Summer 2008 DMP Project at Brown

Past the halfway mark

I have now been here for more than half of my time. Crazy to think of! Now really got to get cranking on the project to get an amazing result. Today I found out how to deal with bridges... in good simulation mode! Oh the joys of fastscan. I will simulate bridges by knowing the values for each pattern at each wire. I will then see what would happen if an AND or OR bridge was added in those places, and if it would change the result. I can't do this for the whole circuit, so I will do it for a series of 'non feedback bridges'. One of the grad students has such a list and I accidentally asked her for the wires that are next to each other rather than the wires that are not connected before... woops! But I am hoping to finish with the bridge fiasco by the end of the week. :)


Waiting for data

So I have written all the scripts I will need to parse the data when I know the wires that I need to connect. I will check to see if each vector finds the fault in the wire. Then I will generate a vector list of each vector and what faults it detects (well a file for each vector). Then I can use the vector lists I made before to see how many faults each of the vectors detect. :) Yay? Problem is, I don't know what wires to connect. I can only connect 'non feedback edges', and its hard to find them. I asked Stella if her code generates something like that, and she said she'll look at it, and from the sound of it, its hard to parse and it will take a while, and I'll need to do it. So hopefully she can give it to me soon and I can play with it. I'm also terrified of how long these scripts will take to run on all the data. I can't make them run more efficiently since well... there is a lot of data, just gotta deal with that. But it will take forever! I don't want to be stuck because I am waiting for them to run... On that note, I got some papers from Prof Dworak that I can do some 'background reading' in for my final report. I guess that will be my 'downtime' activity. :)


Hours away from progress! (in a good way)

I got the code from Stella on her distance measurements, and she said that all I had to do was switch the BFS search for a DFS search. I did that but then realized that her code produces combinations of wires that are within a certain distance of each other... but I need the opposite: wires that are not connected at all. So I modified her code to find all wires that are in both ways not connected at all together. Now that is running. I tested it on c432, which is a smaller circuit, and it worked (!!), and now it is running on colorconverter, which is huge... I'd actually set it running overnight, but found that the file it was creating was too huge to open (this seems to be a common problem...). So today I modified it to only look for a certain amount of bridges. I figure... I don't need ALL the bridges in the circuit, just enough to see how it looks. So thats running now and I hope it will finish by the time I have to leave today. :)