Abstract :
Accelerator physicists today have access to computers that are far more powerful than those available just 10 years ago. In the early 1980´s, desktop workstations performed less than one million floating point operations per second (Mflops), and the realized performance of vector supercomputers was at best a few hundred Mflops. Today vector processing is available on the desktop, providing researchers with performance approaching 100 Mflops at a price that is measured in thousands of dollars. Furthermore, advances in Massively Parallel Processors (MPP) have made performance of over 10 gigaflops a reality, and around mid-decade MPPs are expected to be capable of teraflops performance. Along with advances in MPP hardware, researchers have also made significant progress in developing algorithms and software for MPPs. These changes have had, and will continue to have, a significant impact on the work of computational accelerator physicists. Now, instead of running particle simulations with just a few thousand particles, we can perform desktop simulations with tens of thousands of simulation particles, and calculations with well over 1 million particles are being performed on MPPs. In the area of computational electromagnetics, simulations that used to be performed only on vector supercomputers now run in several hours on desktop workstations, and researchers are hoping to perform simulations with over one billion mesh points on future MPPs. In this paper we will discuss the latest advances, and what can be expected in the near future, in hardware, software and applications codes for advanced simulation of particle accelerators
Keywords :
digital simulation; parallel processing; particle accelerators; particle beam diagnostics; physics computing; vector processor systems; 10 GFLOPS; 100 MFLOPS; advanced simulation; computational accelerator physicists; computational electromagnetics; desktop simulations; massively parallel processors; particle accelerators; particle simulations; vector processing; vector supercomputers; Application software; Computational electromagnetics; Computational modeling; Computer simulation; Hardware; Particle accelerators; Physics computing; Software algorithms; Supercomputers; Workstations;