Microsoft incorporating physics on GPU into DirectX?

[InorganicMatter]

Lord, I hope not. Seems they have put out an ad for a physics persont to join the DX squad:

http://www.hardocp.com/news.html?news=MTk2NTEsLCxobmV3cywsLDE=

If M$ is really beind this, I take back everything I said. Half-physics will prevail, and once again the true innovator (AGEIA) will get left in the dust in favor of the large corporation.

 

[MrNasty]

That's been floating around for a while, it's a way off yet and no-one knows what hardware path it will take - if any...

 

[Netrat33]

Lord, I hope not. Seems they have put out an ad for a physics persont to join the DX squad:

http://www.hardocp.com/news.html?news=MTk2NTEsLCxobmV3cywsLDE=

If M$ is really beind this, I take back everything I said. Half-physics will prevail, and once again the true innovator (AGEIA) will get left in the dust in favor of the large corporation.

Quite honestly, I doubt it would/will ever prevail unless it became really cheap. $200 is a lot for "Just physics" which yoru everyday shopper would be thinking considering games have been using Havok for so long and see physics in games all the time.

 

[RazorWind]

I take back everything I said. Half-physics will prevail, and once again the true innovator (AGEIA) will get left in the dust in favor of the large corporation.

How do you know? The way I see it, if ANY sort of realistic physics computation is going to be done in hardware, it needs to be part of some sort of API that is interchangeable between different hardware. The best way to do this is via DirectX, since it's the API that almost everyone is using anyway.

Who's to say that future iterations of DirectX won't include the capability for interactive gameplay physics computation, or that some intrepid developer won't figure out how to do that.

Also, this leaves the possibility of having a standardized software emulation of the hardware available for those people who don't want to buy a PPU or extra video card, kind of like how Direct3D has a software renderer that can be used in place of a 3D accelerator. Right now, each game has its own proprietary version of this, but there's not really any reason I see that there can't be a standard one that every game can use.

I could see Microsoft buying out the Havok people for this, and appropriating that technology.

 

[nhusby]

well... the big dirty machine gets its hands into everything. We all knew this was comming...

I thought I heard physics was slotted for DX 11...

I feel that a DX physics solution will solve the physics problem for video games. The question is, would it support hardware acceleration? I hope so. I think it will support GPU hardware acceleration, but remember that ageia has relations with microsoft (they are on ageia's supported list anyhow.) and remember that ageia makes its profit on hardware, not software, so it is possible for them to work together. If DX physics would support Ageia hardware, that would be the best thing for Ageia. Even if it supported both Ageia's PPU and nvidia/ati GPU's I would think Ageia's PPU would outperform.

 

[pxc]

W00T! One standard API to rule them all. How can this not be good? If Ageia's PhysX is as flexible as they claim, owners of the cards can look forward to many more supported titles. People who want to do GPU-based physics (not just effects physics) also win. Fast dual and quad core CPUs will also benefit from a stanard API, just not at the same level.

Depending on how much Ageia marketeers away from a standard MS Direct Physics, the more likely it isn't going to work with DP. In that case, early PhysX adopters don't win.

 

[Moofasa~]

W00T! One standard API to rule them all. How can this not be good? If Ageia's PhysX is as flexible as they claim, owners of the cards can look forward to many more supported titles. People who want to do GPU-based physics (not just effects physics) also win. Fast dual and quad core CPUs will also benefit from a stanard API, just not at the same level.

Depending on how much Ageia marketeers away from a standard MS Direct Physics, the more likely it isn't going to work with DP. In that case, early PhysX adopters don't win.

Doesn't appear to be a win for Ageia.

It is interesting to note that Microsoft intends to process physics solely on GPUs, not specially designed physics processing units (PPU) or multi-core central processing units (CPUs). Theoretically, it means that Microsoft shares vision of Havok with its Havok FX engine that allows computing physics effects using GPU power. This may pose some threat to Ageia, who is pushing its own environment for physics processing along with its hardware PhysX approach.

 

[InorganicMatter]

Doesn't appear to be a win for Ageia.

Just as I predicted. The little guy loses again

 

[bildad]

They should have included multiple cpus as well and not just GPUs. Alas Ageia...we knew thee...no wait, we didn't. Thanks for stopping by anyway.

 

[cyks]

To say DirectPhysics will only work for a GPU is asinine.

What is next guys? Direct3D will only work for Nvidia, not ATI?

Vista will only work on an Intel, not an AMD?

C'mon!!! What sort of silly relationship is next??

 

[|CR|Constantine]

I'm glad this is happening. With MS incorporating a global API for physics in their Direct X solution, it means we will see the same standardization as video and sound.

Which means everyone is on the same page and we get to see new products quicker.

Secondly, using an older video card purely for physics is downright the best idea in a long time. I am stoked knowing that every year when I upgrade to the latest and greatest graphic behemoth that I can still use my prior one for a great physics card.

Instead of purchasing the latest and greatest physics processor card every year as well.

Sorry Ageia but you should have known better than to think you could corner the market and have everyone follow you instead of MS and Direct X.

But thanks for getting the ball rolling.

 

[vidor]

I
Secondly, using an older video card purely for physics is downright the best idea in a long time. I am stoked knowing that every year when I upgrade to the latest and greatest graphic behemoth that I can still use my prior one for a great physics card.

Actually, it really isnt. Because using a GPU for physics is not apples to oranges. To say it again, ad nauseum, GPUs cant transfer data to and from the CPU effectively. Hence, GPU powered solutions aren't the same as Ageia's. Its simply not made for floating point calculations / vector calculations. What your saying is similar to "hey, I can use my old sound card as a math co-processor instead of chucking it!".

Instead of purchasing the latest and greatest physics processor card every year as well. Sorry Ageia but you should have known better than to think you could corner the market and have everyone follow you instead of MS and Direct X.

Ageia's solution would be something that has more staying power than year to year. I would guess more along the timetable of a hard drive or a CPU. I sincerely doubt the PPU market will EVER be as volatile as the video card market (because NOTHING else is).

Lastly, like others have said, this isn't necessarily a kill for Ageia. Did Direct3d kill the voodoo cards? Or...does creative have problems getting their sound cards to work for microsoft? Theres no reason to think that Ageia's HARDWARE will be nullified by a universal API.

 

[Terra]

Lastly, like others have said, this isn't necessarily a kill for Ageia. Did Direct3d kill the voodoo cards? Or...does creative have problems getting their sound cards to work for microsoft? Theres no reason to think that Ageia's HARDWARE will be nullified by a universal API.

QFT.
Cellfactor runs fine..even if there is NO "physcis" DX9 API...

And don't think that MS and AGEIA are strangers.
Xbox360 titles use PhysX(Or at least some do)
And this thread...in this forum shows even more:PPU Forum - Microsoft Uses PhysX Engine

The Microsoft Robotics Studio delivers three areas of software:

1. A scalable, extensible runtime architecture that can span a wide variety of hardware and devices. The programming interface can be used to address robots using 8-bit or 16-bit processors as well as 32-bit systems with multi-core processors and devices from simple touch sensors to laser distance finding devices.
2. A set of useful tools that make programming and debugging robot applications scenarios easier. These include a high quality visual simulation environment that uses the Ageia Technologies™ PhysX™ engine.
3. A set of useful technology libraries services samples to help developers get started with writing robot applications.

Terra - Wonder when the paided ATI-trolls will start posting inhere

 

[InorganicMatter]

Actually, it really isnt. Because using a GPU for physics is not apples to oranges. To say it again, ad nauseum, GPUs cant transfer data to and from the CPU effectively. Hence, GPU powered solutions aren't the same as Ageia's. Its simply not made for floating point calculations / vector calculations. What your saying is similar to "hey, I can use my old sound card as a math co-processor instead of chucking it!".

QFT!

Ageia's solution would be something that has more staying power than year to year. I would guess more along the timetable of a hard drive or a CPU. I sincerely doubt the PPU market will EVER be as volatile as the video card market (because NOTHING else is). I.

Correct, AGEIA said the PPU would operate on a 2-year cycle.

Lastly, like others have said, this isn't necessarily a kill for Ageia. Did Direct3d kill the voodoo cards? Or...does creative have problems getting their sound cards to work for microsoft? Theres no reason to think that Ageia's HARDWARE will be nullified by a universal API.

A universal API that specifically locks them out. This is the equivalent of Microsoft locking out Matrox from rendering D3D.

 

[Terra]

QFT!


Correct, AGEIA said the PPU would operate on a 2-year cycle.


A universal API that specifically locks them out. This is the equivalent of Microsoft locking out Matrox from rendering D3D.

Please explain how Microsoft(and why would they?) would lock out PhysX?
PhysX got nothing to do with any directX version.
It's an API...just like Havok is.
And when did Microsoft lock out Matrox?
For the Parhelia on their website it states:

OpenGL® and Microsoft® DirectX® compliant

Terra...

 

[Moofasa~]

QFT.
Cellfactor runs fine..even if there is NO "physcis" DX9 API...

And don't think that MS and AGEIA are strangers.
Xbox360 titles use PhysX(Or at least some do)
And this thread...in this forum shows even more:PPU Forum - Microsoft Uses PhysX Engine



Terra - Wonder when the paided ATI-trolls will start posting inhere

I am guessing that DirectPhysics will be embedded into the graphics portion of DirectX, which would make it almost impossible for an Ageia card to hardware accelerate it. The only possible way would be if Ageia somehow emulated all of DirectX on their card (making it a very weak video card); I don't even think that is possible. Even then I would guess the performance would be almost equivalent to a CPU emulating DirectX (perhaps Ageia would have slightly better performance).

If this is true about Ageia not being able to support DirectPhysics on their card, then it surely is the end of Ageia. Though this is all based on rumors right now, so anything could happen. Either way we can all at least agree on that Ageia started the whole "Physics Revolution", even if it won't see the end of it. Oh well, I had high hopes too.

 

[Terra]

I am guessing that DirectPhysics will be embedded into the graphics portion of DirectX, which would make it almost impossible for an Ageia card to hardware accelerate it. The only possible way would be if Ageia somehow emulated all of DirectX on their card (making it a very weak video card); I don't even think that is possible. Even then I would guess the performance would be almost equivalent to a CPU emulating DirectX (perhaps Ageia would have slightly better performance).

If this is true about Ageia not being able to support DirectPhysics on their card, then it surely is the end of Ageia. Though this is all based on rumors right now, so anything could happen. Either way we can all at least agree on that Ageia started the whole "Physics Revolution", even if it won't see the end of it. Oh well, I had high hopes too.

Then please explain to me why the Physx works now?
Or why OpenGL games can run under XP(and Vista too)?
DirectX is not involved at alll in the PhysX as we speak..so making a "physcis"-APi under DX wouldn't mean squat for the PhysX...
Games don't need to be fully DirectX...

Look at Doom3, Quake 4 and the upcomming Prey.

Terra - AFAIK OpenAL also has nothing to do with DirectX

 

[Moofasa~]

Then please explain to me why the Physx works now?

It has it's own card that has dedicated transistors to supports its API.

Or why OpenGL games can run under XP(and Vista too)?

Who said anything about it not being able to work for Windows? I am a little confused at this point. But to hardware accelerate OpenGL, you need specific cards to dedicate transistors to support it natively (like our current video cards). Aegia cards don't have dedicated transistors to support DirectX or OpenAL, only their own physics API.

DirectX is not involved at alll in the PhysX as we speak..so making a "physcis"-APi under DX wouldn't mean squat for the PhysX...

Sure it does. If most games in the future use DirectPhysics that can only be hardware accelerated through GPU's, there is little point in investing in an Ageia card or Ageia physics (talking from a PC stand point).

Games don't need to be fully DirectX...

True. Companies could make up their own API if they wanted, but ATi and Nvidia cards will only be able to hardware accelerate DirectX or OpenGL games.

Look at Doom3, Quake 4 and the upcomming Prey.

Yeah all OpenGL games, which works because video cards natively support OpenGL.

Terra - AFAIK OpenAL also has nothing to do with DirectX

Because Creative has dedicated transistors on it's Audigy 2 and X-Fi cards to support OpenAL (all other cards use a DirectSound3D wrapper to support OpenAL, this has to be done in software though [on a CPU], which in turn decreases performance, something Ageia may have to do).

I am not sure what you are getting at. If Ageia wishes to survive, it will have to find a way to support DirectPhysics with decent performance. The problem is, at this time, it does not seem feasible.

 

[nhusby]

correct me if I am wrong here, but this doesnt change anything...

Havok offers to use current technology to implement soft physics. (can I call it "soft physics" without being slammed?)

Its general consensus that current GPU's cannot sufficiently communicate to do [H]ard physics (rigid bodies interacting with players and whatnot.)

I'm just speculating here, but MS is not doing full blown [H]ard phyiscs. (still ageia vs havok argument) Either that or MS knows something about upcomming GPU's we dont. (duh)
so its it for havok.....

 

[Moofasa~]

correct me if I am wrong here, but this doesnt change anything...

Havok offers to use current technology to implement soft physics. (can I call it "soft physics" without being slammed?)

Its general consensus that current GPU's cannot sufficiently communicate to do [H]ard physics (rigid bodies interacting with players and whatnot.)

I'm just speculating here, but MS is not doing full blown [H]ard phyiscs. (still ageia vs havok argument) Either that or MS knows something about upcomming GPU's we dont. (duh)
so its it for havok.....

I thought DirectX 10 was able to communicate with the CPU better? I guessing Microsoft will be able to mutate DirectX in order to allow for gameplay physics. Or maybe not. Who knows.

 

[Terra]

It has it's own card that has dedicated transistors to supports its API.

A free APi...

Who said anything about it not being able to work for Windows? I am a little confused at this point. But to hardware accelerate OpenGL, you need specific cards to dedicate transistors to support it natively (like our current video cards). Aegia cards don't have dedicated transistors to support DirectX or OpenAL, only their own physics API.

In was brought up in OrganicMatters's post:
"A universal API that specifically locks them out. This is the equivalent of Microsoft locking out Matrox from rendering D3D."
And neither DirectX or OpenGL have anything to do with the PhysX.
Graphics != Physcis.
And the PhysX dosn't need DirectX to run.

Sure it does. If most games in the future use DirectPhysics that can only be hardware accelerated through GPU's, there is little point in investing in an Ageia card or Ageia physics (talking from a PC stand point).

Not understood?
OpenGL works independant from DirextX
OpenAL works independant from DirectX
PhysX Novodex works independant from DirectX
The Physx API gives direct access to the PhysX...outside DirectX.

Even if Microsoft made a physcis directX API that did not support the PhysX...it wouldn't matter for the PhysX...
You have a free developerkit(as long as you add support for the PhysX) with it's own API to the hardware accelerated Physcis.
Look at the list of developers that support the PhysX...with out DirectX involved.
DirectX API is not required at all...it runs fine today without it...

True. Companies could make up their own API if they wanted, but ATi and Nvidia cards will only be able to hardware accelerate DirectX or OpenGL games.

Sorry?
Intel, S3, Matrox and others...what have they done you?

Yeah all OpenGL games, which works because video cards natively support OpenGL

Natively?
Without a driver?
I think not...
Microsoft tried to wrap OpenGL via DirectX 10..but stopped as the profressional industry said "What the fuck?"
But then it' good that PhysX has "native" support for Novodex then..problem solved

Because Creative has dedicated transistors on it's Audigy 2 and X-Fi cards to support OpenAL (all other cards use a DirectSound3D wrapper to support OpenAL, this has to be done in software though [on a CPU], which in turn decreases performance, something Ageia may have to do)

I see no loss(FPS) bigger than under Directx via OpenAL sound?
And yes I do own a X-Fi...speaking from personal experince.
But the same reply as above would apply to this.
And you think the CPU isn't involed in hardware call under directx?

I am not sure what you are getting at. If Ageia wishes to survive, it will have to find a way to support DirectPhysics with decent performance. The problem is, at this time, it does not seem feasible.

They have the BEST Physics preformance so the part about "decent preformance" leaves me lost for words?

Terra...

 

[Terra]

I thought DirectX 10 was able to communicate with the CPU better? I guessing Microsoft will be able to mutate DirectX in order to allow for gameplay physics. Or maybe not. Who knows.

DirectX 10 is supposed to give less CPU overhead from what I understand...

Terra...

 

[psychoace]

Not understood?
OpenGL works independant from DirextX
OpenAL works independant from DirectX
PhysX Novodex works independant from DirectX
The Physx API gives direct access to the PhysX...outside DirectX.

OpenGL isn't used that much in games
OpenAl isn't use that much in games
Physx isn't use that much in games
Direct X has a stake in 90% of games.

If publishers see that they have an easy and fairly big market by going with directx based physics then I doubt they will branch out much from there. The only way they can win is directx physics cost to much and developers find it hard to code for.

 

[Moofasa~]

In was brought up in OrganicMatters's post:
"A universal API that specifically locks them out. This is the equivalent of Microsoft locking out Matrox from rendering D3D."
And neither DirectX or OpenGL have anything to do with the PhysX.
Graphics != Physcis.
And the PhysX dosn't need DirectX to run.

I never said PhysX needed DirectX to run. So I don't quite see your point. But if many future games use only DirectPhysics API instead of PhysX API (which they have a good chance of doing), and Ageia's card can't support the DirectPhysics API, it will hurt Aegia's chances in the future.

Not understood?
OpenGL works independant from DirextX
OpenAL works independant from DirectX
PhysX Novodex works independant from DirectX
The Physx API gives direct access to the PhysX...outside DirectX.

See above message.

Even if Microsoft made a physcis directX API that did not support the PhysX...it wouldn't matter for the PhysX...
You have a free developerkit(as long as you add support for the PhysX) with it's own API to the hardware accelerated Physcis.
Look at the list of developers that support the PhysX...with out DirectX involved.
DirectX API is not required at all...it runs fine today without it...

Also see above message. I find it hard to believe most developers would use PhysX over a standardized physics API. What happens if the PhysX API becomes unpopular and obsolete?

Sorry?
Intel, S3, Matrox and others...what have they done you?

When I say ATi/Nvidia, I mean video cards today in general.

Natively?
Without a driver?
I think not...
Microsoft tried to wrap OpenGL via DirectX 10..but stopped as the profressional industry said "What the fuck?"
But then it' good that PhysX has "native" support for Novodex then..problem solved

Natively as in a DirectX 8.1 card cannot not run DirectX 9 features in a game. A card needs to have dedicated transistors to support DirectX and OpenGL. Your OpenGL wrapper analogy has no correlation to the conversation.

That's great PhysX supports Novodex, but that isn't the problem. The problem is right now it cannot support DirectPhysics API (the API I am assuming will be the most popular, though I could be wrong).

I see no loss(FPS) bigger than under Directx via OpenAL sound?
And yes I do own a X-Fi...speaking from personal experince.
But the same reply as above would apply to this.
And you think the CPU isn't involed in hardware call under directx?

If you actually read my post, you would have known I said X-Fi and the Audigy 2 series are the only cards that do not use the DirectSound3D wrapper because it has dedicated transistors toward OpenAL. So you could not experience it personally.

They have the BEST Physics preformance so the part about "decent preformance" leaves me lost for words?

For PhysX, but if the future is DirectPhysics, it doesn't matter how well it handles PhysX, it will matter how well it handles DirectPhysics (assuming it even can).

 

[nhusby]

Not that the world is fair, but in all fairness microsoft has an obligation (to its customers) to make its software compatible with available hardware. To ignoir the PPU despite its performance advantages over the GPU would be moronic. Consumers will be the loosers if DX physics ignoir Ageia's PPU.

 

[Terra]

OpenGL isn't used that much in games
OpenAl isn't use that much in games
Physx isn't use that much in games
Direct X has a stake in 90% of games.

And?
They work just fine?
And are you trying to call Doom3, Quake 4, Quake Wars and Prey for "small" titles?

If publishers see that they have an easy and fairly big market by going with directx based physics then I doubt they will branch out much from there. The only way they can win is directx physics cost to much and developers find it hard to code for.

Good thing publishers don't code games then...
That is up the the developer...
And what a hard choice.
Pay $250.000 for HaovkFX....or pay nothing for more?

Terra -Though one

 

[Moofasa~]

Not that the world is fair, but in all fairness microsoft has an obligation (to its customers) to make its software compatible with available hardware. To ignoir the PPU despite its performance advantages over the GPU would be moronic. Consumers will be the loosers if DX physics ignoir Ageia's PPU.

I agree, but perhaps that way they are implementing their Physics API makes it impossible for Ageia's PPU to support it. We will see how it pans out.

Pay $250.000 for HaovkFX....or pay nothing for more?

No one is talking about Havok's API; we are discussing Microsoft's standardized physics API. And between the three, I will bet most developers will choose Microsoft's API. This will hurt Havok just as much as it will hurt Ageia.

 

[Terra]

I never said PhysX needed DirectX to run. So I don't quite see your point. But if many future games use only DirectPhysics API instead of PhysX API (which they have a good chance of doing), and Ageia's card can't support the DirectPhysics API, it will hurt Aegia's chances in the future.

Not if the DirectX API only will do eyecandy-physics.
Then the PhysX via Novodex would prefrom better regardless...outside DirectX

Also see above message. I find it hard to believe most developers would use PhysX over a standardized physics API. What happens if the PhysX API becomes unpopular and obsolete?

If the want real gameplay physics, a Eye Candy API would force them too...
And looking at hte numbers I'd saay in 2 years my PhysX will still be kicking and being very much alive..and able to preform.
But with +60 developers supporting the PhysX your "theory" has little foundation in reality, sorry

When I say ATi/Nvidia, I mean video cards today in general.

Kinda hard to read form you statement, but okay.

Natively as in a DirectX 8.1 card cannot not run DirectX 9 features in a game. A card needs to have dedicated transistors to support DirectX and OpenGL. Your OpenGL wrapper analogy has no correlation to the conversation.

Or the driver needs to be able to tell DirectX what hardware capabilites it has.
I remember when the first DirectX came out.
Next driverupdate for my graphics card was DX-certified.
Even if the card was made before any mention of directX...thus lacking the "dedicated" hardware.

That's great PhysX supports Novodex, but that isn't the problem. The problem is right now it cannot support DirectPhysics API (the API I am assuming will be the most popular, though I could be wrong).

All we know is that the "physcis" API will be only for eyecandy...not real physcis.
So far the most popular physics API among developers is the AGIEA's Novodex for the PhysX.

If you actually read my post, you would have known I said X-Fi and the Audigy 2 series are the only cards that do not use the DirectSound3D wrapper because it has dedicated transistors toward OpenAL. So you could not experience it personally.

It can do both..via it's drivers.
No driver support...no hardware calls...

For PhysX, but if the future is DirectPhysics, it doesn't matter how well it handles PhysX, it will matter how well it handles DirectPhysics (assuming it even can).

Again...effects-physcis only by the rumours so far...
Full flegde gameplay physics not included..you think developers will settle for that?
I don't..and the up comming games listed with PhysX support would also suggest otherwise...

Terra...

 

[Moofasa~]

I understand what you are saying, if this DirecX Physics API is only eye candy, PhysX would be a better solution. I agree with you on this. But what if DirectX Physics API is gameplay physics too? Then Ageia's in big trouble.

You riding awfully a lot on the chance that the DirectX Physics API is eye candy physics only.

 

[Terra]

I understand what you are saying, if this DirecX Physics API is only eye candy, PhysX would be a better solution. I agree with you on this. But what if DirectX Physics API is gameplay physics too? Then Ageia's in big trouble.

You riding awfully a lot on the chance that the DirectX Physics API is eye candy physics only.

That the only information we have so far.
So I have to go by it.
If that information changes, I will rewise my point of view accordingly to the new information

Terra...

 

[vidor]

You will be a member of the core engine team who will be primarily responsible for working closely with our Direct3D team, helping to define, develop and map optimized simulation and collision algorithms onto data structures that are optimized for the GPU. Extensive experience with graphics shading languages such as HLSL is expected as well as a good understanding of modern graphics hardware and associated algorithms.

Sounds like GPU half-physics to me.

Edit: further thoughts
Seems to me microsoft would be silly to bet on doing DX10 just for people with multi-gpu solutions. I think the sector of the market willing to buy an additional PCI add-in (PhysX) is a whole lot bigger than those running SLi-capable motherboards AND willing to buy an additional video card.

 

[Terra]

Sounds like GPU half-physics to me.

Yup...
But do you wanna bet that the PR-spin on this will be "real physcis"

Terra - Tired of PR-Bull...

 

[nhusby]

Sounds like GPU half-physics to me.

LoL

 

[Farseer]

All of this pure speculation. We have only seen one ad for one engineer. For all we know Microsoft hired the game-play physics engineer internally. In any case, I find it highly unlikely that Microsoft would implement a new API and not include hardware support for existing hardware. But, that’s just my opinion.

 

[Terra]

This was interesting:
http://members.microsoft.com/careers/search/details.aspx?JobID=6b94ac4f-0627-4851-8e6a-633186d96261

The Windows Graphics and Gaming Technology group is looking for a software design engineer to join a growing team responsible for developing Direct Physics. This team is responsible for delivering a great leap forwards in the way game developers think about integrating Physics into their engines. Physics and real time, accurate simulation is a key part of the next generation gaming experience, bringing increased realism, greater immersion and more interesting experiences.
You will be a member of the core engine team who will be primarily responsible for working closely with our Direct3D team, helping to define, develop and map optimized simulation and collision algorithms onto data structures that are optimized for the GPU. Extensive experience with graphics shading languages such as HLSL is expected as well as a good understanding of modern graphics hardware and associated algorithms.
You should have excellent C/C++ skills with at least 4 years of development experience. You should have a good understanding of 3D Math as well as current CPU and GPU architecture. Knowledge of DirectX and shipping game applications is strongly recommended. An understanding of Havok, Ageia, MathEngine, Meqon or ODE would be ideal. A BA/BS degree in Computer Science, Math, Physics or equivalent experience is required.

If it's only GPU-physics, why would it be ideal if you know something about AGEIA?

Terra...

 

[HighTest]

It's ok for Games, GPU physics that is, but for real life robotics Aegia is the way to go

http://news.yahoo.com/s/zd/20060620/tc_zd/181304


That's right, an honest to goodness robotics development solution from MS.

Check out this quote

Perhaps the most powerful part of this new development environment is the 3D graphic simulation environment. It integrates an Aegia 3D physics engine to apply realistic physics to a computer-generated robot image. The on-screen robot can be a representative of a robot connected directly to the development PC or a virtual one. Routines can be completely worked out in advance and then off-loaded to the robot or used to develop a new one. Trower explained the code is ready to help work with and build robots as varied as 8-bit Lego Mindstorm robots, PC-based robots, and on up to industrial workhorses.

Note that the solution needs to model a real life simulation environment as the Robotics will need to interact in the real environment. Aegia's solution is the one used.

Now I'll speculate that xbit and the others may be off base some. Why go through the development effort, release a product and not reuse some of that development work for DirectX? If this is the case then Aegia has the win.

More links to the "AEGIA based Microsoft Robotics Studio":

http://www.pcmag.com/article2/0,1895,1979616,00.asp
http://www.pcmag.com/article2/0,1895,1978565,00.asp
http://www.engadget.com/2006/06/21/microsoft-reveals-multi-platform-robotics-studio/

Oh and finally, not to leave it to speculation, from the source itself - MS:

http://msdn.microsoft.com/robotics/
http://www.microsoft.com/presspass/press/2006/jun06/06-20MSRoboticsStudioPR.mspx

Check out this little tasty quote, note the "hardware acceleration support"!!

Microsoft has licensed the PhysX™ engine from AGEIA™, a pioneer in hardware-accelerated physics, enabling real-world physics simulations with robot models. The PhysX simulations can also be accelerated using AGEIA hardware.

I'll say it again, why waste programming dollars and effort for another API when you have one finished already? Looks like the one that is finished is...... Havok.

 

[mastrmndv2]

I think one of the main facts you guys are completely looking over is if DirectX physics is going to even be available for DirectX 10.....I'd imagine we'll be waiting for a while.

 

[Farseer]

I think one of the main facts you guys are completely looking over is if DirectX physics is going to even be available for DirectX 10.....I'd imagine we'll be waiting for a while.

Well, according to this Inquirer article there is some chatter that DX10.1 is already being considered. The article states that there is no word as to what the new additions might be, but I would think that DirectPhysics would be a prime candidate.

http://www.theinquirer.net/?article=32524

 

[mastrmndv2]

Well, according to this Inquirer article there is some chatter that DX10.1 is already being considered. The article states that there is no word as to what the new additions might be, but I would think that DirectPhysics would be a prime candidate.

http://www.theinquirer.net/?article=32524

Granted I had not read that article but the one that I did read stated "DirectX 10 will not include physics capability."

http://www.tgdaily.com/2006/06/21/microsoft_directx_physics/

Granted it is Toms, however, its still in the News section and I would like to think that they had something concrete to go on before making a statement as bold as that, versus the rumors going around at The Inquirer. However, at the same time I'm not ruling it out as its far to advanced to tell yet.

 

[RazorWind]

Just as I predicted. The little guy loses again

But if you think about it, everyone else wins.

The fact is that graphics cards are something that every single person who will care about hardware physics has at least one of. Right now, the appropriation of an extra one for physics is kind of an ugly hack, but what if you altered the GPUs such that they were more capable in this regard. It seems like the GPUs are kind of moving in that direction anyway.

In a few years, it's entirely possible, in my opinion, that a GPU used to do this could perform every bit as well as a purpose built physics processor, but not necessitate the purchase of another $300+ part, especially if it's done in such a way that the last generation card can be used for physics, or if Ageia gets bought up by nVidia or ATI, and that technology is simply added to the graphics cards, instead of appropriating the existing GPU.