• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Nvidia APEX

Z

Zogrim

n00b
Joined
Oct 30, 2006
Messages
46
Anything that uses it won't be out for a year or two.
 
This was impressive back in 2005...
It was neat that video cards were capable of it in 2006/2007.

Now, who cares? This doesn't increase the game quality at all and unless fairly high end rigs are used, very few people will actually even use this feature (to its full extent).
 
The cloth physics is nice but not feasible in a game where it would actually matter--like in an RPG. Too many character models, too many combinations of items, clothing, and attack animations. Realistic cloth physics is probably 10 years away. The funny thing is, we said the same thing in 1998 and we're still in the same damn spot. Who knows? Maybe it'll never happen.
 
astrallite said:
The cloth physics is nice but not feasible in a game where it would actually matter--like in an RPG. Too many character models, too many combinations of items, clothing, and attack animations. Realistic cloth physics is probably 10 years away. The funny thing is, we said the same thing in 1998 and we're still in the same damn spot. Who knows? Maybe it'll never happen.
Click to expand...

For your sake I hope you're joking. So in 1998 we had the same type of physics in game the we are seeing now. 10 years away? Hardly. Maybe for complete environmental simulation, but not for realistic cloth physics :rolleyes:
 
OpenCL will likely render PhysX obsolete. What has been learned and these Apex tools won't be useful unless they work for OpenCL.

OpenCL is the new standard.
 
ElMoIsEviL said:
OpenCL will likely render PhysX obsolete. What has been learned and these Apex tools won't be useful unless they work for OpenCL.

OpenCL is the new standard.
Click to expand...

What are you smoking? OpenCL is an API like PhysX. To build one on top of the other adds another layer of indirection not to mention horrendous inefficiency.

Besides, OpenCL lacks the high-level API of CUDA and therefore won't be as popular with non-techy people, such as those in research institutes who are most prone to use GPGPU to accelerate tasks.
 
Elledan said:
What are you smoking? OpenCL is an API like PhysX. To build one on top of the other adds another layer of indirection not to mention horrendous inefficiency.

Besides, OpenCL lacks the high-level API of CUDA and therefore won't be as popular with non-techy people, such as those in research institutes who are most prone to use GPGPU to accelerate tasks.
Click to expand...

I really don't like it when people who have no clue what they're talking about comment on a board.

OpenCL allows software to talk to hardware (It is not yet an API but rather a standard which will become an API in due time). Effectively it currently allows the Havok API to run on GPUs.

Havok is the API currently ported to OpenCL (a few subsets were run on OpenCL). It is now an industry standard having been adopted by all the big players. OpenCL is also open ended.

http://www.pcper.com/comments.php?nid=6954

This means that PhysX and/or CUDA subsets could also be ported to run on OpenCL (which is what nVIDIA is working on right now). OpenCL allows both GPU and CPU code to be written effectively allowing developers to use a single too in order to work on physics effects in game. As for it's GPGPU capabilities, it is also being pushed by Intel for it's larrabee project.

PhysX per-say and CUDA per-say are dead solutions. Goodnight and Good luck. It's over :) (anybody remember 3dfx Glide?
 
GPGPU is about efficiency, hence API -> hardware is the most efficient way. API -> OpenCL -> hardware is most decidedly less efficient. NVidia has admitted already that their PhysX on OpenCL implementation was too inefficient to be usable at this point. AMD's implementation of Havok on OpenCL suffered from the same inefficiency (though they were less quick to admit this, I wonder why).

I also found your comment on my knowledge of this subject to be somewhat insulting, as I have been developing with CUDA for quite a few months now, and have done research on other forms of GPGPU, including AMD's solution, OpenCL and GLSL.
 
ElMoIsEviL said:
I really don't like it when people who have no clue what they're talking about comment on a board.

OpenCL allows software to talk to hardware (It is not yet an API but rather a standard which will become an API in due time). Effectively it currently allows the Havok API to run on GPUs.
Click to expand...
Oh, the iron knee.
 
You must log in or register to reply here.
Back
Top