|High-Performance, High-Precision Computing in Math, Physics and Engineering |
Mechanical and Aerospace Engineering
|David H. Bailey, Lawrence Berkeley Laboratory|
3:30 PM, Engineering Room B-120
Large-scale scientific computing, aided by relentless advances in hardware, algorithms, parallelization, performance and visualization, is now well-established as a third form of scientific discovery, after theory and experiment. Most of these calculations are performed using standard 64-bit (16-digit) IEEE arithmetic, but in recent years a number of applications have arisen that require substantially higher numeric precision. One of these, in a study co-authored with the late Richard Pelz, investigated vortex roll-up using 64-digit arithmetic. Other applications of highprecision computation include supernova simulations, planetary dynamics, n-body Coulomb atomic systems and electromagnetic scattering. Of particular interest are the numerous recent discoveries of new results in mathematical physics, facilitated by extremely high precision computations coupled with the "PSLQ" integer relation algorithm. This talk gives a brief overview of the techniques used and a survey of some of the recent results.