topics
success story 01:
BoxClusterML implemented by Prof. Antony Jameson
Benchmark test performed at Aeronautics and Astronautics
laboratory
at Stanford University
The benchmark test has been undertaken by the research staffs with Prof.
Antony Jameson. The main purpose is to test the system can perform its expectation
with the actual program code that he is using.
The BoxClusterML is designed for a turnkey solution so it should be able to start the system with very short time after open the carton, power on the system, copy the source code into the system and then to compile, build the code. This time, it took about 20mins after opening the carton box to actually finished the compilation to start the benchmark test. The system that they have purchased previously took about 4 hours to start using as they have to re-install the OS, setup the compiler and libraries to use as a cluster MPI environment.
The benchmark test was performed using “FLO” software that Prof. Antony Jameson has been developed. The best result from the laboratory was about 800sec. This time, the result was 350sec and all of them could not believe the result that they checked the time stamp and run several times to triple check the results. The picture shows the Prof Antony Jameson and his staffs checking the result of the benchmark test.
BoxClusterML installation
Based on the benchmark test result, Prof Antony Jameson has decided to purchase one system for him self and later Stanford University has purchased two systems for the lab.
He actually used the newly purchased BoxClusterML to produce 7 aircraft wing re-design within 1 day and this result was published at AIAA, 46th Aerospace Sciences Meeting and Exhibit with the title of “AIAA 2008-0145 Aerodynamic-Structural Design Studies of Low-Sweep Transonic Wings”.
His copy of presentation can be found at:
http://aero-comlab.stanford.edu/Papers/AIAA-2008-145-822.pdf
[Written by Akira Sano]