Damn my lousy programming skills
Thursday, April 2nd, 2009I just found out that the experiments I used to figure out the feasibility of analysing the CoalHMM runs through a database were buggy. I noticed that the query time once we started filling in actual data in our database were very inconsistent with the results I got earlier, so I started reading through the prototype code and comparing it with the actual database code.
It turns out that I had a bug in the prototype that speeded the query time up a lot, but gave the wrong results. The actual code, that was properly tested, doesn’t have the bug, but the query time is much slower.
I guess I have to do the feasibility analysis all over again…
–
92-114=-22