It’s been known for a while that NVIDIA was interested in getting into the super computer business, but I am not sure if anyone knew it was going to happen so soon. GPUs are capable of blowing away desktop processors for certain operations, so now NVIDIA is putting that into action with their personal supercomputer solution called Tesla.
Similar to an 8-series card, the Tesla GPU features 128 parallel processors and delivers up to 518 gigaflops of parallel computation. Throw four in a system and it will offer you a staggering 2 teraflops of computation to deal with. It doesn’t stop there though, as you can daisy chain machines to tackle computations together, essentially offering up to 8 teraflops. These GPUs are not for the home user, but the need is obvious and many examples are laid out on the site. The GPUs themselves are pricey, reportedly starting off at $1,500 each.
Revolutionary NVIDIA Tesla high performance computing (HPC) solutions arm scientists, engineers and other technical professionals with the power to solve previously unsolvable problems. A dedicated, high performance GPU computing solution, Tesla brings supercomputing power to any workstation or server and to standard, CPU-based server clusters.
Source: NVIDIA Tesla Page