contification paper: benchmarks

[Stephen Weeks]

> Opinions on how to present the running times.  Currently, I have the
> running time in seconds and the running time normalized to the running
> time of the dominator version.  But, I'm considering doing the
> normalization to the none version.  If dominator was always the clear
> winner, then I would keep that, but when the ratios are all over, it seems
> a little arbitrary to choose dom.

I agree.  normalize to none.