2003-11-27

fourier

this is nifty! i dumped the 15-minute load average history (from rrd) of eco for the last 3 weeks to a file, loaded that file into matlab, and did a fft / power spectrum analysis on it. much to my surprise it actually worked! HUGE spike at (186400) Hz (i.e., 1 day) … over 5x larger than anything else