Thanks for the feedback. I will fix the test accordingly.
>> +# Create a table to control scavenge cycles.
>> +# The new scavenger will scavenge for every (250Mb / 50) = 5Mb of record data.
> Just out of curiosity: why 50? Where does that number originate from?
The previous scavenger used 20 age groups, or 250 / 20. I wanted better
accuracy in how much the scavenger would scavenge, so that we would be
able to push the scavenge threshold up to 90% or higher, and so it would
run more often. So I chose 50, or 2% of cache for the age group size
instead of 5%.