| CS378, History Of Computing Week 8, Weekly Report Due 2004 Oct 14 Reports for History of Computing CS378
Introduction During Dr. Clines talk he brought up many good points about and uses for numerical analysis computing. His lecture showed how floating points make numerical computing difficult. Taking this into considering I would like to look at how two different technologies have helped progress numerical analysis. The first topic I will look at is Mathematica. Dr. Cline briefly discussed this topic, speaking of its ability to solve complex numerical problems, such as the programming of oil refinery expectations. The second topic I will discuss is parallel computing, Dr. Cline touched on the subject when we began discussing how numerical computation was used for weather prediction and that parallel computing was necessary for such computations. Mathematica Mathematica 1.0 was introduced in 1988 during a time when there were many standard languages such as FORTRAN and PASCAL, but no easy way to integrate them into technical work flows. Mathematica is an application that brought the computing community a turnkey solution for integrating complex computing systems, which resulted in better data integration and representation of results. For example the ability to calculate complex computational data such as weather patterns, process the findings and produce simple 3D model representations did not have such an intergrated workflow until Mathematica was created.. Wolfram Research, the maker of Mathematica, says: “the visionary concept of Mathematica was to create once and for all a single system that could handle all the various aspects of technical computing in a coherent and unified way.” (1) Since 1988 this vision has become a reality for many of users of Mathematica, current uses of Mathematica for numerical analysis are: financial forecasts, astronomical models, weather forecasts and many other things. Parallel computing Parallel computing has been used for decades to compute complex algorithms. It’s origins date back to the late 1950’s when Cray built one of the first multi processor super computers.(2) Before parallel computing became to be what it is today; super computers, such as Cray super computers were used to solve complex algorithms. These devices used fast processing power in order to crunch numbers quickly, however they were very expensive and this greatly effected the scalability. There are basically two levels of parallel computing, super parallel and normal parallel. Examples of super parrallel computing systems are the NEC Earth simulator and IBM’s recent super computer. These parallel systems consist of highly customized multi processor/computer systems. A normal parallel system usually consists of off the shelf PC’s that are then connected together through a high speed backbone, which enables them to be combined for greater processing power. These systems have many numerical computing uses, such as the calculation of vectors for rendering 3D models and forecasting financial and weather models. Conclusion With software solutions such as Mathematica and hardware solutions like parallel computing, numerical analysis has progressed forward, not only answering old questions but creating new ones. These questions require more processing power and complex software solutions. I feel numerical computing has helped progress all of technology forward by pushing the state of the art. Currently one technology that might help progress numerical computing is dual core processor technology. This is where a CPU uses dual processors on a single die, creating a high bandwidth communication system that allows for high speed parallel processing. Dual core processors are due in the mid year of 2005, with proper software these systems will help create new findings and questions about numerical computing. Notes 1. "The History of Mathematica." Wolfram Research. http://www.wolfram.com/company/history/ (wolfman.com, Online, accessed 2004 Oct 13) 2. "Super Computers" Calle, Dan. http://ei.cs.vt.edu/~history/SUPERCOM.Calle.HTML (Calle, Dan, Online, accessed 2004 Oct 13) |