|
I arrived in Providence on Monday evening. Tuesday was my first day at Brown. I met Dr. Jennifer Dworak, and I feel very fortunate to have been paired with her. I got my key to the computer lab and my Brown ID card (my key to everything else). I started getting acquainted with Fastscan, our circuit simulation program. I started out with a large circuit with a couple hundred logic gates. This was definitely not the best choice. I spent some time being really confused before I decided to take it slow, until I got more familiar with the software. I took it down a notch and continued with a very small circuit, with only 5 inputs and 6 logic gates--I definitely thought it would make things easier. It did--somewhat. I still had to go through the 800+ page manual to find the commands I needed (of course, I didn't read the whole thing...just a LOT of searching). Fastscan is starting to become more familiar. |
|
|
Since I was more familiar with the software, I began using Fastscan to get some data for my project. Fastscan will automatically generate input patterns for the circuit simulation. It took awhile, but I finally figured out how to make it use many random patterns, rather than just the few Fastscan "decided" were the most effective for fault detection. Fastscan also allows me to set constraints on the circuit inputs. I can tell it which of my inputs are to be known and which are to be unknown. I can also view the output of each logic gate within the circuit (these are the interior values my project concerns). It displays all of these output values either on the command line screen or in an output file. For the small circuit with 6 gates, it was plausible to go through this data by hand, and determine what it meant for my project. However, this is a teeny tiny circuit. With only 12 gates and slightly more than 5 inputs, the amount of data to go through grows very quickly. I quickly saw that if I wanted to use this same method on even a medium sized circuit, it would take all summer! I needed a solution. It's called: Perl. I need to write a perl script to parse these huge output files, so I can actually get something useful out of the simulations. The problem is, I've never learned Perl. Well, I have to start somewhere...hopefully it doesn't take too long! |
|
