This page contains a Flash digital edition of a book.
HPC comment fast lane


When speed and reliability are


of the essence, professional


racing teams need a computing system


Egginton, chief engineer at Team Lotus, explains


they can rely on, as Jody In the W


hen our team was set up in 2009 we asked ourselves what we needed from our strategy software, and decided


to develop that particular software in- house to ensure it could be tailored to our specifi c needs. As a team, we know how fast our car can go and have quite a good idea of everyone else’s speed. Add that to the observations of competitors’ performance during the weekend before the race and we know roughly how their tyres will behave – and we can plot how the cars will act on the track. All of this information can be entered into a database and then, using the software, we repeatedly crunch the numbers until we have a prediction of how the race will pan out. We then start throwing different


scenarios into that mix, such as the probability of there being a safety car, so that our strategy can take a variety of possibilities into account. The variables are then added in and we calculate the possible outcomes – some of which can be discounted immediately based on experience – before refi ning them via the software. As the race starts certain things will happen, so we’re constantly updating the information. For example, the car we predicted would stop for tyres between laps 15 and 20 stops instead at lap fi ve. Its pace has to then be adjusted as it will be quicker with new tyres, but having them fi tted earlier in the race means it will be heavier due to the amount of fuel still on board. Throughout the race the software is


constantly running algorithms to ensure we’re updated on all possible outcomes. Of course, while the software can predict what can happen it is still up to the individual to react to what actually does happen and to narrow those scenarios down. That’s how that software works and it is massively processor-hungry, which is why we will be upgrading to 64-bit machines for these types of applications, and from two Gb to four. The software is under constant refi nement and, as our


Kit List


The computational fl uid dynamics software can simulate test conditions found in wind tunnels and on the track, enabling the team to fi lter down their design ideas


www.scientific-computing.com


The Team Lotus factory uses a scalable Dell high-performance computing (HPC) solution consisting of Dell PowerEdge M1000e blade enclosures, Dell PowerEdge M610 blade servers and Dell PowerEdge R610 and R710 rack servers. The trackside offering includes the Dell Precision M6400 mobile workstation and the Dell Precision T7500 tower workstation, in addition to a mobile datacentre which collects and processes data during each lap of the race. The data is then sent to Dell EqualLogic PS6000X iSCSI storage area network arrays (SANs) for further analysis at the factory.


FEBRUARY/MARCH 2011 21


engineers continue to ask more of it, our hardware needs to keep up. Another in-house software package


we use deals with simulations. We code it with everything it needs to know about our car, from the aerodynamic ability and drag levels to how the tyres behave in different conditions. We enter parameters, such as the level of the wing, and the software predicts how the car will sit on the track. From that, we calculate if we are in the optimum ride height for aerodynamics. It generally takes less than a minute to do a lap in simulation, produce times and see how the car is behaving. Computing power is crucial as we might want to run a batch of 30 simulations to get a stepped answer, and without adequate processing we could easily waste an entire day waiting for results. We are all very impatient and if something takes longer than 60 seconds, people generally start complaining. It is a massively repetitive process


in which we are basically running and re-running the same calculations with one variable slightly changed in order to arrive at the best solution. The software is also used trackside as each car in race confi guration has 150 sensors relaying data to the datacentre. Once processed, that information is used to conduct continuous simulations and both the hardware and software have to be powerful and robust to handle the complex mathematical calculations in real time. Further refi nements are done offl ine. Computers are fundamental to all


of this. It simply wouldn’t be possible for the team to function without computing power as the cars are far too complicated to be run competitively without the ability to do all these calculations within the necessary timeframe. On the pit wall we generally have less than 90 seconds to make a decision and so don’t have time to delve into situations. We need answers quickly and without central computer processing power we would struggle.


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44