By the way, this problem with fslpspec is under OSX 10.5.2. Looking at the process, after 37 minutes
of CPU time, I have had over 500 million memory faults (that can't be good). This is surprising, given
that this machine has 22GB of RAM (of which 15GB is currently free). The data set that I'm analyzing
is 64x64x45x220 time points.
Doing a process sample shows that the it is in fact, stepping through the 3D data, doing "RealFFT"
over and over again - it's just doing it very slowly...
Blaise
|