LifeFullOfGames

The project showcase went very well, even more than I expected! It seems like the project had proved itself to be quite interesting even for the professional game developers, I'm so thrilled! :D

Today I have met so many people. Since the showcase started I don't actually have time to break which I'm so happy. there are our tutors, classmates, friends from RTW, students and parents come visiting and listen to my presentation. Hopefully my work will inspired some young people as similar as I had from the work of the previous generation.

Dissertation Update

• All main sections have been done, waiting for last review form the technical supervisor
• Appendices and CD (DVD) Companion should be done in a next few days
• Don't forget to check with the print shop, how long it gonna take for binding and printing!

22nd May Showcase

• Poster need to be made in A2 size ( 4 - A4 Slides could do)
• Will bring my laptop for presentation ( videos and more...)
• DirectX 10 PC is essential!

Activities Update

• My first completed draft of the dissertation has been finished, and already been sent out for a peer reviews
• This week, the plan is to clean up the prototype, and try some new ideas for optimization.
• Implement "Full screen" mode selection, and XML configuration script
• Try to finish a first draft of the final presentation before end of this week

Activities Update

• This week main task is still pretty much the same, trying to finish the dissertation and submit to my supervisor regularly
• The introduction, literature review, methodology and experiments have been made solid. Only reorganizing might be requested
• Critical discussion, conclusion and abstracts will be the next
• Have a rough layout, and video for the final presentation ready

Activities Update

• Roughly, the first few sections on the dissertation have been in-placed. Some reviews and basic knowledge has been organized in the way I like to present and this draft version has already made their way to my supervisor for the first proof
• Writing a good dissertation might take longer than I expected. Good grief, that I've already had the application that could be used in the showcase.
• Keep on writing, and finish it as planned

Activities Update

• Keep working as planned, this will definitely help avoiding the coming crunch time.
• Until the next few weeks, the main activity will be the dissertation writing. So far, all evidiences and research papers are in-placed. The hardest part will be writing them up into an academic research format
• This week I will spend time doing the "introduction" and "literature reviews" section, around 3000 words as the rough target
• It should be enough experiments that I've prepared in order to prove my research methodology
• It's also good to know that at least I have done many other works regarding to this busy time, and managed to set them aside for a while. This week will be dedicated only to the dissertation!
• XML Configuration loader has been already tested and come out with a very impressive result. After the first draft of the dissertation, I will integrate some XML script to the final application to make things more controllable

Project update

• Improved Rain Rendering, textured
• Imporved Scene Lighting, Shading and Textures on the army are more correct
• Improved Grass Rendering, draw more grass blade repeatly onto the bigger billboard (3x)
  
• The grass billboard will be resized as their LOD changing. This has been done under Geometry Shader stage where I tried to implement "Early Z-Test" to reduce the Pixel Shader load.
• Add simple FPS calculation, and print screen function
  

Videos

• 110k Animated Armymen, 110k GPU Rain Particles and 110k Animated Grassblades (updated 31 Mar 09)
  
  Adobe Flash Player not installed or older than 9.0.115!
  

Performance Update

• From last week, we had implemented a number of new stuffs that have been mentioned above. After remeasure the GPU load, we found the GPU has to do a lot more works on the PS Stage.
• Fragment Processing (Texturing, Pixel Shader and Frame buffer) become the system bottleneck
• The size of GPU billboard is effectively related to the scene performance. 3-5x bigger billboard can effect 6-7 fps
  
• As I thought, the bigger billboard will result more fragments need to be filled, and this is clearly the large performance hit.
• The solution, I found that we could perform the "Early Z test" in Geometry stage and adapt the billboard size based on the distance from the viewer
• Smaller billboard size will produce a lot less fragments that need to be processed, and the result FPS has been bumped back to normal (~30fps)
• Frame buffer stage, reducing the viewport size has slightly effected the rendering performance as well. 1024p -> 800p help improve 2-3 fps when the post processing is on.