM
So why all the fuss about it then? Why do people get there panties in a bunch screaming about DX10 making it to XP when odds are this project might never see the light of day? :p

My point: I just wish they'd let this dead horse R.I.P. ;)

I guess all the fuss is about freedom of choice and not needing Microsoft to run DX10 software. Many people don't want to touch Vista (some with very good reasons) but still want to be able to use DX10 stuff.

Wine is free software and can be ported to run native on almost any operating system running on x86 compatible hardware. There's a hell of a lot of non-Microsoft systems in the world and some are growing pretty fast.

The Wine project is commercially supported through CodeWeavers and is very much alive so it's far from being a dead horse. DX10 is just not a priority at the moment, but will probably be sooner or later, and they have already started to implement the infrastructure needed for it. AFAIK DX7 support is feature complete and a lot of DX9 stuff is already working.

Of course it would be better if all game companies ditched DX and started to always support open API's but even when (hopefully not if) they do lots of people will most likely have a lot of legacy stuff depending on Vista/DX10 around that they might want to use on their more modern systems (that's where Wine would help).

My point: No need to hurry but what's wrong with freedom? ;)

JOPS
10-28-07, 08:50 AM
You're actually both right: In theory you could write a completely new interface for Direct X in an OS other than Vista that would work, but without the source it's like building a car engine from scratch, in the dark. In other words, it's a dead horse. :)

So how come we have working DX7 support and much of DX9 in Wine already? And yes, it is reverse engineering and takes lots of time but IMO that's not the same thing as being a dead horse.

JOPS
10-28-07, 08:54 AM
itll never happen. only way it could be possible is to write it out of the view of the public eye and then release it through torrents.

Free software is most often developed in view of the public eye, that's the whole idea. You can download the source code for Wine today but there's probably not much DX10 related stuff in it yet, maybe a few headers, but more will come.

LycosV
10-28-07, 09:03 AM
So how come we have working DX7 support and much of DX9 in Wine already? And yes, it is reverse engineering and takes lots of time but IMO that's not the same thing as being a dead horse.

Hey, I didn't say it was impossible. Some people are very good at building things in the dark. :)

I meant that it's a dead horse because either A. it will only ever come out after it's market desire has passed (ie: nobody is running XP anymore) or B. with a significant performance decrease over running DX10 in Vista which eliminates the whole reason people want it in XP.

ViN86
10-28-07, 01:42 PM
Free software is most often developed in view of the public eye, that's the whole idea. You can download the source code for Wine today but there's probably not much DX10 related stuff in it yet, maybe a few headers, but more will come.
i know, but some of those links posted here are about companies trying to make a DX10 port. they will get a cease and desist VERY quickly if they make any progress at all.

JOPS
10-28-07, 02:59 PM
I meant that it's a dead horse because either A. it will only ever come out after it's market desire has passed (ie: nobody is running XP anymore) or B. with a significant performance decrease over running DX10 in Vista which eliminates the whole reason people want it in XP.

It doesn't matter because Vista is slow itself and it will also be obsolete one day. In the future we will also have machines with much better performance than today and we don't even know what kind of operating systems we will use.

But I assume we still want to play some of todays games (which might have become classics over time) on them (like people who still play DOS, C64 or Nintendo NES games). Wine will most likely make it easier to do that even when both XP and Vista are gone.

There's always a desire for compatibility with legacy applications and who knows when Microsoft suddenly stops supporting their current stuff? ;)