Generating Efficient Data Movement Code for Heterogeneous Architectures with Distributed-Memory

Programming for parallel architectures that do not have a shared address space is extremely difficult due to the need for explicit communication between memories of different compute devices. A heterogeneous system with CPUs and multiple GPUs, or a distributed-memory cluster are examples of such systems. Past works that try to automate data movement for distributed-memory architectures can lead to excessive redundant communication. In this paper, we propose an automatic data movement scheme that minimizes the volume of communication between compute devices in heterogeneous and distributed-memory systems. We show that by partitioning data dependences in a particular non-trivial way, one can generate data movement code that results in the minimum volume for a vast majority of cases. The techniques are applicable to any sequence of affine loop nests and works on top of any choice of loop transformations, parallelization, and computation placement. The data movement code generated minimizes the volume of communication for a particular configuration of these. We use a combination of powerful static analyses relying on the polyhedral compiler framework and lightweight runtime routines they generate, to build a source-to-source transformation tool that automatically generates communication code. We demonstrate that the tool is scalable and leads to substantial gains in efficiency. On a heterogeneous system, the communication volume is reduced by a factor of 11X to 83X over state-of-the-art, translating into a mean execution time speedup of 1.53X. On a distributed-memory cluster, our scheme reduces the communication volume by a factor of 1.4X to 63.5X over state-of-the-art, resulting in a mean speedup of 1.55X. In addition, our scheme yields a mean speedup of 2.19X over hand-optimized UPC codes.

Improving the Rigor and Consistency of the Thermodynamic Theory for Clathrate Hydrates through Incorporation of Movement of Water Molecules of Hydrate Lattice

Current applications of statistical thermodynamic theories for clathrate hydrates do not incorporate the translational and rotational movement of water molecules of the hydrate lattice,in a rigorous manner. Previous studies have shown that the movement of water molecules has a significant effect on the properties of clathrate hydrates. In this Article, a method is presented to incorporate the effect of water movement with as much rigor as possible. This method is then used to calculate the Langmuir constant of the guest species in a clathrate hydrate. Unlike previous studies on modeling of clathrate hydrate thermodynamics, the method presented in this paper does not regress either the intermolecular potentials or the properties of the empty hydrate from clathrate phase equilibria data. Also the properties of empty hydrate used in the theory do not depend on the nature and composition of the guest molecules. The predicted phase equilibria from the resulting theory are shown to be highly accurate and thermodynamically consistent by comparing them with the phase equilibria computed directly from molecular simulations.

GBNV encoded movement protein (NSm) remodels ER network via C-terminal coiled coil domain

Plant viruses exploit the host machinery for targeting the viral genome-movement protein complex to plasmodesmata (PD). The mechanism by which the non-structural protein m (NSm) of Groundnut bud necrosis virus (GBNV) is targeted to PD was investigated using Agrobacterium mediated transient expression of NSm and its fusion proteins in Nicotiana benthamiana. GFP:NSm formed punctuate structures that colocalized with mCherry:plasmodesmata localized protein la (PDLP la) confirming that GBNV NSm localizes to PD. Unlike in other movement proteins, the C-terminal coiled coil domain of GBNV NSm was shown to be involved in the localization of NSm to PD, as deletion of this domain resulted in the cytoplasmic localization of NSm. Treatment with Brefeldin A demonstrated the role of ER in targeting GFP NSm to PD. Furthermore, mCherry:NSm co-localized with ER-GFP (endoplasmic reticulum targeting peptide (HDEL peptide fused with GFP). Co-expression of NSm with ER-GFP showed that the ER-network was transformed into vesicles indicating that NSm interacts with ER and remodels it. Mutations in the conserved hydrophobic region of NSm (residues 130-138) did not abolish the formation of vesicles. Additionally, the conserved prolines at positions 140 and 142 were found to be essential for targeting the vesicles to the cell membrane. Further, systematic deletion of amino acid residues from N- and C-terminus demonstrated that N-terminal 203 amino acids are dispensable for the vesicle formation. On the other hand, the C-terminal coiled coil domain when expressed alone could also form vesicles. These results suggest that GBNV NSm remodels the ER network by forming vesicles via its interaction through the C-terminal coiled coil domain. Interestingly, NSm interacts with NP in vitro and coexpression of these two proteins in planta resulted in the relocalization of NP to PD and this relocalization was abolished when the N-terminal unfolded region of NSm was deleted. Thus, the NSm interacts with NP via its N-terminal unfolded region and the NSm-NP complex could in turn interact with the ER membrane via the C-terminal coiled coil domain of NSm to form vesicles that are targeted to PD and there by assist the cell to cell movement of the viral genome complex. (C) 2015 Elsevier Inc. All rights reserved.

Experimental Study on Flow and Sediment Movement in Cavity Circulations

This paper reports a flume experiment of flow and sediment movement in a cavity. The flow velocity, sediment concentration and the mechanism of hydraulic sorting in the circulation flow are discussed. The quantity and patterns of sediment deposition in the circulation area are studied as well.

Gust-Induced Movement of Dusts Suspended in Atmosphere

在气固两相流的单向耦合模型框架下,研究大气悬浮沙尘在阵风作用下的运动轨迹及其随大气风速、沙尘粒径的变化,揭示了在Stokes阻力、Saffman力和重力共同作用下,沙尘悬移可能转换为跃移的现象。

Experimental Investigation Of Particle Velocity Distributions In Windblown Sand Movement

With the PDPA (Phase Doppler Particle Analyzer) measurement technology, the probability distributions of particle impact and lift-off velocities on bed surface and the particle velocity distributions at different heights are detected in a wind tunnel. The results show that the probability distribution of impact and lift-off velocities of sand grains can be expressed by a log-normal function, and that of impact and lift-off angles complies with an exponential function. The mean impact angle is between 28 degrees and 39 degrees, and the mean lift-off angle ranges from 30 degrees to 44 degrees. The mean lift-off velocity is 0.81-0.9 times the mean impact velocity. The proportion of backward-impacting particles is 0.05-0.11, and that of backward-entrained particles ranges from 0.04 to 0.13. The probability distribution of particle horizontal velocity at 4 mm height is positive skew, the horizontal velocity of particles at 20 mm height varies widely, and the variation of the particle horizontal velocity at 80 mm height is less than that at 20 mm height. The probability distribution of particle vertical velocity at different heights can be described as a normal function.

A model experiment of the plate movement

Detailed analysis of some difficult aspects has been made from modeling the platemovement. A new method of using differential density of material (instead of differentialtemperature) has been developed in the experiments. The effect of convection of mantleon the plate movement has been studied using a centrifugal technique, and a patternshowing the recurrence of the plate movement has been successfully obtained. In this paper, a criterion De=Dm is presented for the similarity of the model to thecounterpart of the original mantle. According to the criterion, what happens in the modelin a span of ten minutes suggests a process of the "original model" going on in geologi-cal time of three million years.

Changes in Behavior, Movement, and Home Ranges of Largemouth Bass Following Large-scale Hydrilla Removal in Lake Seminole, Georgia

About 1,200 ha of hydrilla ( Hydrilla verticillata L.f. Royle) was eliminated in the Spring Creek embayment of Lake Seminole, Georgia, using a drip-delivery application of fluridone (1- methyl-3-phenyl-5-[3-(trifluoromethl) phenyl]-4(1H)-pyridinone) in 2000 and 2001. Two groups of 15 and 20 largemouth bass (Micropterus salmoides Lacepede) were implanted with 400-day radio tags in February 2000 and 2001 to determine changes in movement and behavior before and after hydrilla reduction.(PDF contains 8 pages.)

A pilot study of hogfish (Lachnolaimus maximus Walbaum 1792) movement in the Conch Reef Research Only Area (northern Florida Keys)

The largely sedentary behavior of many fishes on coral reefs is well established. Information on the movement behavior of individual fish, over fine temporal and spatial scales, however, continues to be limited. It is precisely this type of information that is critical for evaluating the success of marine reserves designed for the conservation and/or management of vagile fishes. In this pilot study we surgically-tagged eight hogfish (Lachnolaimus maximus Walbaum 1792) with coded-acoustic transmitters inside the Conch Reef Research Only Area (a no-take marine reserve) in the northern Florida Keys National Marine Sanctuary. Our primary objective was to characterize the movement of L. maximus across Conch Reef in the vicinity of the reserve. All fish were captured, surgically-tagged and released in situ during a saturation mission to the Aquarius Undersea Laboratory, which is located in the center of the reserve. Movement of tagged L. maximus was recorded for up to 95 days by three acoustic receivers deployed on the seafloor. Results showed clear diel patterns in L. maximus activity and regular movement among the receivers was recorded for seven of the eight tagged fish. Fidelity of tagged fish to the area of release was high when calculated at the scale of days, while within-day fidelity was comparatively low when calculated at the scale of hours. While the number of fish departures from the array also varied, the majority of departures for seven of the eight fish did not exceed 1-hr (with the exception of one 47-day departure), suggesting that when departures occurred, the fish did not travel far. Future efforts will significantly expand the number of receivers at Conch Reef such that fish movement behavior relative to the reserve boundaries can be quantified with increased temporal and spatial resolution. (PDF contains 22 pages.)

Movement of yellowtail snapper (Ocyurus chrysurus Block 1790) and black grouper (Mycteroperca bonaci Poey 1860) in the northern Florida Keys National Marine Sanctuary as determined by acoustic telemetry

We tagged a total of 14 yellowtail snapper (Ocyurus chrysurus Bloch 1790) and black grouper (Mycteroperca bonaci Poey 1860) inside the Conch Reef Research Only Area (a no-take marine reserve) in the northern Florida Keys National Marine Sanctuary in November 2001. Both species are heavily exploited in the region. Our objective was to characterize site fidelity and movement behavior along the reef tract to the north and south of the release point. Fishes were collected by baited hook and line from the surface, surgically-tagged with coded-acoustic transmitters, and returned to the reef by snorkelers. Tracking of fish movement behavior was conducted by five acoustic receivers deployed on the seafloor from Davis Reef in the south to Pickles Reef in the north. Fishes were tracked for up to eight months. Results indicated that the majority of signal detections for individual fish from both species were recorded at the two Conch Reef receivers. Limited movement from Conch Reef to Davis Reef was recorded, but no signal detections were recorded at the two sites to the north of Conch Reef. These results suggest that both species show site fidelity to Conch Reef. Future studies will seek to characterize this site fidelity with increased temporal and spatial resolution at Conch Reef. (PDF contains 25 pages.)