I have a new Core i7 - 4930K hexa-core computer whose performance is very disappointing. I'm running it in a cluster of 3 computers, the others being Core i7 quad cores (4770 and 3770K). All are running Ubuntu 12.04 and OpenMPI. With 12 MPI processes on the 4930K and 8 MPI processes on each of the quad cores, the 4930K consistently runs a factor of 2.5 slower than the quads. That is, each process on the 4930K takes 2.5 times longer to complete a lengthy cpu-bound task. (90 minutes versus 35 minutes). The 4930K is on the ASUS Sabertooth motherboard. I've tried various tuning options provided in the ASUS bios with little effect. Any ideas?