Author: Danny Mathieson

Goals for the Week:

• Integrate the final automation calculation with writing to BRAM storage for output to a monitor through VGA
• Begin working with multiple Pmod amplifiers
• Presentation

Progress:

• Calculation hardware is finished
• Integrations has begun, but isn’t as far as we would like it to be at this point
  • We gave ourselves an extra week just in case we fell a little behind like this so we are not worried about the deadline
• We have previous experience with the Pmod amplifiers, and have an idea on how we could use multiple amps at the same time
  • This would involve multiplexing the output between which amp we want to use and when
• We have finished researching which automation rules have which properties, and once integration is complete we can see which rules provide interesting visual and audio effects

Current Challenges:

• There are some bugs in integration between the final calculation and writing the final values to the BRAM
  • Not too worried here because we have dealt with similar problems before.  This just takes time
• Calculation Speed
  • I thought that I would be able to increase the speed tremendously by working with hardware instead of my original software implementation, but it isn’t really that much faster.  I will continue to try to improve this once we have a working design
• Playing multiple Pmod notes at the same time
  • This is the difficult challenge and will take most of my time in the coming week

Goals For Next Week:

• Work out the bugs we are currently experiencing in integration
• Work with multiple Pmod setup
• Website – For expo
  • Post pages on interesting rule algorithms in cellular automation
  • Post design strategy page
  • Work on overall aesthetic design
• (If Time) Improve calculation speed

Goals for the Week:

• Finish Cellular Automation Hardware design in Verilog
• Create mid-project presentation

Progress:

• I still have not finished the hardware for automation
  • Calculating border conditions in hardware is much less trivial than doing so in software
  • Running the calculations in parallel is more difficult than what we did for the julia-set creation
• The presentation will be finished shortly

Current Challenge:

• Running a parallel calculation with multiple values that are subject to previous change is causing me the following problems:
  • Timing.  I am trying to optimize the amount of time needed for each generation to be computed by having hardware specific to each cell.
  • Hardware usage.  I am using specific hardware for each cell, and each cell also has 2 flip-flops per cell.  Also, because cells all depend on the previous values of other cells, wires between previous flip flop values are required as well.
• I don’t see these as huge problems, but I will bring them up in our presentation tomorrow to get ideas from the rest of the class on how to solve them

Plan for next week:

• Finish the hardware implementation by Wednesday.
• Begin implementing with Sarah’s BRAM on Wednesday, and begin displaying the output of the cellular automation via VGA
• Spend the weekend debugging my calculation implementation in hardware
  • I won’t be able to debug except for synthesize errors because I won’t be able to see the output until then

Quick Note: Hopefully this ends up in the right places…

Goals for the week:

• Start working on creating a cellular automation creator with a  preset starting state.
• Possibly begin BRAM integration

Progress:

• I didn’t know how I was going to test it with a hardware implementation, so I implemented the automation in C
• I didn’t get the results that I expected, but this is fine as the hardware version will be easier to implement due to the case sensitive nature of the the calculations
• Hardware implementation has begun as well, but I am not very far into it.

Current Challenge:

• I am not 100% sure still that the way that I am calculating the values for each cell is entirely correct.
  • This is because none of the research that I have done supports whether every cell is computed one after the other with a newly changed cell affecting the cell to its immediate right, or if all of the cells are to be computed in parallel (much like a hardware system would do anyway) using the states of the previous cell to its immediate left.
  • Border conditions are also troublesome to compute, but with the way that I am computing them (using only itself and its only neighboring cell) This computation is going quite smoothly
• It will take further testing to confirm whehter my calculations are being done correctly, but these can be minor fixes during implementation in the coming weeks.

Plan for next week:

• Finish hardware implementation of calculations and have it program the board without errors
• If time allows, coordinate with Sarah and begin integration of the BRAM parts and the automation, hopefully resulting with a finished cellular automation through VGA displayed on a monitor

I just wrote an info page on Elementary Cellular Automation.  Please respond if you want more information on the subject and I’ll add to what I wrote.