Intel To Acquire Havok

Any game really, not just multiplayer.

If a building explodes and rains down objects that can hurt you, changes in the number of objects change how the game plays. If you change the number of objects based on CPU power some players will have it easier than others because there are less objects that can hurt you.

Developers don't want the intended game play experience to change between hardware configurations. They will let you scale graphics and effect physics because it has no impact on game play, but for anything that does have a impact they have to consider the lowest CPU performance gamers may have at that particular time and build around that.

When the vast majority of gamers are guaranteed to have dual cores, then we will be seeing games utilise that extra core for game play physics. Which comes down to my original point about the delay in the release of hardware to when we see games properly take advantage of it, what comes out today or tomorrow is irrelevant because it won't have a impact for a long time.
 
XamediX said:
I also don't like the fact that Intel and/or AMD might be pushing to remove Nvidia from the market. If they disappear, who else is gonna come out with superior products for them to catch up on? Ok, sorry that was borderline flamebait, but you get the gist of my point.
Click to expand...

Nvidia just came to my college(Purdue University) for a job fair and I signed up for a presentation with them. Trust me, Nvidia is going no where. They are in so many other markets, its just they are mostly known for the Geforce video cards. Even in the video card market, they are pretty strong in and will continue to be strong in. From what we seen so far, hardware physics acceralation isnt really on game developers to do list. All games mostly use the software based Havok engine.

It be interesting to see where Intel goes with this though.
 
Lazy_Moron said:
Nvidia just came to my college(Purdue University) for a job fair and I signed up for a presentation with them. Trust me, Nvidia is going no where. They are in so many other markets, its just they are mostly known for the Geforce video cards. Even in the video card market, they are pretty strong in and will continue to be strong in. From what we seen so far, hardware physics acceralation isnt really on game developers to do list. All games mostly use the software based Havok engine.

It be interesting to see where Intel goes with this though.
Click to expand...

Nobody is immune from being taken out of the market, we have Voodoo to thank for that lesson.

Also, IU is gonna win the bucket this year
 
Will be interesting to see what intel has been planning. Get some exclusives Kyle! ;)
 
Anybody hear of Hydracore?

Havok’s HydraCore™ software technology is designed to help game developers tap into the true power of these systems, by distributing real-time computation of physics and character animations flexibly and evenly across multiple processors.”

HydraCore is an adaptive middleware technology that will allow game developers to flexibly assign and remove threads from the computationally-heavy tasks that next-generation games will require – from computing physics simulations for dynamic and destructible worlds, to helping game characters respond more intelligently to each other and their environments. HydraCore accelerates these elements of game-play by dividing up the work load and dispatching it to any number of central or secondary processing units such as the AMD Athlon™ 64 processor family. HydraCore ensures Havok-powered games can harvest the native scalability of the next wave of consumer PC game platforms – without requiring specialized add-on hardware.
 
riot8ap said:
Anybody hear of Hydracore?

Havok’s HydraCore™ software technology is designed to help game developers tap into the true power of these systems, by distributing real-time computation of physics and character animations flexibly and evenly across multiple processors.”

HydraCore is an adaptive middleware technology that will allow game developers to flexibly assign and remove threads from the computationally-heavy tasks that next-generation games will require – from computing physics simulations for dynamic and destructible worlds, to helping game characters respond more intelligently to each other and their environments. HydraCore accelerates these elements of game-play by dividing up the work load and dispatching it to any number of central or secondary processing units such as the AMD Athlon™ 64 processor family. HydraCore ensures Havok-powered games can harvest the native scalability of the next wave of consumer PC game platforms – without requiring specialized add-on hardware.
Click to expand...

Ah so maybe they bought them so they could change "AMD Athlon™ 64" to "Intel Core2Duo/Core2Quad" ;). Does sound like something they would be interested in having though.
 
riot8ap said:
I agree that hydracore has been around for a couple years. What I found interesting is a link on Intel's site that I originally meant to put in my first post but couldn't check it out cuz I was at work..dunno how old it might be.

http://game-on.intel.com/eng/geekout/havok/default.aspx
Click to expand...
That video has a very interesting part right near the end, I've typed up what I think I heard.

I think one of the key points that we have seen from this so far is that software needs hardware solutions, especially a system like Havok which is computationally heavy and a real time system.

So the software solutions that we have been provided by Intel, being primarily the Intel core 2 duo and the Intel core 2 quad based systems, really allow us to deliver the astonishing and immersive game play experience that you just can't do on a single core based system.
Click to expand...
Am I the only one who finds it odd that he admitted there needs to be a hardware solution for physics, but then goes on the praise Intel? From my interpretation they are saying beggars cannot be choosers, they use what they can get but would prefer a hardware solution.

Of course that could just be me, anyone else wants to interpret?

As for what the video itself showed; it looks like a interesting method, but I imagine identifying groups of objects for separating into individual tasks adds a little bit of a additional performance hit. Maybe not on a small scale; but when you are dealing with thousands or tens of thousands of objects (such as shown in Ageia & Havok FX demo's) that must be separated into groups on the fly, I can imagine seeing a noticeable hit compared to if it was threaded directly in code. However despite that, this method is certainly more adaptable and easier for developers to implement, which is most likely a good trade off.
 
LeTUOtter said:
Ah, reprise the late 90s video card wars. ^_^
I think gamers are the only ones who don't want NVIDIA gone.
Click to expand...


I would agree with you on both points.

Interestingly though, NVIDIA realizes that their past strength (focusing on video cards alone) is a terrible weakness in the coming market. NVIDIA desperately wants a "platform" but no matter how much they say it, they don't have a CPU.

Don't think for one second that Intel cannot build a badass video gamers card. They have hired up all the help in the industry. People from NVIDIA, AMD, ATI, Matrox, SGI, etc etc etc. They have put together a solid team and given the profit dollars they have to invest in research I do not think they will end up with a meaningless popcorn fart. Put a solid Intel video card together with their already stellar chipsets, CPUs, and integrate Havok's hardware physics engines into their video cards and allow you to scale up to 4 cards across PCIe2.0 and that IS a platform. Much the same model that AMD is spearheading.
 
AnnoyedDragon said:
That video has a very interesting part right near the end, I've typed up what I think I heard.


Am I the only one who finds it odd that he admitted there needs to be a hardware solution for physics, but then goes on the praise Intel? From my interpretation they are saying beggars cannot be choosers, they use what they can get but would prefer a hardware solution.

Of course that could just be me, anyone else wants to interpret?
Click to expand...

I'm assuming he said "hardware" when talking about the Core architecture solutions from Intel. It would be strange to call it a software solution. ;)

Interesting comment you found, though!

As for what the video itself showed; it looks like a interesting method, but I imagine identifying groups of objects for separating into individual tasks adds a little bit of a additional performance hit. Maybe not on a small scale; but when you are dealing with thousands or tens of thousands of objects (such as shown in Ageia & Havok FX demo's) that must be separated into groups on the fly, I can imagine seeing a noticeable hit compared to if it was threaded directly in code. However despite that, this method is certainly more adaptable and easier for developers to implement, which is most likely a good trade off.
Click to expand...

If i understand what is being described here, this is a long-time trick in MD (molecular dynamics) simulations, where the overhead to partition up the volume into patches and/or calculate "interactive groups" in an ongoing fashion is much lower than the computations required to treat all atoms as interacting particles (which, technically, they are -- fortunately, you can chose a distance where their interaction is negligible). That facilitates spreading the simulation over multiple cores/systems. Of course, it requires some tricks for special cases... Graphics rendering is the same way except that it's all rendered/rejected according to observer position (one "patch"), but the comparison was probably more apt with MD because it's physics.
 
One of the interesting things I think some of you have overlooked is that, as Intel it moving into the GPU market, and Havok had all these deals for Havok FX, nVidia and ATI may no longer want to work with Havok as they would then be contributing to their competition. Intel buying up Havok puts them in a really good position, where if nVidia and ATI stick with Havok, Intel makes money. If nVidia and ATI decide to go somewhere else, they may end up with "second best" Ageia of if a totally different company buys up Ageia in response to this, nVidia and ATI may be left out in the cold on the physics side of things.

As for Ageia going anywhere, I think they have enough valuable code base and potential, that someone will buy them up if it comes to that. Even if a company just wants their software or hardware side, and dump the rest, they would be worth the purchase.
 
Chee42 said:
As for Ageia going anywhere, I think they have enough valuable code base and potential, that someone will buy them up if it comes to that. Even if a company just wants their software or hardware side, and dump the rest, they would be worth the purchase.
Click to expand...

I don't think PhysX is going anywhere.
I have heard "AGEIA is dying" since the PhysX card came out....kinda like the AMD-forum when the Core 2 Duo came out....it's silly. :rolleyes:
 
Atech said:
I don't think PhysX is going anywhere.
I have heard "AGEIA is dying" since the PhysX card came out....kinda like the AMD-forum when the Core 2 Duo came out....it's silly. :rolleyes:
Click to expand...

People believe what they want, many people are simply against the PPU because they don't want yet another item on their upgrade list. Whether or not the PPU becomes successful is irrelevant, they prefer the future promises of CPU and GPU physics than the working here and now solutions the PPU has to offer. They set their standards unreasonably high for a product in the circumstances the PPU is under, one person I met said it would take '5 AAA titles' in 'one year' to justify a positive image which is of course ridicules.

Any solution that doesn't require buying another component is looked upon fondly, even if they have nothing out and exist on future promises, I know that isn't the only reason for being against the PPU but it is a ever present one with a large portion of gamers. For instance there is a PPU thread currently running in the UT3 PC forum at GameSpot; the ignorance and assumptions being made is mind numbing. You can only read “dont you guys realize that the 8800gtx has a physics processor in it” so many times before you feel like clobbering them.

As for what Chee42 said I really don't see the need for GPU companies to get involved with Agiea. They would probably just be better off refining their GPGPU architecture which from what I have seen makes the GPU a better physics processor than Havok FX. Of course my understanding of GPGPU is minimal so there could be something I am missing.
 
AnnoyedDragon said:
As for what Chee42 said I really don't see the need for GPU companies to get involved with Agiea. They would probably just be better off refining their GPGPU architecture which from what I have seen makes the GPU a better physics processor than Havok FX. Of course my understanding of GPGPU is minimal so there could be something I am missing.
Click to expand...


HavokFX is GPU-physics? :confused:
 
Atech said:
HavokFX is GPU-physics? :confused:
Click to expand...
Well yeah but it is limited, Havok FX still relies on the CPU for making all the game play calculations for physics. GPGPU, if I understand it, allows for better code acceleration by the GPU and hence *should* be able to run physics affects without having to rely on the CPU so much.

Can I get a confirmation/correction from someone with a better understanding of GPGPU?
 
AnnoyedDragon said:
People believe what they want, many people are simply against the PPU because they don't want yet another item on their upgrade list
Click to expand...

And yet we have some flaming the PPU, while advocating GPU-physics...using a more expensive, power-hungry and yet used by any games solution...

Using the CPU for physics also has a major drawback:
Physics is massive multithreaded, even the 16 threads(8 cores with HT) of "Nehalem" is way from optimal...

Sorry for the "late" afterthought.
 
GPU physics sure seems like a great idea, what with Crysis already fully loading the GPU :rolleyes:
 
[RCKY] Thor;1031593424 said:
GPU physics sure seems like a great idea, what with Crysis already fully loading the GPU :rolleyes:
Click to expand...

No,no...you have have to buy a new mobo and a 3rd top of the line card and the pursuade someone to use GPU-physics, then you are alle set :rolleyes: ;):D

I want my CPU to compute, my GPU to do the GFX, my NIC to do the networking, my soundcard to do the soundprocessing and my PPU top do the physcis...simple as that.
 
I want my CPU to compute, my GPU to do the GFX, my NIC to do the networking, my soundcard to do the soundprocessing and my PPU top do the physcis...simple as that.
Click to expand...

Sometimes the simplest ideas are the hardest to grasp ;)

Dedicated hardware is the way to go imho, integrated solutions are good for mainstream, not for tech head like us, problem is... tech head and people live in the same world :D
 
You must log in or register to reply here.
Back
Top