With the additional of the Tesla K20 GPU to the NVIDIA Maximus system, the compute work is channeled away from the CPU as well as the Quadro K5000. The Maximus set of software and drivers is able to transparently and automatically assign visualization, simulation or rendering work to the right GPU. According to NVIDIA, the 2nd generation Maximus technology is capable of reducing overheads to the system's CPU as well. The feature highlights of the Tesla K20 include the following aspects that enthusiasts would have already been familiar from the desktop GeForce 6 series:-
SMX promises to deliver more performance, due to its higher number of CUDA cores, than the Fermi architecture and it will deliver the stipulated level of performance with less power; in fact, it is touted to be three times as efficient as the previous generation SM.
The Tesla K20 GPU supports Dynamic Parallelism and Hyper-Q technologies, making it easier for GPU developers to speed up all parallel recursive loops, resulting in the Kepler GPU being able to create its own jobs without having to go back to the CPU for such instructions. This will allow the compute work to be sped up, while at the same time reducing computing overheads at the system's CPU.