High Performance Computing (HPC) has always been the dream of many scientists around the world, for it can provide answers to problems that reach (or even strech beyond) Universal dimensions.

For example, even Blue Gene L, IBM’s brainchild and probably the most advanced computer cluster in the world, is still overwhelmed when it comes to crunching the trillions of teraflops which describe how much energy is involved when two massive galaxies collide or how proteins fold and misfold in living organisms.

Paradoxically, one solution for the need of HPC came from the gaming industry, after Sony launched its next-gen PlayStation 3 console, powered by IBM’s Cell CPU. In collaboration with Stanford University, Sony started the Folding@Home program for PS3, which dramatically increased the processing power.

Now, Nvidia launches its own HPC-dedicated GPU, called Tesla, which will be integrated from simple workstations to server clusters.

Key features of the new GPU- as mentioned by Nvidia- are: massively multi-threaded architecture with a 128-processor computing core, the ability to seamlessly integrate with existing HPC environments and over 500 gigaflops of peak floating point performance added to a simple workstation,.

Nvidia will also provide scientists and enthusiasts with its proprietary CUDA software development kit, which includes a standard C compiler, hardware debugger tools, and a performance profiler for simplified application development.

To be able to use the GPU’s raw power, customers need an Nvidia Tesla certified system and a Windows XP (32-bit) or Linux (64 and 32-bit, including Red Hat Enterprise and SUSE) operating systems.