Пікірлер
@ericO141 Жыл бұрын
This is the most niche thing the KZbin algorithm has recommended me ever
@jhanolaer8286 2 жыл бұрын
Sir, could you make a video tutorial how to make this?🙏
@LegitimaDefensaPazYLibertad 2 жыл бұрын
agustin
@Mainycrafty 7 жыл бұрын
buen video pero sube más minecraft :)
@Lenakeiz 9 жыл бұрын
Very interesting, especially because you have done in CPU. Is there ant chance to have a look at the source code?
@luisAfp9 9 жыл бұрын
Hi, nice smiulation. What did you use? CUDA?. I am working in this kind of simulation in my university. But i have problems with the render of surface
@TinoRS1917 9 жыл бұрын
No, I didn't use CUDA. All is running in CPU. For rendering i used standard openGL commands and the mesh was constructed by marching cube method. I coded wiyh openCL another fluid simulator based on eulerian method,
@luisAfp9 9 жыл бұрын
Agustín Rodríguez Serrano Excelent. I'm using Cuda. Do you have some information that you can share? Thx
@BiancaWong 9 жыл бұрын
Did you do anything to smooth the mesh? Thank you!
@TinoRS1917 9 жыл бұрын
What do you mean with "smooth the mesh" ? The mesh is rendered by openGL Thank you
@BiancaWong 9 жыл бұрын
Agustín Rodríguez Serrano Thank you for the reply. I was getting these ugly jittering effects with low resolution meshes. Yours looks so smooth! :-) So I wondered if you did anything to control the mesh quality. I followed the algorithm on Paul Bourke's website for the marching cubes implementation. - Or maybe you are doing something different with grid point sampling. How did you sample the grid point values in your implementation? Thank you for your time and have a nice day!
@qgu 10 жыл бұрын
Is it real-time?
@blinkofnight 11 жыл бұрын
Thank you, I'll check that out. I wish you luck in your research.
@TinoRS1917 11 жыл бұрын
And ... Best regards from Spain, ;-)
@TinoRS1917 11 жыл бұрын
You're right about incompressibility, ;-). This video was the first ones I've generated. Also, this simulation don't calculate the mass per particle correctly. No realtime, no GPU, everything is done on CPU. About WCSPH I've followed the Markus Becker & Matthias Teschner paper. You can seen the results in other videos from my own In any case,now, I'm researching how to couple eulerian (grid-based) and lagrangian (SPH)
@blinkofnight 11 жыл бұрын
It seem to be compressible little bit (WCSPH?), but looks nice. I also wonder whether that's realtime and whether GPU programming was invoked. Best regards from Ukraine.
@TinoRS1917 11 жыл бұрын
:A mi me costó ajustar parámetros (masa por partícula, densisdad, viscosidad, tensión superficial) para que fuera estable, el core de la simulación no tanto. Supongo que habrás visto multitud de papers al respecto, pero a mi me sirvió de mucho este libro: Smoothed Particle Hydrodynamics: A Meshfree Particle Method, de M.B. Liu
@luisAfp9 11 жыл бұрын
Esta excelente, mi tesis de pregrado es justamente hacer eso, aunque aun no lo logro :(
@TinoRS1917 12 жыл бұрын
It needs anisotropic smoothing kernels in the reconstruction of the surface, i.e., in the calculation and definition of the implicit distance function
@TinoRS1917 12 жыл бұрын
Ok, thank you vey much. I just arrived to the "CFD world" since February 2012. It's very exciting and interesting.
@VelRyphon 12 жыл бұрын
Nice work. GPU implementation will be tricky, specially concerning spatial hashing optimization. Unless you first implement brute force :)
@TinoRS1917 12 жыл бұрын
About 5hours. Yes, it runs physics simulation step on the CPU. I'll try use openMP for multithreading, and then, maybe I'll try port ing on GPU