Background

As design complexity has increased and computational power has become more available, leading manufacturers in many industries have adopted simulation as part of the design process. The technology enables companies in the electronics industry to replace costly hardware prototyping and testing with virtual design. Using simulation, virtual prototypes can be analyzed for performance and reliability, and optimized early in the design process.

ANSYS, Inc. is a global innovator of simulation software and technologies designed to optimize the product development process. The ANSYS software solutions address the needs of organizations in a wide range of industries including electronics, aerospace, automotive, chemical processing, civil engineering, defense, medical, and others.

By combining state-of-the-art simulation technology with high performance computing, unprecedented levels of detail and realism can be accurately accounted for in the early stages of a design. The latest Appro servers and clusters solutions based on the AMD Opteron processor deliver outstanding performance and scalability for ANSYS simulation software and related EDA applications. Together, Appro, AMD and ANSYS create end-to-end industry-leading solutions. This partnership ensures a solution that responds to your specific needs, regardless of whether you are a global enterprise or small business. From software tuning and benchmarking, through computer system specification and delivery, Appro, AMD, and ANSYS will work together to ensure your satisfaction.

Appro Benchmark Report - 5/25/07

The Fluent benchmark suite consists of nine tests of three problem sizes (small, medium, and large). The problem size is determined by the number of cells that are solved for specific problem. In CFD a cell is an area of volume used to define the flow around an object. In general, the more difficult the problem (or more precise a result) needed the larger the number of cells and hence the larger the amount of computation. The benchmark numbers are defined as the number of benchmarks that can be run on a given machine (in sequence) in a 24 hour period. Thus the bigger the benchmark the longer it takes and the lower its rating. Ratings increase as processors are added to the computation.
Small class problems contain less than 100,000 cells. (FL5S1,FL5S2,FL5S3)
Medium class problems contain between 100,000 and 500,000 cells. (FL5M1,FL5M2,FL5M3)
Large class problems contain more than 500,000 cells. (FL5L1,FL5L2,FL5L3)
For instance test FL5L2 Exterior flow around a passenger sedan. You can find background on the tests below.

Benchmark Results - Recent Results posted 5/25/07

How to read results: ---There are three sets of benchmarks: Small, Medium and Large Class. If you want to compare our benchmark numbers against the competition, you would need to click on the SL5S1, FL5S2 or FL5S3 for small class and do the same to the other classes. Higher rating and speed is best and higher % numbers means better efficiency.

>>http://www.fluent.com/software/fluent/fl5bench/new.htm

>>http://www.fluent.com/software/fluent/fl5bench/flbench_6.3/fullres.htm

Older Benchmark Results - 10/4/06

>> http://www.fluent.com/software/fluent/fl5bench/flbench_6.2/fullres.htm