[Kst] Some Benchmarking Results

[Barth Netterfield]

Hi,

Based on the feeling that kst-1.0.0 is a little slow, I have done some 
benchmarks comparing cvs of today, 0.97, and 0.94.  Absolute values are, of 
course irrelevant, given the dodgy approach, but relative is intersting:

1)  Real time plot of 9 100 Hz fields, 200ms update, dirfile data source:
Total CPU usage (kst, X, everything else)
0.94:  35%
0.97:   30%
CVS: 33%

So in real time operation, we are currently between 0.94 and 0.97.... We have 
also deleted a lot of plotting bugs since 0.97, so , this is probably not 
surprising.

2) Load and plot 3 10E6 sample fields in three separate plots, and exit:
time:
0.94: 13.8 s
0.97: 12.5 s
CVS: 16.25s

We have regressed a lot in loading/plotting big data files since 0.97.  This 
is a big problem!  Don't know if it is in plotting, or in loading.

cbn