Another way would be to run a VM with 4 threads on top of your native client with 4 threads. They will all run at half speed, taking twice as long, but they will think they are running at full speed.