Casting Pearls Ballistically: Efficient Massively Parallel Simulation of Particle Deposition
✍ Scribed by Boris D. Lubachevsky; Vladimir Privman; Subhas C. Roy
- Book ID
- 102584967
- Publisher
- Elsevier Science
- Year
- 1996
- Tongue
- English
- Weight
- 425 KB
- Volume
- 126
- Category
- Article
- ISSN
- 0021-9991
No coin nor oath required. For personal study only.
✦ Synopsis
particles. Here for expository simplicity we eliminated the Y-coordinate thereby reducing this process to two dimen-We simulate ballistic particle deposition wherein a large number of spherical particles are ''cast'' vertically over a planar horizontal sions, X and Z, and then we generated the deposition over surface. Upon first contact (with the surface or with a previously a segment of length 10 particle-diameter units. Unlike this deposited particle) each particle stops. This model helps material example, interesting simulations are in 3D. In order to scientists to study the adsorption and sediment formation. The yield statistical confidence, the actual runs involve many model is sequential, with particles deposited one by one. We have found an equivalent formulation using a continuous time random millions of particles and a substrate area in millions of process and we simulate the latter in parallel using a method similar square units. The 19 balls shown in Fig. 1 (right) is a tiny to the one previously employed for simulating Ising spins. We augfragment cut out of such a large configuration.
ment the parallel algorithm for simulating Ising spins with several
The main interest in these simulations arises in the field techniques aimed at the increase of efficiency of producing the of deposition of submicron particles on substrates. Referparticle configuration and statistics collection. Some of these techniques are similar to earlier ones. We implement the resulting algoences to the underlying physical systems in colloid chemisrithm on a 16K PE MasPar MP-1 and a 4K PE MasPar MP-2. The try, in biology, and in other fields can be found in [10, 13, parallel code runs on MasPar computers nearly two orders of magni-17]. Numerical studies of these and similar systems in 3D tude faster than an optimized sequential code runs on a fast and 2D are surveyed in Section 6. All previously developed workstation.