Monday, November 30, 2009

Have to leave soon for work. But want to still get a quick post up as I am not doing many during the week. Hung out with ML over the weekend and we played around with scratch holograms. Pretty cool stuff. Fairly crazy that it works.

Other than that just working on some programming stuff. Truing to get better with openGL and the such. Need to rewrite the graphics engine for Game of life so that it uses vector buffer objects instead of the depricated immediate vertex display mode. Won't be terribly difficult, but will take a little thought. Think the easiest way is instead of changing the verticies based on the state of the board, will change the display color. The only thing is that means that each cell will actually be drawn, while right now some are and some aren't. Think that will be easier than rewriting both the vertex and color vectors each tick/cycle of the 'game'. Especially since that will involve destroying a dynamically allocated array, and creating a new one each time. With changing the colors it just means changing the contents of the color vector. Will have to see. Probably start working on that and a few other little ones after work.

Sunday, November 22, 2009

Ok, the simple toroidal board version also works now...... Think the cool interface is where it actually looks like a torus will take a long long time to do.

Saturday, November 21, 2009

Wrote an implementation today of John Conway's Game of Life with openGL for the display. It is pretty cool Right now it uses a simple flat board. I want to add the option to choice a toroidal board. Also would be cool if that board could also be displayed as a torus and the viewpoint in space could be navigated with the mouse. That however will be much more difficult; I am going to need much more graphics programming experience before that takes place. Though this is pretty cool. A vast improvement compared to the one I wrote back in high school that used ASCII art to display the state of the board and the such. Think this one might even be shorter. Wish I still had the code somewhere from my old version. I know I had printed out the source way back when, but I doubt it is still around. Have a bunch of things I want to do to enhance this though. Everything from the data structure (using a simple set of 2x 2D arrays right now), graphics, options, and add some openMP segments as well to utilize multicore processors. Not that it really needs to go faster. I had to add in a 'do nothing' type function to slow it down a bit so that it is a bit less of a blur on the screen. I want to do something better with that though. So the original idea of using this as a practice project for CUDA is sort of being scrapped. Can't even imagine the size board I would need to require that much multithreaded performance.

Monday, November 16, 2009

Spent the weekend working on some openGL tutorial programs, and playing around with the code a bit. Did the basic stuff like drawing a circle with lines, and a spiral. Also worked on a version of a squire that you can move around the screen with the WASD keys as direction keys. Have to now take the stuff I learned from that and apply it to the simple version of Game of Life I wrote. Want to change a few things. Going to probably switch from an array of bools to a bitmap using an array of 64bit unsigned ints.

This will be important when I write the CUDA version. Also I think there is a way that I can use that directly to define the image (if each cell is represented by a single pixel) without doing a memory transfer from the graphics card to main memory for processing with openGL. Just have to read up a bit on it. That would be so critical with such large datasets, since that would very quickly become a bottleneck.

Also should get off my arse and do some more arduino stuff. Haven't done much of that lately.

Monday, November 9, 2009

Have ganglia up and running now. have to work on some of the other tools i installed, get them all configured, then start building compute nodes. Should be interesting.

Tuesday, November 3, 2009

I can actually see my desk now..... and dual monitors are great when they are not on different desks. I am very pleased with how things are coming together, even though I am still a bit short on space and need pick up some 3 or 4 U rackmountable cases. Can probably take the NTI KVM out of the rack as well, give me some more space there.

Monday, November 2, 2009

Looks like I am going to have to bite the bullet and finally learn openGL for the program I am working on right now. From what I have heard it is going to be painful, but well... need to actually display the results in graphic form. Figure it cant be more mind bending than CUDA. None of the "wait, how many threads did that line of code just make? where is a calculator...." situation.

Sunday, November 1, 2009

CUDA workstation

The core2 duo workstation is actually running correctly now. Still a couple minor things to deal with, but overall, I am pleased. It is running Ubuntu 9.10, with the latest version of the cuda toolkit. I ran some test code, and it works. The next steps probably will involve a few upgrades, as well as actually utilizing the system to practice doing cuda programming. It should be an adventure.