Summer 2008 DMP Project at Brown

Starting to get the hang of things...

This weekend was my first weekend in Providence. I was afraid of not having anything to do and being bored out of my mind, but things came and up and it turned out well! Went up to Boston for one evening, exploring Providence on Saturday and to the beach on Sunday. Looks like there's things to do in this town! It has also gotten unbearably hot. I am in need of a fan in my non-air conditioned room. For now, spending more time in the lab because it has air conditioning. :)

Work wise, things look like they will be more productive. I spent a good part of the day refreshing my knowledge of Perl. Its' amazing how much one forgets! Then Yiwen arrived and helped me get her files on the fault dictionary, so now I have something to go on for the project. I am looking at a plot of the different faults and how many times they are detected when various types of tests are run on them to try to find a threshold, above which, you don't need to consider the faults, because they will most likely be detected anyways, since they are detected by so many tests. I am not sure how to set said threshold, but working on it.


A system can only do so much

So I figured out that I need a file that Yiwen (one of the grad students) has that is a matrix of all of the vectors run and all of the faults that each of them detects. I need this file in order to see what would happen if faults that are always detected are not worried about. Problem is... this file is huge. So huge she can't put it on my computer. Waiting on the system administrator to make that available to me. I hate how you can get so bogged down by either people not being here, or technological issues...


I am moving along!!!

Professor Dworak returned yesterday and I have things to work on now!! YES!!! I can't describe how excited I am to actually have a project to work on, and sort of know what I am planning on doing! I am working on what I was working on before: finding a threshold for the number of times a fault is detected, above which one does not need to worry about the fault anymore since it was detected so many times, odds are it will be detected anyways. This would allow more focus on detecting the faults that are not detected often. I am writing some code that will decided based the various vectors that the circuit is tested with and the criticality of each fault (how bad it is if it is not found), that will decide which vectors should and should not be used.

The big breakthrough was actually... finding where the vector files are. Yiwen was hesitant about giving them to me because they are so huge. It turns out she has them linked from her home directory, which I can access, and so its not a problem! I can just route a folder to her machine and there is no problem! There was a lot of 'rediscovery' of various UNIX commands that I had forgotten with that. Its a useful thing! So now, I have a project, and I am ready to write some Perl! YES!


End of Week 2

So Week 2 has been significantly more productive than Week 1. I feel like I am settling into the project and can make some actual progress. I have written a script that takes faults that are detected below a certain threshold, and chooses vector files (pattern test sets) that would detect those faults only. Then, with the chosen vector files, it checks which faults are detected. There are some undetected faults, which needs to be worked on. I also want to look at criticalities of faults.

Right before leaving for the weekend I discovered a very frustrating problem... Open office, which I have used to sort my data bafore I put it into text files for parsing, can only handle 32,000 rows... while my data is about 65,000 rows in length. I had not noticed that it had cut off half my data... so my tests have not been producing the actual results. Will have to work on that next week.... :(