Synergistic Execution of Stream Programs on Multicores with Accelerators

The StreamIt programming model has been proposed to exploit parallelism in streaming applications oil general purpose multicore architectures. The StreamIt graphs describe task, data and pipeline parallelism which can be exploited on accelerators such as Graphics Processing Units (GPUs) or CellBE which support abundant parallelism in hardware. In this paper, we describe a novel method to orchestrate the execution of if StreamIt program oil a multicore platform equipped with an accelerator. The proposed approach identifies, using profiling, the relative benefits of executing a task oil the superscalar CPU cores and the accelerator. We formulate the problem of partitioning the work between the CPU cores and the GPU, taking into account the latencies for data transfers and the required buffer layout transformations associated with the partitioning, as all integrated Integer Linear Program (ILP) which can then be solved by an ILP solver. We also propose an efficient heuristic algorithm for the work-partitioning between the CPU and the GPU, which provides solutions which are within 9.05% of the optimal solution on an average across the benchmark Suite. The partitioned tasks are then software pipelined to execute oil the multiple CPU cores and the Streaming Multiprocessors (SMs) of the GPU. The software pipelining algorithm orchestrates the execution between CPU cores and the GPU by emitting the code for the CPU and the GPU, and the code for the required data transfers. Our experiments on a platform with 8 CPU cores and a GeForce 8800 GTS 512 GPU show a geometric mean speedup of 6.94X with it maximum of 51.96X over it single threaded CPU execution across the StreamIt benchmarks. This is a 18.9% improvement over it partitioning strategy that maps only the filters that cannot be executed oil the GPU - the filters with state that is persistent across firings - onto the CPU.

Nuclear Magnetic-Resonance Spectra Of Oriented Biologically Important Molecules - The Structure Of And The Internal-Rotation In N,N-Dimethyluracil

From the proton NMR spectra of Nfl-dimethyluracil oriented in two different nematic solvents, the internal rotation of the methyl groups about the N-C bonds is studied. It has been observed that the preferred conformation of the methyl group having one carbonyl in the vicinity is the one where a C-H bond is in the ring plane pointing toward the carbonyl group. The results are not sensitive to the mode of rotation of the other methyl group. These data are interpreted in terms of the bond polarizations.

Study of internal motions through N.Q.R. in 6-Chloropyridin-2-ol

Chlorine-35 n.q.r, has been observed for the first time in 6-chloropyridin-2-ol and its temperature dependence has been studied from 77 K to room temperature. The torsional frequencies and their temperature dependences have been calculated by using Bayer's theory with and without Tatsuzaki's modification.

Analysis and computation of the oil film thickness between the piston ring and cylinder liner of an internal combustion engine

A study of the essential features of piston rings in the cylinder liner of an internal combustion engine reveals that the lubrication problem posed by it is basically that of a slider bearing. According to steady-flow-hydrodynamics, viz. Image the oil film thickness becomes zero at the dead centre positions as the velocity, U = 0. In practice, however, such a phenomenon cannot be supported by consideration of the wear rates of pistion rings and cylinder liners. This can be explained by including the “squeeze” action term in the

Conformation of the LL and LD hairpin bends with internal hydrogen bonds in proteins and peptides

The conformation of three linked peptide units having an internal 4 → 1 type of hydrogen bond has been studied in detail, and the low energy conformations are listed. These conformations all lead to the reversal of the chain direction, and may therefore be called as “hairpin bends” or “U-bends”. Since this bend can occur at the end of two chains hydrogen-bonded in the antiparallel β-conformation, it is also known as the “β-bend”. Two types of conformation are possible when the residues at the second and third Cα atoms are both of type L (the LL bend), while only one type is possible for the LD and the DL bend. The LL bend can also accommodate the sequences LG, GL, GG (G = glycine), while the LD bend can accommodate the sequences LG, GD and GG. The conformations for the sequences DD and DL are exact inverses (or mirror images) of those for the sequences LL and LD, respectively, and have dihedral angles (phi2, ψ2), (phi3, ψ3) of the same magnitudes, but of opposite signs as those for the former types, which are listed, along with the characteristics (length, angle and energy) of the hydrogen bonds. A comparison of the theoretical predictions with experimental data (from X-ray diffraction and NMR studies) on proteins and peptides, show reasonably good agreement. However, a systematic trend is observable in the experimental data, slightly deviating from theory, which indicates that some deformations occur in the shapes of the peptide units forming the bend, differing from that of the standard planar peptide unit.

Effect of viscosity on internal waves from a source in a wall

The solution for a line source of oscillatory strength kept at the origin in a wall bounding a semi-infinite viscous imcompressible stratified fluid is presented in an integral form. The behaviour of the flow at far field and near field is studied by an asymptotic expansion procedure. The streamlines for different parameters are drawn and discussed. The real characteristic straight lines present in the inviscid problem are modified by the viscosity and the solutions obtained are valid even at the resonance frequency.

Primary structure of a Thomsen-Friedenreich-antigen-specific lectin, jacalin [Artocarpus integrifolia (jack fruit) agglutinin]. Evidence for the presence of an internal repeat

Jacalin [Artocarpus integrifolia (jack fruit) agglutinin] is made up of two types of chains, heavy and light, with M(r) values of 16,200 +/- 1200 and 2090 +/- 300 respectively (on the basis of gel-permeation chromatography under denaturing conditions). Its complete amino acid sequence was determined by manual degradation using a 4-dimethylaminoazobenzene 4'-isothiocyanate double-coupling method. Peptide fragments for sequence analysis were obtained by chemical cleavages of the heavy chain with CNBr, hydroxylamine hydrochloride and iodosobenzoic acid and enzymic cleavage with Staphylococcus aureus proteinase. The peptides were purified by a combination gel-permeation and reverse-phase chromatography. The light chains, being only 20 residues long, could be sequenced without fragmentation. Amino acid analyses and carboxypeptidase-Y-digestion C-terminal analyses of the subunits provided supportive evidence for their sequence. Computer-assisted alignment of the jacalin heavy-chain sequence failed to show sequence similarity to that of any lectin for which the complete sequence is known. Analyses of the sequence showed the presence of an internal repeat spanning residues 7-64 and 76-130. The internal repeat was found to be statistically significant.

An algorithm to find similar internal sequence repeats

In recent years, identification of sequence patterns has been given immense importance to understand better their significance with respect to genomic organization and evolutionary processes. To this end, an algorithm has been derived to identify all similar sequence repeats present in a protein sequence. The proposed algorithm is useful to correlate the three-dimensional structure of various similar sequence repeats available in the Protein Data Bank against the same sequence repeats present in other databases like SWISS-PROT, PIR and Genome databases.

Boundary-layer effects on internal flow choking in dual-thrust solid rocket motors

Theoretical studies have been carried out to examine internal flow choking in the inert simulators of a dual-thrust motor. Using a two-dimensional k-omega turbulence model, detailed parametric studies have been carried out to examine aerodynamic choking and the existence of a fluid throat at the transition region during the startup transient of dual-thrust motors. This code solves standard k-omega turbulence equations with shear flow corrections using a coupled second-order-implicit unsteady formulation. In the numerical study, a fully implicit finite volume scheme of the compressible, Reynolds-averaged, Navier-Stokes equations is employed. It was observed that, at the subsonic inflow conditions, there is a possibility of the occurrence of internal flow choking in dual-thrust motors due to the formation of a fluid throat at the beginning of the transition region induced by area blockage caused by boundary-layer-displacement thickness. It has been observed that a 55% increase in the upstream port area of the dual-thrust motor contributes to a 25% reduction in blockage factor at the transition region, which could negate the internal How choking and supplement with an early choking of the dual-thrust motor nozzle. If the height of the upstream port relative to the motor length is too small, the developing boundary layers from either side of the port can interact, leading to a choked,flow. On the other hand, if the developing boundary layers are far enough apart, then choking does not occur. The blockage factor is greater in magnitude for the choked case than for the unchoked case. More tangible explanations are presented in this paper for the boundary-layer blockage and the internal flow choking in dual-thrust motors, which hitherto has been unexplored.

1H NMR study of internal motions and quantum rotational tunneling in (CH3)4NGeCl3

(CH3)4NGeCl3 is prepared, characterized and studied using 1H NMR spin lattice relaxation time and second moment to understand the internal motions and quantum rotational tunneling. Proton second moment is measured at 7 MHz as function of temperature in the range 300-77 K and spin lattice relaxation time (T1) is measured at two Larmor frequencies, as a function of temperature in the range 270-17 K employing a homemade wide-line/pulsed NMR spectrometers. T1 data are analyzed in two temperature regions using relevant theoretical models. The relaxation in the higher temperatures (270-115 K) is attributed to the hindered reorientations of symmetric groups (CH3 and (CH3)4N). Broad asymmetric T1 minima observed below 115 K down to 17 K are attributed to quantum rotational tunneling of the inequivalent methyl groups.

Three-level inverter fed open-end winding IM drive with common mode voltage elimination and reduced power device count

A three-level space phasor generation scheme with common mode elimination and with reduced power device count is proposed for an open end winding induction motor in this paper. The open end winding induction motor is fed by the three-level inverters from both sides. Each two level inverter is formed by cascading two two-level inverters. By sharing the bottom inverter for the two three-level inverters on either side, the power device count is reduced. The switching states with zero common mode voltage variation are selected for PWM switching so that there is no alternating common mode voltage in the pole voltages as well as in phase voltages. Only two isolated DC-links, with half the voltage rating of a conventional three-level neutral point clamped inverter, are needed for the proposed scheme.

Reliability assessment of internal stability of reinforced soil structures: A pseudo-dynamic approach

A methodology for reliability based optimum design of reinforced soil structures subjected to horizontal and vertical sinusoidal excitation based on pseudo-dynamic approach is presented. The tensile strength of reinforcement required to maintain the stability is computed using logarithmic spiral failure mechanism. The backfill soil properties, geometric and strength properties of reinforcement are treated as random variables. Effects of parameters like soil friction angle, horizontal and vertical seismic accelerations, shear and primary wave velocities, amplification factors for seismic acceleration on the component and system probability of failures in relation to tension and pullout capacities of reinforcement have been discussed. In order to evaluate the validity of the present formulation, static and seismic reinforcement force coefficients computed by the present method are compared with those given by other authors. The importance of the shear wave velocity in the estimation of the reliability of the structure is highlighted. The Ditlevsen's bounds of system probability of failure are also computed by taking into account the correlations between three failure modes, which is evaluated using the direction cosines of the tangent planes at the most probable points of failure. (c) 2009 Elsevier Ltd. All rights reserved.

Software Pipelined Execution of Stream Programs on GPUs

The StreamIt programming model has been proposed to exploit parallelism in streaming applications on general purpose multi-core architectures. This model allows programmers to specify the structure of a program as a set of filters that act upon data, and a set of communication channels between them. The StreamIt graphs describe task, data and pipeline parallelism which can be exploited on modern Graphics Processing Units (GPUs), as they support abundant parallelism in hardware. In this paper, we describe the challenges in mapping StreamIt to GPUs and propose an efficient technique to software pipeline the execution of stream programs on GPUs. We formulate this problem - both scheduling and assignment of filters to processors - as an efficient Integer Linear Program (ILP), which is then solved using ILP solvers. We also describe a novel buffer layout technique for GPUs which facilitates exploiting the high memory bandwidth available in GPUs. The proposed scheduling utilizes both the scalar units in GPU, to exploit data parallelism, and multiprocessors, to exploit task and pipelin parallelism. Further it takes into consideration the synchronization and bandwidth limitations of GPUs, and yields speedups between 1.87X and 36.83X over a single threaded CPU.