John Calsbeek

Los Angeles, CA

johncalsbeek.com

Age: 27

I’m interested in physics, data-oriented design, parallelism, lock-free programming, SIMD programming, cache optimization, compilers, and procedural content generation.

Feb
5
awarded Good Answer
Feb
1
awarded Popular Question
Dec
13
awarded Guru
Dec
12
revised What does the MSBuild -> (arrow/dash greater/->) operator do?
added a bit of extra info
Dec
12
answered What does the MSBuild -> (arrow/dash greater/->) operator do?
Dec
4
comment How much processing power does stenciling actually save?
@Llamageddon The canonical example for stencil these days is discarding all visible sky when doing a lighting full-screen pass. Lighting is expensive, and the sky often large.
Dec
3
comment How much processing power does stenciling actually save?
@Llamageddon Pixel shaders are always packed together in 2x2 quads, because a pixel shader that doesn't sample at least one texture is a rare thing indeed. But if you are doing depth-only rendering, that matters less. Also, for what it's worth, shadow rendering is traditionally bound by vertices, so in traditional high-vertex-count situations this probably won't help. Also don't forget that you'd be writing to every pixel's stencil just to skip writing to that same pixel's depth (unless your pixel shader is complex).
Nov
25
comment With modern fillrates and deferred rendering, is occlussion culling still relevant?
@Llamageddon The chat room for this site chat.stackexchange.com/rooms/26589/the-cornell-box is probably a good venue.
Nov
24
comment With modern fillrates and deferred rendering, is occlussion culling still relevant?
@Llamageddon Generally you just cull them against a conservative volume. Or several smaller volumes that you deform with the mesh (for example, when skinning you can attach culling volumes to joints).
Nov
24
comment With modern fillrates and deferred rendering, is occlussion culling still relevant?
@Llamageddon On modern GPU architectures, the vertex shader is always run, and is perfectly capable of deforming meshes—that's just transforming vertices non-rigidly. A more expensive vertex shader that does more work is, of course, more work that will be skipped by culling. I don't have any papers on GPU occlusion culling, academia seems to not be very enthralled with it. It is not a built-in feature of GPUs, just a creative use of compute.
Nov
23
answered With modern fillrates and deferred rendering, is occlussion culling still relevant?
Nov
8
accepted How does Smith multiple scattering interact with diffuse subsurface scattering?
Nov
5
comment Are draw calls executed in parallel or sequentially or both?
It's entirely possible to run the pixel shader for overlapping triangles at the same time. All that matters is that the blending happens sequentially. Pixel shaders can kick off a lot of work to the blending unit in any order, which can then reorder all those blend operations depending on the triangle that it came from.
Nov
4
awarded Yearling
Nov
4
awarded Yearling
Nov
4
asked How does Smith multiple scattering interact with diffuse subsurface scattering?
Nov
3
reviewed Approve suggested edit on Why integrate over a hemisphere (and not a sphere) to solve the rendering equation?
Nov
3
revised How can I generate procedural noise on a sphere?
added a link to Spherical Fibonacci Mapping
Oct
22
revised C++: 2 templates vs. 1 template specialization on complex class
added 10 characters in body
Oct
22
answered C++: 2 templates vs. 1 template specialization on complex class
1 2 3 4 5