TreadMarks supports parallel computing on networks of workstations by providing the application with a shared memory abstraction. Shared memory facilitates the transition from sequential to parallel programs. After identifying possible sources of parallelism in the code, most of the data structures can be retained without change, and only synchronization needs to be added to achieve a correct shared memory parallel program. Additional transformations may be necessary to optimize performance, but this can be done in an incremental fashion. We discuss the techniques used in TreadMarks to provide e cient shared memory, and our experience with two large applications, mixed integer programming and genetic linkage analysis.

Abstract-We present an approach to support massively multi-player games on peer-to-peer overlays. Our approach exploits the fact that players in MMGs display locality of interest, and therefore can form self-organizing groups based on their locations in the virtual world. To this end, we have designed scalable mechanisms to distribute the game state to the participating players and to maintain consistency in the face of node failures. The resulting system dynamically scales with the number of online players. It is more flexible and has a lower deployment cost than centralized games servers. We have implemented a simple game we call SimMud, and experimented with up to 4000 players to demonstrate the applicability of this approach.

China is undergoing a rapid transition from a rural to an urban society. This societal change is a consequence of a national drive toward economic prosperity. Rapid urbanization impacts on infrastructure, environmental health and human wellbeing. Unlike many cases of urban expansion, Chinese urbanization has led to containment, rather than to increase, in the spread of infectious diseases. Conversely, the incidence of chronic conditions such as cardiovascular and metabolic diseases has risen, with higher rates occurring in urban regions. This ruralÀurban gradient in disease incidence seems not to be a reflection simply of more aggressive diagnosis or healthcare access. Other diseases exhibit little rural versus urban differences (e.g., liver cancer or respiratory disease), or even occur at a higher rate in the rural population (e.g., esophageal cancer). This article examines the impact of this changing demographic on environmental health and human wellbeing in China. Lessons learned from epidemiological studies mostly carried out in Europe and the U.S. may not be directly transferable to China. We advocate that there is now a need to establish robust systems of accurate data collection, a Chinese biobank network to facilitate the profiling of human health effects, and relevant randomized controlled trials to identify effective interventions in the Chinese urbanized setting. Such studies could allow for the future implementation of disease-preventive strategies.

Most massively multiplayer game servers employ static partitioning of their game world into distinct mini-worlds that are hosted on separate servers. This limits cross-server interactions between players, and exposes the division of the world to players. We have designed and implemented an architecture in which the partitioning of game regions across servers is transparent to players and interactions are not limited to objects in a single region or server. This allows a finer grain partitioning, which combined with a dynamic load management algorithm enables us to better handle transient crowding by adaptively dispersing or aggregating regions from servers in response to quality of service violations.Our load balancing algorithm is aware of the spatial locality in the virtual game world. Based on localized information, the algorithm balances the load and reduces the cross server communication, while avoiding frequent reassignment of regions. Our results show that locality aware load balancing reduces the average user response time by up to a factor of 6 compared to a global algorithm that does not consider spatial locality and by up to a factor of 8 compared to static partitioning.

We derive a generalized Gross-Pitaevskii density-functional equation appropriate to study the Bose-Einstein condensate ͑BEC͒ of dimers formed of singlet spin-half Fermi pairs in the BEC-unitarity crossover while the dimer-dimer scattering length a changes from 0 to ϱ. Using an effective one-dimensional form of this equation, we study the phenomenon of dynamical self-trapping of a cigar-shaped Fermi superfluid in the entire BECunitarity crossover in a double-well potential. A simple two-mode model is constructed to provide analytical insights. We also discuss the consequence of our study on the self-trapping of an atomic BEC in a double-well potential.

We study the expansion dynamics of a one dimensional polarized Fermi gas after sudden release from confinement using both the mean-field Bogoliubov-de Gennes and the numerically exact Time-Evolving Block Decimation methods. Our results show that experimentally observable spin density modulations directly related to the presence of a Fulde-Ferrel-Larkin-Ovchinnikov (FFLO) state develop during the expansion of the cloud, providing incontrovertible evidence of this long-sought state.

In this paper, we present the first system that implements OpenMP on a network of shared-memory multiprocessors. This system enables the programmer to rely on a single, standard, shared-memory API for parallelization within a multiprocessor and between multiprocessors. It is implemented via a translator that converts OpenMP directives to appropriate calls to a modified version of the TreadMarks software distributed sharedmemory (SDSM) system. In contrast to previous SDSM systems for SMPs, the modified TreadMarks system uses POSIX threads for parallelism within an SMP node. This approach greatly simplifies the changes required to the SDSM in order to exploit the intranode hardware shared memory.We present performance results for seven applications (Barnes-Hut, CLU, and Water from SPLASH-2, 3D-FFT from NAS, Red-Black SOR, TSP, and MGS) running on an SP2 with four four-processor SMP nodes. A comparison between the thread implementation and the original implementation of TreadMarks shows that using the hardware shared memory within an SMP node significantly reduces the amount of data and the number of messages transmitted between nodes and consequently achieves speedups that are up to 30 0 better than the original versions. We also compare SDSM against message passing. Overall, the speedups of multithreaded TreadMarks programs are within 7 30 0 of the MPI versions. Academic Press

Recent groundbreaking experiments studying the effects of spin polarization on pairing in unitary Fermi gases encountered mutual qualitative and quantitative discrepancies which seem to be a function of the confining geometry. Using novel numerical algorithms we study the solution space for a 3-dimensional fully self-consistent formulation of realistic systems with up to 10 5 atoms. A study of the three types of solutions obtained demonstrates a tendency towards metastability as the confining geometry is elongated. One of these solutions, which is consistent with Rice experiments at high trap aspect ratio, supports a state strikingly similar to the long sought Fulde-Ferrel-Larkin-Ovchinnikov state. Our study helps to resolve the long-standing controversy concerning the discrepancies between the findings from two different experimental groups and highlights the versatility of actual-size numerical calculations for investigating inhomogeneous fermionic superfluids. [3]. Primarily due to lack of sufficient experimental evidence, the issue remains largely unresolved even though it is central to many forms of matter such as superconductors, neutron stars and color superfluids in the quark-gluon plasma [3]. Ultra-cold samples of two component degenerate Fermi gases [6][7][8][9][10] have reenergized the debate because of their exquisite controllability.In this paper, we focus on apparently contradictory results on spin-imbalanced unitary Fermi gases from recent experiments between two leading groups [7][8][9][10]. In both experiments it was observed that, consistent with earlier predictions [1,2], the trapped superfliud responds to polarization by phase separating into an inner core with negligible polarization surrounded by a polarized outer shell. However, the Rice experiments [8,9] performed in cigar-shaped traps with total particle numbers N ∼ 10 5 observed a significant and unexpected deformation of the central superfluid core, indicating a clear violation of the local density approximation (LDA). In addition, these results also suggest a much higher superfluid to normal (Chandrasekhar-Clogston) transition than the MIT experiments [7,10] in which no deformations were observed. The excellent quantitative agreement with theory [11,12] for the MIT experiments conducted at much lower trap aspect ratio and with higher particle numbers N ∼ 10 6 , hints that there might be unexpected physics at work in the Rice experiment. In addition, the concurrence of experiments performed in Paris [13] with the MIT experiments also suggest a crucial role of the trapping geometry. This impasse has inspired speculation about the possible role of exotic phases such as the FFLO state in the observed discrepancies, and stirred much discussion and debate over the past few years by the cold atom community.The apparent contradiction between the Rice and MIT experiments reflects theoretical difficulties within trapped geometries: since the effective chemical potential (µ) varies in space, several phases may co-exist within a trapped sample. ...