Hi,
are there any d3d related projects for SOC 2008 ? Who is the Mentor ? How can I join ?
Best regards Artur
Hi Artur,
2008/3/14, Artur Szymiec artur.szymiec@gmail.com:
are there any d3d related projects for SOC 2008 ? Who is the Mentor ? How can I join ?
Definitely, look at the idea page to get an ide aof what we think is good. It's even better if you suggest something YOU want to do here. It doesn't have to come from the idea page.
Cheers, Maarten.
Am Freitag, 14. März 2008 20:04:38 schrieb Artur Szymiec:
Hi,
are there any d3d related projects for SOC 2008 ? Who is the Mentor ? How can I join ?
For any D3D(not d3dx) gsoc project Henri, Roderick or I would be the mentor. If you want to do a project this list is a good place to announce your interest, later when the student registration is open you can officially apply at google's SOC pages.
I have a few rough ideas which I think might work:
*) Fixed function pipeline replacement shader using ARB shaders or GLSL *) Pixel shader 1.x support using GL_ATI_fragment_shader(for shader support on ATI r200,r250 and r280 cards) *) Pixel shader 1.x support using GL_NV_texture_shader and GL_NV_register_combiners(For shader support on Geforce 3 and Geforce 4 cards) *) Get native d3drm.dll running. That's a good pool for finding ddraw/d3d7 bugs.
A few othe ideas which might be hard for someone who doesn't know the code yet:
*) Prepare WineD3D for OpenGL 3(is that even out yet???) by splitting opengl rendering code from general management code. That is done already for surfaces, but needs to be done for IWineD3DDevice and buffers
*) Move WineD3D to a d3d10-like architecture: Restructure the textures, surfaces, index and vertex buffers around the D3D10/OpenGL resource model. To make use of that implement ATIs R2VB(render to vertex buffer) hack in d3d9.dll.
Maybe also worth a consideration: *) Take an advanced game(e.g. Team Fortress 2 in dxlevel 95, Bioshock or Crysis) and a graphics driver which doesn't yet work as intended(e.g. fglrx or any MacOS driver) and isolate the problems that prevent it from working on the chosen driver. Then either find workarounds or report the bug to the driver vendor. The problem here is that we cannot find a timeframe for a debugging project. It could be done withhin 5 minutes or take years.
*) Similarly, in cooperation with the Xorg guys: Find the bug that causes those pesky random memory corruptions when using WineD3D with Mesa/DRI drivers.