Power and cooling are some of the major challenges for high-performance 
computing, impacting datacenter managers on many levels. In this talk, 
we discuss these challenges and solutions, both hardware and software, 
that can address these challenges. We share the lessons we learned from 
developing the ultra energy-efficient SuperBlade server systems and 
successful deployment at CERN for the well-known Large Hadron Collider 
(LHC). An example of ultra energy-efficient personal supercomputer will 
be given as well.