Computer Science Honours Project - 2009

So after spending plenty of time thinking about creating a website, I eventually got round to doing it. I started yesterday, the 23rd March on creating my project website using HTML and CSS. I knew absolutely nothing about CSS, fortunately Bad Brad told me about a website to get free CSS templates from: http://www.freecsstemplates.org/. After fiddling around with the CSS and HTML template, I managed to come up with my website. I am very pleased with how it turned out!

This post has been created to provide some insight into the purpose of my Dev Log. I plan to use this dev log as a kind of informal blog which is related to the development of my project, "High Level OpenCL Implementation". Comments that are displayed in this blog are my express opinions which are not intended for the purpose of harming anyone. So please do not take offense!

So today I presented my project to my fellow Computer Science Honours students and some Computer Science staff members. I felt as though the presentation went relatively well. I must admit that I "freaked" out a little while preparing the presentation, but actually ended understanding my project better! The presentation is available here.

So I finally managed to get some code working with the help of Shaun. It is not correct by a long shot but at least it is something! The code is available here.

I finally completed my literature survey having handed in a draft on Thursday (28th of May). It seemed like such alot of work for so few marks, but at least it will help when it is time for the final project writeup! The literarure survey is available here.

Finally exams are finished, now I need to prepare for my internship at Korbitec in Cape Town.

Shaun and I worked on getting the OpenCL libraries working again today. However we ran into a few stumbling blocks such as a pointer to a pointer, etc. Hopefully I will be able to obtain permission to use NVIDIA's implementation of OpenCL so that I can work with code which is correct and actually compiles and runs. The code is available here.

Finally got OpenCL to execute sequentially, with the kernel being dynamically linked into the executing program. Most of this was possible thanks to the use of AMD examples which came with their OpenCL implementation. My working code is available here.

So we tested the OpenCL code again with different samples and ran into a few problems. Fortunately we managed to sort these problems out and now we can run a kernel which displays a Mandelbrot Fractal! My working code is available here.

Ross and Benji showed me this interesting game called "Monkey Kick Off", my best attempt is here.

We managed to get the Mandelbrot fractal program to execute in parallel. We did notice that our implementation is not the most efficient. Hopefully we can fix this up later if there is time.

I eventually finished my implementation today in the middle of writing my thesis. Fortunately it didn't cause to many problems making the changes to devices, command queues and events. The final version of the code can be downloaded here. Once I had finished my code, I managed to do the last of my experiments, these can be downloaded here.

I presented my final project presentation to the department today. I thought it went well and Shaun said that it was very good. So hopefully my second marker has a better idea of my project now! The presentation can be downloaded here.

Finally 5pm arrived today and I handed my thesis in on time, well almost! I had 3 hours of sleep last night trying to finish it off, which I kind of did. But I am sure that I could still have done with more time! It is a relief to have finished my project, now all I hope is the effort I put in pays off.