This benchmark is from "malloc() Performance in a Multithreaded Linux Environment":

http://www.citi.umich.edu/techreports/reports/citi-tr-00-5.pdf

Try the following parameters, where P = 1 and then the number of
processors on your system:

./linux-scalability 512 10000000 P
./linux-scalability-hoard 512 10000000 P