actually, this test prog was written after there were problems
with multithreaded domestic application server, that was tested under load.
this one is using intensively hash manipulations.
it is develped for redhat 7.3 and tested there.
it was interesting for me, how it will behaves on fbsd in
linux emulation - it was fine (note: there is linix pthreads).
then we wrote this test prog to investigate the threads with
heap manipulations sumultaneousely. It was interesting for
me to investigate the same in freebsd. that's how it was. :)
Hmm, libc_r and libpthread handle spinlock differently which malloc
uses to protect itself, some real world benchmarks are better than this.