Some Performance Results of Interest from Salmon and Warren
16 Pentium Pro's in a cluster (cost $30K today) reach about 1 gigaflop on a tree code with 107 particles
6800 Pentium Pro's (ASCI Sandia machine) with around 300 million particles
- 600 gigaflops with simple O(N2) code
- 400 gigaflops with tree code
But tree code is 105 times as efficient as O(N2) code on this problem
- i.e if Execution Time(naïve code) is c . N2
- then Time (tree code) is 3000 c N
- and tree code is more efficient than O(N2) code for more than 3000 particles