A wealth of information which could be very useful for Cycles development.

Or you could say, a possible treasure trove of information I stumbled upon which could greatly benefit those who are interested in developing and optimizing the Cycles engine. (which could be good for the Mango team as well)

The recent articles include two-way pathtracing, GPU-compatible bidirectional path tracing, portals, and more.

So if you’re interested in developing Cycles or are a curious user, go ahead and check it out.

And maybe we can even get in contact with this dev. to help with the Cycles project the same way we contacted Mike Farny, which would be even better:cool:

Thanks for sharing, very useful, luckily i open random paper and hit exact that part i have trouble to do proper MIS.

The key idea allowing us to use both CPU and GPU effi-
ciently is to sample populations of NC camera paths and NL
light paths independently on CPU, and then combine each
camera path with each light path.

Damn that guy is faster, that was one of my sectet plans how to smash Arion/Octane and other who dare to compete, need to invernt something even more crazy.

Don’t go too crazy, otherwise we may never see your work get into trunk

Just fix up and polish the implementations you currently have and submit the patches to Brecht for inclusion into trunk, then you can think about crazier ideas.:wink:

just out of curiosity: why peolple like this Simon (people willing to “play” with advanced pathtracing algorithms) always start up a new renderer? I mean, he’s currently bounded to spheres and cubes but already has bi-dir stuff going on, perfect caustics… etc. Wouldn’t it be easier for a coder to pick up (uhm… let’s say) Blender which has already a lot of rendering code already done (bvh, meshes and stuff) and start playing on top of that? (yeah, I guess, that the “lot of rendering code already done” actually might be a problem coding-wise).
I’m just asking because it doesn’t seem like he’s building his own new big renderer, but instead he’s just willing to implement on GPU all that eyecandy we wish cycles has…

oh, btw, i’m a optimistic fan of storm’s work :wink:

oh, btw, i’m a optimistic fan of storm’s work :wink:

+1 to that entire post, but extra for that part :slight_smile:

You would think these devs would want to see people actually using their hard work, but maybe some care more about the challenge than the end product.

Yep, maybe. I only wish sometimes we could (as we say in my country) “catch two pigeons with a single fava bean”, which stands for internet slang “a win-win

hey Simon, c’mon on board! :slight_smile:

Someone should make Simon an offer he can’t refuse :stuck_out_tongue:

Nah but it would be cool if we could recruit him to work on the Cycles project, I would be willing to donate to the cause if that’s the case :slight_smile:

That won’t happen: Simon already commented about Cycles here: http://www.sjbrown.co.uk/about/#comment-24272