Crucial to the team's success is the Opteron processor. Why? Because so much of the car's design and engineering relies on high-powered computing systems. AMD is being used for:

- Telemetry: During every race and practice session, data from the onboard computers are sent back from the car to the pit garage. This data needs to be crunched and analysed very quickly so that the engineers can extract race-defining information. Pit stop strategies, tyre decisions and fuel loads are all based on computer simulations of the race extrapolated from telemetry data.

- Car design: The precision required to design every millimetre of the car obviously requires a high-powered CAD/CAM programme running on a meaty workstation.

- Wind tunnel: Crucial to the speed of the car is the aerodynamics of the wings and the body shape. To test this, the team uses a wind tunnel and then records how the air flows over the car. By recording the airflow, they can work out how to make the car sleeker. This data recording is called Computational Fluid Dynamics, or CFD for short.

AMD is involved in all these areas. Opteron machines populate the pit garage, and AMD's pair of technicians are always visible hunched over machines on race day. When it comes to design, Opteron-based workstations run the sophisticated software required. Wind tunnel data is analysed in a new state of the art facility built especially for the purpose, populated with a mammoth 400-node cluster of dual-core Opteron machines.

The software running in the data centre is called Fluent, which is the leader in CFD analysis. With the help of AMD, it's been written to take advantage of the extra memory allowed by the 64-bit processors that AMD has. AMD has also been able to significantly up Ferrari's computing performance over the last year. Just as many of you will have upgraded to a dual core AMD processor, merely flashing the BIOS to enable the extra functionality, Ferrari has done exactly the same thing - times 400. Because Fluent is a threaded application from the ground up, dropping in effectively another 400 processors literally doubles performance. That's a pretty cool real world application of the technology.