[Shootout-list] plugin test

[Bengt Kleberg]

Isaac Gouy wrote:
> Maybe the plugin test is devious - but what does it have to do with
> performance measurements?

i think it is trying to measure the performance when changing plugins.


> The Erlang CPU timings on this test for N=3 and N=17 differ by 0.001s
> (which is beyond what we can reliably measure). 

i do not know the reason behind the values for n in this test. i think a 
wide range of n would be a good thing.

this is a test where a large n should not make the time it takes to do 
each step grow unless the language keeps the old plugin around forever. 
such a failure will hopefully be visible in the graph.


moreover, the webpage is showing the code for the third plugin module, 
not the program.


bengt