As an entry for the 1997 Gordon Bell performance prize, we present results from two methods of solving the gravitational N-body problem on the Intel Teraflops system at Sandia National Laboratory (ASCI Red). The first method, an algorithm, obtained 635 Gigaflops for a 1 million particle problem on 6800 Pentium Pro processors . The second solution method, a treecode which scales as , sustained 170 Gigaflops over a continuous 9.4 hour period on 4096 processors, integrating the motion of 322 million mutually interacting particles in a cosmology simulation, while saving over 100 Gigabytes of raw data. Additionally, the treecode sustained 430 Gigaflops on 6800 processors for the first 5 timesteps of that simulation. This treecode solution is approximately times more efficient than the algorithm for this problem. As an entry for the 1997 Gordon Bell price/performance prize, we present two calculations from the disciplines of astrophysics and fluid dynamics. The simulations were performed on two 16 Pentium Proprocessor Beowulf-class computers (Loki and Hyglac) constructed entirely from commodity personal com- puter technology, at a cost of roughly $50k each in September, 1996. The price of an equivalent system in August 1997 is less than $30k. At Los Alamos, Loki performed a gravitational treecode N-body simulation of galaxy formation using 9.75 million particles, which sustained an average of 879 Mflops over a ten day period, and produced roughly 10 Gbytes of raw data. During the initial 10 hours of the simulation, Loki sustained 1.19 Gigaflops. This simulation is nearly identical to that which won a Gordon Bell performance prize in 1992 on the 512 processor Intelelta. At Caltech, Hyglac performed a simulation of the fusion of two vortex rings using a vortex particle method which took place over a 20 hour period, sustaining about 950 Mflops over that time span. Loki and Hyglac were connected together on the floor of Supercomputing’96 in November 1996, obtaining 2.19 Gigaflops on an N-body treecode benchmark.
Warren, M. S., Salmon, J. K., Becker, D. J., Goda, M. P., Sterling, t., & Winckelmans, G. (1997). Pentium Pro inside: I. A treecode at 430 Gigaflops on ASCI Red; II. Price/performance of $50/Mflop on Loki and Hyglac. Supercomputing 97, San Jose, California. https://hdl.handle.net/2078.5/184140