[Shootout-list] benchmark design, hard limit for how many 'n'?

[Brent Fulgham]

--- Bengt Kleberg <bengt.kleberg@ericsson.com> wrote:
> > ndata.csv holds the data for individual tests 
> > (i.e., each value of N for which a test was run),
> > for the 'side-by-side' comparison.
> 
> are there any limitations for this data set? (ie, do
> all languages have to have the same amount of 'n'
> for them to be displayed?).

You have a line of data in 'ndata.csv' for every test
run, up until the first failure.  So, if
'ackermann.mzscheme' runs all tests (and has a row for
N=7, N=8, N=9, and N=10), but ruby fails on N=9, ruby
will have a row for N=7, N=8, and N=9 (showing the
failure), but no N=10 value.

-Brent