GPGPU journey 2: Feeling my way around

Done

Setup development environment in windows
Figure out hardware requirements
Do a first compile run
Find out how to draw a 3d object

I read a lot and played around with several IDE. Sat down with someone who has a thesis on that and realized that the only way to draw/display is to use Open GL. I am having a hard time finding an example of a simple “Here, draw this value on to pixel (1,1) of your screen”. Pretty frustrating. Now that I have isolated a path, I should be able to get on with development.

To do

Setup development environment in Linux
GPGPU and GT
Sigma Delta
Find and install OpenGL IDE

Progress is slow. I have started the conversion of Verilog code into C on spare times. I was really shocked when I read the code from 5 years ago. I kept on thinking: “And I understood these? Wrote these?” There are around 1500 lines of code for the core logic and about the same amount of code for wiring and connections between components. It really brings back memories of back when I was designer. Multiple windows open for on the whim tracing of a wire through all the core logics. Like playing a game of connect the dot in the dark.

Setting up a development environment with Radeon HD2900 GT has failed miserably. Especially after installing ATI’s drivers. It looks like I will have to install a Linux partition at home to test it with the HD2400 to see if the same thing happens. I am guessing that this black screen problem is caused by using AMD’s Barcelona core with stepping #1. The graphic card basically fails completely if I try to bring up the X server. Switching back into terminal does not bring the display back. Fixing this will take too long, so I am just going to switch cards.

It looks like I will have to learn OpenGL. I have no experience what so ever in it. Can anybody suggest a placce to start?