Water! PhysX Position Based Dynamics

[Davros]

I would like to see a performance comparison chart.

 

[Priyadarshi]

Thanks for the link. I found one more review which is interesting : http://www.youtube.com/watch?v=s5IxYyeppU0

Metro 2033 benchmark results :

GTX 680 + GTX 560Ti (PhysX with PCIe x16) : 61 fps
GTX 660 + GTX 560Ti (PhysX with PCIe x8) : 54 fps
GTX 680 (Render + PhysX) : 52 fps

PCIe seems to be the main bottleneck in this case as performance scales more with bandwidth changes. Since there is a performance improvement by adding a dedicated device I think that there is an overlap between memory transfers and computations. This will be different across different applications though.

 

[jlippo]

Very impressive, would love to see this in games.

 

[gamervivek]

Context switching between CUDA and Graphics should only be a matter of microseconds as you mentioned.

well yeah, anything can happen in a matter of microseconds.

In one frame, the context switching only needs to happen once. Do the CUDA stuff first and then render your results. This is a lot less than transferring your results over the PCIe bus though.

and switching over to CUDA at the start of the frame doesn't incur any performance penalty? or there is simply no switching involved?

since pcie transfers are involved with a dedicated card, it might be illuminating to know how the frame times compare and the overall smoothness of gameplay.

 

[Malo]

Looks fantastic, will enjoy seeing it in games 5+ years from now.

 

[Priyadarshi]

well yeah, anything can happen in a matter of microseconds

Ahemm. Reaction time.

and switching over to CUDA at the start of the frame doesn't incur any performance penalty? or there is simply no switching involved?

Yes that is again one switch. It should be around 10-20 microseconds I have heard. Modern games do context switch between Compute Shader and Graphics so it's not a big deal.

since pcie transfers are involved with a dedicated card, it might be illuminating to know how the frame times compare and the overall smoothness of gameplay.

This should be FCATed I concur.

 

[Davros]

I think that there is an overlap between memory transfers and computations.

There is a lot of stuff ihv's can do to hide latency

 

[Arwin]

Looks fantastic, will enjoy seeing it in games 5+ years from now.

Or, seeing what they did in Uncharted 3, perhaps it will appear somewhere in the next Uncharted, like a not dissimilar type of simulation that's somewhere in a little swimming pool.

 

[almighty]

Thanks for the link. I found one more review which is interesting : http://www.youtube.com/watch?v=s5IxYyeppU0

Metro 2033 benchmark results :

GTX 680 + GTX 560Ti (PhysX with PCIe x16) : 61 fps
GTX 660 + GTX 560Ti (PhysX with PCIe x8) : 54 fps
GTX 680 (Render + PhysX) : 52 fps

PCIe seems to be the main bottleneck in this case as performance scales more with bandwidth changes. Since there is a performance improvement by adding a dedicated device I think that there is an overlap between memory transfers and computations. This will be different across different applications though.

No the performance is from running a GTX 680 over a GTX 660....

You clearly lack any practical know how on this subject so best to leave it at that as everything you've said is completely wrong.

 

[Brad Grenz]

That 660 is a typo. In the link it's only a 680 being used.

 

[gamervivek]

Ahemm. Reaction time.

of course nvidia could have been like totally willing to sell another card over streamlining their code.
Probably that concurrent kernel execution on fermi helped too, nvidia went ballyhoo over it while ati had had that for a long time, at least in the rv770, was looking into it for overcoming the hd48xx disadvantage in LDS when compared to gtx285.
Apparently it wasn't exposed in their software stack, should've known.

ati support

Yes that is again one switch. It should be around 10-20 microseconds I have heard. Modern games do context switch between Compute Shader and Graphics so it's not a big deal.

but they are part of the same API

This should be FCATed I concur.

took a look at hd7990 review yesterday, nvidia are doing too well on that front.

http://www.hardwarecanucks.com/foru...60843-amd-hd-7990-review-malta-arrives-7.html

There is a lot of stuff ihv's can do to hide latency

he used to do something like that for one of them. Dunno if he has changed teams or is still playing lately.

No the performance is from running a GTX 680 over a GTX 660....

You clearly lack any practical know how on this subject so best to leave it at that as everything you've said is completely wrong.

lol no

 

[fellix]

 

[KKRT]

Yeah, i had the same reaction when i saw performance

 

[pharma]

Nvidia Feature advancements - PhysX FleX

PhysX FleX is a particle based simulation technique for real-time visual effects. So far we showed examples for rigid body stacking, particle piles, soft bodies and fluids. In this video we'll be showing some of the new advancements within PhysX FleX.

 

[homerdog]

Really wish smoke effects in games looked even 1/10th as good as that.

 

[Scott_Arm]

Very cool. I've been desperately awaiting a cereal simulator.

 

[Billy Idol]

Are they using smoothed particle hydrodynamics (SPH) methods here?

 

[London Geezer]

Damn! Differents dev have been showing similar demos for bloody centuries but not a single game has showed something like that in real time.

 

[Malo]

Doesn't Black Flag have interactive smoke like that?

The cereal demo was seriously impressive though. Needs to be added to I Am Bread.

 

[AlNom]

TryX are for kids.