so i came across this video on youtube that was posted yesterday, its about a new physics tech coming from nvidia, and i would like to know what you guys think of this? i know it probably wont be implemented in vam2, but it may be possible way way down the road.
-
Hi Guest!
We are extremely excited to announce the release of our first Beta1.1 and the first release of our Public AddonKit!
To participate in the Beta, a subscription to the Entertainer or Creator Tier is required. For access to the Public AddonKit you must be a Creator tier member. Once subscribed, download instructions can be found here.
Click here for information and guides regarding the VaM2 beta. Join our Discord server for more announcements and community discussion about VaM2. -
Hi Guest!
VaM2 Resource Categories have now been added to the Hub! For information on posting VaM2 resources and details about VaM2 related changes to our Community Forums, please see our official announcement here.
VaM 1.x New Nvidia tech would be awesome in vam
Threads regarding the original VaM 1.x
- Joined
- Jun 1, 2020
- Messages
- 2,116
- Solutions
- 75
- Reactions
Nope. The paper shows the frame time on several examples, it ranges between 6ms to 234ms for the frametime (234ms is for the cloth sheets), this means the cloth sheets demo runs at a frametime 10 times slower than what is needed for a somewhat acceptable 60fps in a game. BUT, with nothing in the scene, no UI, no skeleton physics, no AI, no logic... this means, that with an actual game behind it, it's gonna be even worse.
So uhh... yeah, if "way way down the road" is "in 25 years", maybe
- Joined
- May 11, 2020
- Messages
- 798
- Solutions
- 21
- Reactions
Besides, it claims "penetration free simulation"...
Not suitable for Vam! Lol.
Seriously, though, while you're right about the frame times in the video, @hazmhox , that's for cloth with 60,000 plus vertices. In Vam, it's recommended to keep clothing verts below, what, 1500 or 2000? So applied to Vam, the framerates would be much higher.
Not suitable for Vam! Lol.
Seriously, though, while you're right about the frame times in the video, @hazmhox , that's for cloth with 60,000 plus vertices. In Vam, it's recommended to keep clothing verts below, what, 1500 or 2000? So applied to Vam, the framerates would be much higher.
Last edited:
- Joined
- Jun 1, 2020
- Messages
- 2,116
- Solutions
- 75
- Reactions
I wish @DJ !
To take a basic example: if you do a basic scene in VAM 1.x, with a dude and a girl and only glutes/gen soft bodies, the framerate is somewhat "ok tier" in VR.
Now, do the exact same with two girls cisoring, you will drop down to an abysmal 40/20 depending on your CPU. Nothing changed on average, you just have two soft bodies colliding which each other and this fraks the perf massively.
These papers and techs are often way above the realtime they claim. Like for instance, I don't know if you know Embergen, but it's considered "realtime". Yeah, indeed it is, with a somewhat basic simulation setup and only with that in the scene. But all the mega fancy videos the show up in their trailer with massive simulations, are definitely not running at a what I call "proper realtime framerate" ( 60fps locked and above )
It's the same here. I'm not saying you can't use that paper to improve your realtime sim for a game which would make it better. But a 1:1 exact replica at that quality for a runtime game like VAM, definitely not. And even then, multi-layered physics have and extremely high overhead, mostly why pretty much all games do baked sims or single layer sims ( cape/drape/hairs only etc, and they don't interact with each other ).
I hope it'll be a reality someday... but you know, just doing devil's advocate because people talking about these (like 2min papers, goddamn I can't stand the way that dude speaks), don't often try to contextualize it properly for a use. (and then everyone thinks because there's the Nvidia logo, tomorrow games can have this exactly).
To take a basic example: if you do a basic scene in VAM 1.x, with a dude and a girl and only glutes/gen soft bodies, the framerate is somewhat "ok tier" in VR.
Now, do the exact same with two girls cisoring, you will drop down to an abysmal 40/20 depending on your CPU. Nothing changed on average, you just have two soft bodies colliding which each other and this fraks the perf massively.
These papers and techs are often way above the realtime they claim. Like for instance, I don't know if you know Embergen, but it's considered "realtime". Yeah, indeed it is, with a somewhat basic simulation setup and only with that in the scene. But all the mega fancy videos the show up in their trailer with massive simulations, are definitely not running at a what I call "proper realtime framerate" ( 60fps locked and above )
It's the same here. I'm not saying you can't use that paper to improve your realtime sim for a game which would make it better. But a 1:1 exact replica at that quality for a runtime game like VAM, definitely not. And even then, multi-layered physics have and extremely high overhead, mostly why pretty much all games do baked sims or single layer sims ( cape/drape/hairs only etc, and they don't interact with each other ).
I hope it'll be a reality someday... but you know, just doing devil's advocate because people talking about these (like 2min papers, goddamn I can't stand the way that dude speaks), don't often try to contextualize it properly for a use. (and then everyone thinks because there's the Nvidia logo, tomorrow games can have this exactly).
Similar threads
