Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

We demonstrate that a monolayer graphene membrane is impermeable to standard gases including helium. By applying a pressure difference across the membrane, we measure both the elastic constants and the mass of a single layer of graphene. This pressurized graphene membrane is the world's thinnest balloon and provides a unique separation barrier between 2 distinct regions that is only one atom thick.

/pdf/impermeable-atomic-membranes-from-graphene-sheets-xtis5huy2d.pdf

Impermeable atomic membranes from graphene sheets.

SINCE the invention of the scanning tunnelling microscope1, the value of establishing a physical connection between the macroscopic world and individual nanometre-scale objects has become increasingly evident, both for probing these objects2–4 and for direct manipulation5–7 and fabrication8–10 at the nanometre scale. While good progress has been made in controlling the position of the macroscopic probe of such devices to sub-angstrom accuracy, and in designing sensitive detection schemes, less has been done to improve the probe tip itself4. Ideally the tip should be as precisely defined as the object under investigation, and should maintain its integrity after repeated use not only in high vacuum but also in air and water. The best tips currently used for scanning probe microscopy do sometimes achieve sub-nanometre resolution, but they seldom survive a 'tip crash' with the surface, and it is rarely clear what the atomic configuration of the tip is during imaging. Here we show that carbon nanotubes11,12 might constitute well defined tips for scanning probe microscopy. We have attached individual nanotubes several micrometres in length to the silicon cantilevers of conventional atomic force microscopes. Because of their flexibility, the tips are resistant to damage from tip crashes, while their slenderness permits imaging of sharp recesses in surface topography. We have also been able to exploit the electrical conductivity of nanotubes by using them for scanning tunnelling microscopy.

Nanotubes as nanoprobes in scanning probe microscopy

The Interaction of Water with Solid Surfaces: Fundamental Aspects Revisited

A concept for molecular electronics exploiting carbon nanotubes as both molecular device elements and molecular wires for reading and writing information was developed. Each device element is based on a suspended, crossed nanotube geometry that leads to bistable, electrostatically switchable ON/OFF states. The device elements are naturally addressable in large arrays by the carbon nanotube molecular wires making up the devices. These reversible, bistable device elements could be used to construct nonvolatile random access memory and logic function tables at an integration level approaching 10 12 elements per square centimeter and an element operation frequency in excess of 100 gigahertz. The viability of this concept is demonstrated by detailed calculations and by the experimental realization of a reversible, bistable nanotube-based bit.

/pdf/carbon-nanotube-based-nonvolatile-random-access-memory-for-2m51guev4k.pdf

Carbon nanotube-based nonvolatile random access memory for molecular computing

A general method for optimizing problem layout on the Blue Gene®/L (BG/L) supercomputer is described. The method takes as input the communication matrix of an arbitrary problem as an array with entries C(i, j), which represents the data communicated from domain i to domain j. Given C(i, j), we implement a heuristic map that attempts to sequentially map a domain and its communication neighbors either to the same BG/L node or to near-neighbor nodes on the BG/L torus, while keeping the number of domains mapped to a BG/L node constant. We then generate a Markov chain of maps using Monte Carlo simulation with free energy F =Σi,j C(i, j)H(i, j), where H(i, j) is the smallest number of hops on the BG/L torus between domain i and domain j. For two large parallel applications, SAGE and UMT2000, the method was tested against the default Message Passing Interface rank order layout on up to 2,048 BG/L nodes. It produced maps that improved communication efficiency by up to 45%.

/pdf/optimizing-task-layout-on-the-blue-gene-l-supercomputer-12ir5f4d1a.pdf

Optimizing task layout on the Blue Gene/L supercomputer

A Michelson interferometer has been used as a direct quantitative probe for gas phase plasma formation in the UV excimer laser ablation of solids. Excimer laser fluence thresholds for plasma formation are determined and correlated with optical emission from electronically excited ablation fragments.

Studies of excimer laser ablation of solids using a Michelson interferometer

This paper explores the performance and optimization of the IBM Blue Gene/Q (BG/Q) five dimensional torus network on up to 16K nodes. The BG/Q hardware supports multiple dynamic routing algorithms and different traffic patterns may require different algorithms to achieve best performance. Between 85% to 95% of peak network performance is achieved for all-to-all traffic, while over 85% of peak is obtained for challenging bisection pairings. A new software-controlled algorithm is developed for bisection traffic that selects which hardware algorithm to employ and achieves better performance than any individual hardware algorithm. The benefit of dynamic routing is shown for a highly non-uniform "transpose" traffic pattern. To evaluate memory and network performance, the HPCC Random Access benchmark was tuned for BG/Q and achieved 858 Giga Updates per Second (GUPS) on 16K nodes. To further accelerate message processing, the message libraries on BG/Q enable the offloading of messaging overhead onto dedicated communication threads. Several applications, including Algebraic Multigrid (AMG), exhibit from 3 to 20% gain using communication threads.

Looking under the hood of the IBM blue gene/Q network

A general computer-implement method and apparatus to optimize problem layout on a massively parallel supercomputer is described. The method takes as input the communication matrix of an arbitrary problem in the form of an array whose entries C(i, j) are the amount to data communicated from domain i to domain j. Given C(i, j), first implement a heuristic map is implemented which attempts sequentially to map a domain and its communications neighbors either to the same supercomputer node or to near-neighbor nodes on the supercomputer torus while keeping the number of domains mapped to a supercomputer node constant (as much as possible). Next a Markov Chain of maps is generated from the initial map using Monte Carlo simulation with Free Energy (cost function) F=Σ i,j C(i,j)H(i,j)− where H(i,j) is the smallest number of hops on the supercomputer torus between domain i and domain j. On the cases tested, found was that the method produces good mappings and has the potential to be used as a general layout optimization tool for parallel codes. At the moment, the serial code implemented to test the method is un-optimized so that computation time to find the optimum map can be several hours on a typical PC. For production implementation, good parallel code for our algorithm would be required which could itself be implemented on supercomputer.

Optimizing layout of an application on a massively parallel supercomputer

Calcul des trajectoires de la desorption d'ions positifs a partir d'halogenures alcalins par le mecanisme de l'explosion coulombienne, sur l'exemqple de NaF. Mise en evidence du role determinant des distorsions dynamiques du reseau et de la possibilite de piegeage de l'ion F + meme dans le cas d'une localisation favorable de l'excitation electronique

Robert E. Walkup

Papers

Optimizing task layout on the Blue Gene/L supercomputer

Studies of excimer laser ablation of solids using a Michelson interferometer

Looking under the hood of the IBM blue gene/Q network

Optimizing layout of an application on a massively parallel supercomputer

Classical-trajectory studies of electron- or photon-stimulated desorption from ionic solids.