Slashdot Log In
NVIDIA To Enable PhysX For Full Line of GPUs
Posted by
Soulskill
on Fri Jun 20, 2008 07:08 PM
from the new-toys dept.
from the new-toys dept.
MojoKid brings news from HotHardware that NVIDIA will be enabling PhysX for some of its newest graphics cards in an upcoming driver release. Support for the full GeForce 8/9 line will be added gradually. NVIDIA acquired PhysX creator AGEIA earlier this year.
Related Stories
[+]
NVIDIA To Buy AGEIA 160 comments
The two companies announced today that NVIDIA will acquire PhysX maker AGEIA; terms were not disclosed. The Daily Tech is one of the few covering the news to go much beyond the press release, mentioning that AMD considered buying AGEIA last November but passed, and that the combination positions NVIDIA to compete with Intel on a second front, beyond the GPU — as Intel purchased AGEIA competitor Havok last September. While NVIDIA talked about supporting the PhysX engine on their GPUs, it's not clear whether AGEIA's hardware-based physics accelerator will play any part in that. AMD declared GPU physics dead last year, but NVIDIA at least presumably begs to differ. The coverage over at PC Perspectives goes into more depth on what the acquisition portends for the future of physics, on the GPU or elsewhere.
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
Hentai (Score:5, Funny)
Re:Hentai (Score:5, Funny)
They're having difficulties realistically modelling penetration. Close contact like that tends to lead to numerical instabilities in physics engines. There's not much Physx can do to help, though.
Parent
Re:Hentai (Score:5, Funny)
That's why there are teams of researchers working night and day to improve the state of tentacle modeling.
If you have what it takes to advance the state of the art there could be a big government grant and a PhD in it for you.
Parent
Re:Hentai (Score:4, Funny)
Parent
Re: (Score:3, Funny)
Re: (Score:3, Funny)
Re:Hentai (Score:5, Insightful)
That's just disturbing.
Parent
Re:Hentai (Score:4, Funny)
Not as disturbing as the Chronicles of Goatse.cx Part IV: Rick Astley's Revenge
Parent
Re: (Score:2)
You won this story.
Happy Friday.
Works on just the one card? (Score:4, Interesting)
I read TFA, but it didn't really give many details as to how this works, just some benchmarks that don't really reveal much.
Will this work on single cards or will it require an SLi system where one card does the PhysX and the other does the rendering?
Plus, how does handling PhysX affect framerates? Will a PhysX enabled game's performance actually drop because the GPU is spending so much time calculating it and not enough time rendering it, or are they essentially independent because they're separate steps in the game's pipeline?
Re:Works on just the one card? (Score:5, Informative)
Yes, it works on one card. I have enabled it on my 8800GT earlier today. The CUDA/PhysX layer gets time-sliced access to the card. Yes, it will drop framerates by about 10%.
OTOH if you have 2 cards, you can dedicate one to CUDA and one to rendering so there won't be a hit. The cards need to NOT be in SLI (if they're in SLI, the driver sees only one GPU, and it will time-slice it like it does with a single card). This is actually the preferred configuration.
Parent
Re:Works on just the one card? (Score:4, Informative)
You need the latest unreleased yet drivers for toe GTX2xx series, version 177.39. Then edit the nv4_disp.inf file and add an entry for device ID of 0611 (=8800GT). You will then be able to install the driver on the 8800GT. Next, install the new (also unreleased yet, but google is your friend) 8.06 software for PhysX. That's it.
Parent
Re: (Score:3, Interesting)
According to the Maximum PC Podcast [maximumpc.com] they saw significant framerate hits with single card setups, but that it was much better under SLi. They did stress that they had beta drivers, so things may drastically improve once nvidia gets final drivers out the door.
Re: (Score:2)
Are you sure that's the target audience, though?
See I've only got 1 card and I'd love hardware accelerated physics, but I sure as hell wouldn't buy a separate card for it.
Re:Works on just the one card? (Score:5, Informative)
That's not true at all. It works in a single card configuration as well. Modern GPUs have more than enough spare parallel processing power to chug away at some physics operations. Guys are already modifying the beta drivers to test it out on their Geforce 8 cards. The OP in this thread is using a single card configuration:
http://forums.overclockers.com.au/showthread.php?t=689718 [overclockers.com.au]
Parent
Kind old news (Score:2)
I called it (Score:4, Insightful)
Re:I called it (Score:5, Funny)
Parent
Re:I called it (Score:5, Funny)
Hi
We need an address for your 'Sarcastic Achievement - Level 3' certificate - you'll have to pay postage, but I'm sure you won't mind that, right?
Ned Again
COO - Sarcasm Society
Level 5 Sarcasm Ninja (certified)
Parent
Re:Linux Support (Score:4, Funny)
Parent
Re:Linux Support (Score:4, Funny)
And hopefully when it does I'll get first post in the /. article about it.
Parent
Re:Linux Support (Score:4, Funny)
Parent
Re:Linux Support (Score:5, Funny)
And hopefully the comments in the article won't all be attempts at +5, Funny.
Parent
Re:Linux Support (Score:5, Funny)
And hopefully the story wont be posted 4/1/2009.
-J
Parent
Re:Linux Support (Score:5, Funny)
Parent
Re: (Score:3, Funny)
Re:Linux Support (Score:5, Interesting)
That's not a useless comment at all unless I'm missing something. UT3 hasn't been able to put out the long-promised Linux driver because AGEIA is being so unwilling to release the license grapple hold they have over the PhysX engine. This is a legitimate concern. Unless their stance changes, Linux drivers will not be possible.
Parent
Re:Linux Support (Score:4, Interesting)
What Linux application/game uses Havok?
Parent
Re:I didn't RTFA (Score:4, Informative)
Hardware accelerated physical acceleration, gravity and particlestuff if I remember correctly, atleast old examples used to be throwing away items or exploding walls and such.
Parent
Re:I didn't RTFA (Score:5, Funny)
Mmmmm.. hardware accelerated litter..
Parent
Re:I didn't RTFA (Score:4, Interesting)
It makes City of Heroes look all awesome, particularly if you use Gravity, Storm, Kinetics or Assault Rifle power sets.
Having bullet casings, leaves, newspapers and the like drop and swirl around in response to player actions is actually pretty nifty from an immersion standpoint, particularly for a game that's essentially set in something that resembles the real, modern world.
Parent
Re:I didn't RTFA (Score:5, Funny)
"Having bullet casings, leaves, newspapers and the like drop and swirl around in response to player actions is actually pretty nifty from an immersion standpoint"
That's it. I'm done with immersion games. I'm going outside to stand in the rain. Back later.
--
BM0
Parent
Re:I didn't RTFA (Score:4, Funny)
Parent
Re: (Score:3, Insightful)
Doom is a futuristic shooter. We had it back in this thing called the 90s ;) And an RTS is an RTS. Driving games on the PC have never been quite as prolific as on consoles either.. something I used to lament, but things are improving these days.
Re:I didn't RTFA (Score:4, Interesting)
Basically exactly what it sounds like... its a real-time physics calcuating engine.
Used in games for things like shooting the limbs off of creatures, or even wind on trees, or water...
Likewise for other 3D applications, im not sure how extensive it is, or what its limitations are, but im looking forward to it, and more because calculating physic type things on most 3D software takes a lot of CPU power, so if the GPU can handle that, that takes a great load of the main CPU. (from what I would assume)
Parent
Re: (Score:3, Insightful)
But going from a little physics demo to full blown kick ass 3d game with any meaningful results is a whole 'nother matter.
Re: (Score:3, Insightful)
Re:I didn't RTFA (Score:4, Interesting)
Not exactly true, all of the Unreal Tournament Edition 3 engine games consistantly use all four cores in my Intel Q6600 with over a dozen threads spaced throughout my cores. The most notible examples would be UTE3, Bioshock and Mass Effect, 3 of the biggest games of 2007 and 2008. I can typically max out settings for UTE3 engine games.
On the other hand, performance demanding games like Crysis are total doucebags and peg just one core and sometimes using one more if it feels like it every now and then. Although it's not a very good comparison since there's so many different factors involved, I would gather to say that if crysis took an approach of optimizing better for duo and quad core cpus, their publisher would have far less complaints about performance from gamers.
Parent
Re: (Score:3, Interesting)
"What's next? "Graphic" cards with hardware accelerated AI support?"
Actually this isn't a bad idea, this is a good idea since pathfinding in games like Supreme commander is just a nightmare as you add more units, I've wondered about using the GPU for pathfinding acceleration.
Re: (Score:3, Informative)
Except that general purpose CPUs aren't really particularly great for raytracing. GPUs are simply special-purpose processors designed with raster graphics in mind. The newest fad is, of course, using all that special-purpose horsepower in more imaginative ways, but it's still a raster graphics processor at heart.
Why is it that they're raster graphic special purpose processors? Because raster dominates the playfield. What's the logical conclusion there? As soon as raytraced graphics engines start becoming
Re:PhysX? (Score:5, Informative)
http://en.wikipedia.org/wiki/PhysX [wikipedia.org]
Realtime hardware accelerated physics. Used to be on a separate expensive board which few games supported but Nvidia are implementing it on CUDA so it can run on their graphic cards instead.
Parent
Re:PhysX? (Score:4, Insightful)
nvidia bought out he company so they own it and can put it on their cards, games that decide to add support for it it will benefit nvidia.
Parent
Re: (Score:3, Informative)
The source engine, while "capable" of scaling to multiple cores, does a very poor job on current x86 chips. The games become very unstable with mat_queue_mode 2 on, and there are problems with jerky motion in any sort of latency.
It's a shame, too, because the engine works with multicore on various consoles, and it's a lot faster when it does work on PC.
Re: (Score:3, Insightful)
Re: (Score:3, Informative)
I don't need to Google. Anything built on the Unreal 3 engine has PhysX support built in.
Re:Does anyone else remember... (Score:4, Informative)
Reading comprehension...anything built on the Unreal 3 engine.
Like one of these many licensees:
http://www.unrealtechnology.com/news.php [unrealtechnology.com]
Native PhysX Support:
http://www.theinquirer.net/en/inquirer/news/2007/05/30/unreal-3-thinks-threading [theinquirer.net]
Parent
Re:Does anyone else remember... (Score:5, Insightful)
Unreal 3 is an engine that's used on LOTS of games - technically ALL of them have PhysX support, so no, not "just" Unreal 3, because there is no game called Unreal 3.
Parent
Re: (Score:3, Funny)
Re: (Score:3, Informative)
When we're talking about game worlds in which there could easily be 50 or 100 objects on the screen at once, it makes much more sense to have maybe one physics thread (separate from the render thread, and the AI thread) -- or maybe one per core. I very much doubt one real OS thread per object would work well at all.
Re:Does anyone else remember... (Score:4, Informative)
Um, except if you you have exactly 1 physics thread you have to juggle complex scheduling considerations about who needs how much CPU, handle the prioritization against the render and AI threads, handle intermixing them, etc. You have to implement a task scheduler. ... which is exactly what Quake 1 did. Carmack wrote a userspace thread library, and spawned multiple threads. Since DOS didn't have threads this worked rather well.
An OS thread will give any thread a base priority, and then raise that priority every time it passes it over in the queue when it wants CPU time. It lowers the priority to the base when it runs. If a task sleeps, it gets passed over and left at lowest priority; if it wakes up and wants CPU, it climbs the priority tree. In this way, tasks which need a lot of CPU wind up getting run regularly-- as often as possible, actually-- and when multiple ones want CPU they're split up evenly.
If you make the render thread one thread, you have to implement this logic yourself. Further, the OS will see your thread as exactly one thread, and act accordingly. If you have 10000 physics objects and 15 AIs, keeping both threads CPU-hungry, then the OS will give 1/3 CPU to the physics engine; 1/3 CPU to the AI; and 1/3 CPU to the render thread. This means your physics engine starves, and your physics start getting slow and choppy well before you reach the physical limits of the hardware. The game breaks down.
You obviously don't understand either game programming or operating systems.
Parent