Comment 4 for bug 1173972

Agreed with denixx: an instantaneous (and space-preserving) bar graph per core, plus a combined graph of CPU load history (i.e. what we have now) would be the ideal approach to display all relevant information in a compact way.