Re: Small volumetric path tracer back
Board:
Home
Board index
Raytracing
Visuals, Tools, Demos & Sources
(L) [2013/03/12] [dbz] [Re: Small volumetric path tracer] Wayback!Interesting. How is the performance on cuda?
(L) [2013/03/12] [D-POWER] [Re: Small volumetric path tracer] Wayback!It depends on the scene, but on average there is 8 to 9 times speedup on my nvidia GT220
(L) [2013/03/12] [spectral] [Re: Small volumetric path tracer] Wayback!Looks Nice,
And have you also fixed the version on the git repository ? Or upload the cuda version ? [SMILEY :-)]
(L) [2013/03/12] [D-POWER] [Re: Small volumetric path tracer] Wayback!I have just upload the new source code. Though, keep in mind that the source code is still experimental and lacks proper optimization, considering that my cuda knowledge is quite limited
EDIT: here is the link to the github repository [LINK https://github.com/D-POWER/vpath]
(L) [2013/03/12] [spectral] [Re: Small volumetric path tracer] Wayback!Sure,
It is just to give a try and review...
Thx
(L) [2013/03/13] [D-POWER] [Re: Small volumetric path tracer] Wayback!Any suggestions/improvements is very welcome. I'm always looking for techniques to squeeze more speed [SMILEY :)]
(L) [2013/03/13] [spectral] [Re: Small volumetric path tracer] Wayback!I haven't take a closer look,
But this project seems to implement SSS to and support more kind of primitives : [LINK http://gpupathtracer.blogspot.be/]
Maybe take a look [SMILEY ;-)]
(L) [2013/03/13] [D-POWER] [Re: Small volumetric path tracer] Wayback!Yes, I'm aware of Peter and Karl's gpu path tracer, actually it inspired me a lot in writing my own volumetric path tracer
(L) [2013/03/15] [ost
by D-POWER] [Re: Small volumetric path tracer] Wayback!Testing volumetric caustics
image192605.png
(L) [2013/04/10] [Dade] [Re: Small volumetric path tracer] Wayback!D-Power, I hope you don't mind I borrowed your idea of applying scattering to SmallPT and I added this features to the good old SmallPTGPU (an OpenCL version of Kevin Beason's SmallPT available here [LINK http://code.google.com/p/ocltoys/]):
[IMG #1 Image]
[IMG #2 Image]
[IMG #1]:
![[IMG:#0]](images/7b108bd2cbfc87c876651e17e95e6608718f613f29549e3917e2e111edaeba13.jpg)
[IMG #2]:
![[IMG:#1]](images/e361ba554b44e362c028faa26adc745d511a952641d35c34c94d3010f9311483.jpg)
(L) [2013/04/11] [tby D-POWER] [Re: Small volumetric path tracer] Wayback!Of course I don't mind. That's some very cool renders you've got there [SMILEY :)]
I think you should also update this page: [LINK http://davibu.interfree.it/opencl/smallptgpu/smallptGPU.html]
(L) [2013/04/11] [tby D-POWER] [Re: Small volumetric path tracer] Wayback!By the way, what's the maximum depth did you use when rendering SSS? In my implementation [LINK https://github.com/D-POWER/vpath], I noticed that small values didn't work and I needed to use like 100 or even 200 as the maximum recursion depth.
Thanks
(L) [2013/04/11] [tby Dade] [Re: Small volumetric path tracer] Wayback!>> D-POWER wrote:By the way, what's the maximum depth did you use when rendering SSS? In my implementation [LINK https://github.com/D-POWER/vpath], I noticed that small values didn't work and I needed to use like 100 or even 200 as the maximum recursion depth.
Strange, I get good results even with a max. depth of 6-8. I'm using a glossy reflection (hard-coded to white color) and glossy transmission BSDF. Scattering/absorption is enabled when you get into the sphere and disabled when you get out.
back