48 reputation
3
bio website
location
age
visits member for 1 year
seen Nov 1 '13 at 14:08

Aug
14
comment GPGPU physics & game logic
That's what I'm afraid of, I found two references which both seem to support this story: developer.nvidia.com/content/… and gamedev.stackexchange.com/questions/51309/… Still some links in the discussion below Jovan's answer seems to show some promise. I guess we have to wait and see what the author of bullet physics comes up with..
Aug
14
comment GPGPU physics & game logic
@Alrikai: the problem is that I did.. all cpu physics hit a wall when simulating many objects (for example 1000's objects because of dynamic debris (with gameplay interaction), or 2 way coupling of rigidbodies and sph fluid).. the second problem (why I posted this question) is with gpu physics.. there aren't any actual implementations around to test around with in a game environment. There are these papers popping up but they all seem to circumvent the question of how it will play out when there's the need to synchronize results.. :/
Aug
14
awarded  Supporter
Aug
14
comment GPGPU physics & game logic
@Alrikai: the numbers dont look too bad for small transfers from the report Jovan linked right?
Aug
14
comment GPGPU physics & game logic
@Jovan: Thanks, that actually looks very promising, with a transfer of 8mb incurring about ~11 mircosecond latency!
Aug
14
comment GPGPU physics & game logic
@Alrikai: yes, I imagine the 1 million objects the slides I linked talk about take some time to copy over the bus.. and then there is also the latency of sending inputs to the physics sim I guess?
Aug
14
comment GPGPU physics & game logic
I understand there's a turning point, if simulating 1 million objects on the gpu take 0.01 sec + 0.1 sec synchronization latency vs 1 second sim time on the cpu thats a clear win.. but still not practical for a game where we need to simulate faster than 10hz).. so what latency can we expect on a moderatly modern pc architecture?
Aug
14
asked GPGPU physics & game logic
Aug
14
answered Box2D: Efficient method to create b2Chainshapes for a tile-based map?
Aug
13
awarded  Scholar
Aug
13
accepted Symmetry preserving path simplification
Aug
9
comment Symmetry preserving path simplification
Yes that sounds like the logical thing to do.. Do you know of any methods to find line segments that could be candidates (exposing certain symmetry like features), or would this be too domain specific and implemented per case?
Aug
9
awarded  Student
Aug
9
asked Symmetry preserving path simplification