scholarly journals Data Access Partitioning for Fine-grain Parallelism on Multicore Architectures

Author(s):  
Michael Chu ◽  
Rajiv Ravindran ◽  
Scott Mahlke
2009 ◽  
Vol 17 (4) ◽  
pp. 309-323 ◽  
Author(s):  
P.E. Hadjidoukas ◽  
G.Ch. Philos ◽  
V.V. Dimakopoulos

In this work we present a runtime threading system which provides an efficient substrate for fine-grain parallelism, suitable for deployment in multicore platforms. Its architecture encompasses a number of optimizations that make it particularly effective in managing a large number of threads and with low overheads. The runtime system has been integrated into an OpenMP implementation to allow for transparent usage under a high level programming paradigm. We evaluate our implementation on two multicore systems using synthetic microbenchmarks and a real-time face detection application.


2019 ◽  
Vol 179 ◽  
pp. 24-53 ◽  
Author(s):  
Shiji Bijo ◽  
Einar Broch Johnsen ◽  
Ka I Pun ◽  
S. Lizeth Tapia Tarifa

2010 ◽  
Vol 18 (1) ◽  
pp. 35-50 ◽  
Author(s):  
Hatem Ltaief ◽  
Jakub Kurzak ◽  
Jack Dongarra ◽  
Rosa M. Badia

The objective of this paper is to describe, in the context of multicore architectures, three different scheduler implementations for the two-sided linear algebra transformations, in particular the Hessenberg and Bidiagonal reductions which are the first steps for the standard eigenvalue problems and the singular value decompositions respectively. State-of-the-art dense linear algebra softwares, such as the LAPACK and ScaLAPACK libraries, suffer performance losses on multicore processors due to their inability to fully exploit thread-level parallelism. At the same time the fine-grain dataflow model gains popularity as a paradigm for programming multicore architectures. Buttari et al. (Parellel Comput. Syst. Appl. 35 (2009), 38–53) introduced the concept oftile algorithmsin which parallelism is no longer hidden inside Basic Linear Algebra Subprograms but is brought to the fore to yield much better performance. Along with efficient scheduling mechanisms for data-driven execution, these tile two-sided reductions achieve high performance computing by reaching up to 75% of the DGEMM peak on a 12000×12000 matrix with 16 Intel Tigerton 2.4 GHz processors. The main drawback of thetile algorithmsapproach for two-sided transformations is that the full reduction cannot be obtained in one stage. Other methods have to be considered to further reduce the band matrices to the required forms.


Author(s):  
R. Sinclair ◽  
B.E. Jacobson

INTRODUCTIONThe prospect of performing chemical analysis of thin specimens at any desired level of resolution is particularly appealing to the materials scientist. Commercial TEM-based systems are now available which virtually provide this capability. The purpose of this contribution is to illustrate its application to problems which would have been intractable until recently, pointing out some current limitations.X-RAY ANALYSISIn an attempt to fabricate superconducting materials with high critical currents and temperature, thin Nb3Sn films have been prepared by electron beam vapor deposition [1]. Fine-grain size material is desirable which may be achieved by codeposition with small amounts of Al2O3 . Figure 1 shows the STEM microstructure, with large (∽ 200 Å dia) voids present at the grain boundaries. Higher quality TEM micrographs (e.g. fig. 2) reveal the presence of small voids within the grains which are absent in pure Nb3Sn prepared under identical conditions. The X-ray spectrum from large (∽ lμ dia) or small (∽100 Ǻ dia) areas within the grains indicates only small amounts of A1 (fig.3).


Author(s):  
Harry Schaefer ◽  
Bruce Wetzel

High resolution 24mm X 36mm positive transparencies can be made from original black and white negatives produced by SEM, TEM, and photomicrography with ease, convenience, and little expense. The resulting 2in X 2in slides are superior to 3¼in X 4in lantern slides for storage, transport, and sturdiness, and projection equipment is more readily available. By mating a 35mm camera directly to an enlarger lens board (Fig. 1), one combines many advantages of both. The negative is positioned and illuminated with the enlarger and then focussed and photographed with the camera on a fine grain black and white film.Specifically, a Durst Laborator 138 S 5in by 7in enlarger with 240/200 condensers and a 500 watt Opale bulb (Ehrenreich Photo-Optical Industries, Inc., New York, NY) is rotated to the horizontal and adjusted for comfortable eye level viewing.


Author(s):  
P. J. Lee ◽  
D. C. Larbalestier

Several features of the metallurgy of superconducting composites of Nb-Ti in a Cu matrix are of interest. The cold drawing strains are generally of order 8-10, producing a very fine grain structure of diameter 30-50 nm. Heat treatments of as little as 3 hours at 300 C (∼ 0.27 TM) produce a thin (1-3 nm) Ti-rich grain boundary film, the precipitate later growing out at triple points to 50-100 nm dia. Further plastic deformation of these larger a-Ti precipitates by strains of 3-4 produces an elongated ribbon morphology (of order 3 x 50 nm in transverse section) and it is the thickness and separation of these precipitates which are believed to control the superconducting properties. The present paper describes initial attempts to put our understanding of the metallurgy of these heavily cold-worked composites on a quantitative basis. The composite studied was fabricated in our own laboratory, using six intermediate heat treatments. This process enabled very high critical current density (Jc) values to be obtained. Samples were cut from the composite at many processing stages and a report of the structure of a number of these samples is made here.


Author(s):  
K. Ogura ◽  
T. Suzuki ◽  
C. Nielsen

In spite of the complicated specimen preparation, Transmission Electron Microscopes (TEM) have traditionally been used for the investigation of the fine grain structures of sintered ceramics. Scanning Electron Microscopes (SEM) have not been used much for the same purpose as TEM because of poor results caused by the specimen charging effect, and also the lack of sufficient resolution. Here, we are presenting a successful result of high resolution imaging of sintered alumina (pure Al2O3) using the Specimen Heated and Electron Beam Induced Conductivity (SHEBIC) method, which we recently reported, in an ultrahigh resolution SEM (UHR-SEM). The JSM-6000F, equipped with a Field Emission Gun (FEG) and an in-lens specimen position, was used for this application.After sintered Al2O3 was sliced into a piece approximately 0.5 mm in thickness, one side was mechanically polished to get a shiny plane for the observation. When the observation was started at 20 kV, an enormous charging effect occured, and it was impossible to obtain a clear Secondary Electron (SE) image (Fig.1).


Sign in / Sign up

Export Citation Format

Share Document