From IBM Registered Knowledge Base document 22631216: To simplify the MI performance calculations, CPU cycles are scaled by the Licensed Internal Code (LIC) so that it will appear that a partition is using an integral number of processors. (For example, a partition configured for .5 processor is scaled so that it appears to be using one whole processor.) The excess cycles are charged to a dummy task called HvLpTask. HvLpTask is strictly for bookkeeping purposes and it is not really using any CPU.

Reply With Quote