Computer Science

We present a Cholesky factorization for multicore with GPU accelerators systems. The challenges in developing scalable high performance algorithms for these emerging systems stem from their heterogeneity, massive parallelism, and the huge gap between the GPUs' compute power vs the CPU-GPU communication speed. We show an approach that is largely based on software infrastructures that have already been developed for homogeneous multicores and hybrid GPU-based computing.

Recent activities of major chip manufacturers, such as Intel, AMD, IBM and NVIDIA, make it more evident than ever that future designs of microprocessors and large HPC systems will be hybrid/heterogeneous in nature, relying on the integration (in varying proportions) of two major types of components:

Implementations of the Basic Linear Algebra Subprograms (BLAS) interface are major building block of dense linear algebra (DLA) libraries, and therefore have to be highly optimized. We present some techniques and implementations that significantly accelerate the corresponding routines from currently available libraries for GPUs.

The goal of the Matrix Algebra on GPU and Multicore Architectures (MAGMA) project is to create a new generation of linear algebra libraries that achieve the fastest possible time to an accurate solution on hybrid/heterogeneous architectures, starting with current multicore+ multiGPU systems.

Abstract. We present a Cholesky factorization for multicore with GPU accelerators. The challenges in developing scalable high performance algorithms for these emerging systems stem from their heterogeneity, massive parallelism, and the huge gap between the GPUs' compute power vs the CPU-GPU communication speed. We show an approach that is largely based on software infrastructures that have already been developed for homogeneous multicores and hybrid GPU-based computing.

We present a Hessenberg reduction (HR) algorithm for hybrid systems of homogeneous multicore with GPU accelerators that can exceed 25�� the performance of the corresponding LAPACK algorithm running on current homogeneous multicores. This enormous acceleration is due to proper matching of algorithmic requirements to architectural strengths of the system's hybrid components.

Abstract Dense linear algebra (DLA) is one of the most important softwares in high performance computing. It is also important for it's wide usage in other application domains like machine learning, gaming, speech processing, image processing, etc. The introduction of new machines from vendor provides us opportunities to optimize DLA libraries for the new machines and thus exploit their power. Unfortunately the optimization phase is not straightforward all the time.

Abstract We present an improved matrix���matrix multiplication routine (General Matrix Multiply [GEMM]) in the MAGMA BLAS library that targets the NVIDIA Fermi graphics processing units (GPUs) using Compute Unified Data Architecture (CUDA). We show how to modify the previous MAGMA GEMM kernels in order to make a more efficient use of the Fermi's new architectural features, most notably their extended memory hierarchy and memory sizes.

Abstract In this work we propose a joint energy, thermal and cooling management technique (JETC) that significantly reduces per server cooling and memory energy costs. Our analysis shows that decoupling the optimization of cooling energy of CPU & memory and the optimization of memory energy leads to suboptimal solutions due to thermal dependencies between CPU and memory and non-linearity in cooling energy.

CdS thin films of varying thicknesses were deposited on cleaned glass substrates at room temperature by thermal evaporation technique in a vacuum of about 2 x 10 -5 torr. UV-VIS spectra of the films were studied using the optical transmittance measurements which were taken in the spectral region from 300 nm to 1100 nm. The absorbance and reflectance spectra of the films in the UV-VIS region were also studied. Optical constants such as optical band gap, extinction coefficient, refractive index, optical conductivity and complex dielectric constant were evaluated from these spectra. All the films were found to exhibit high transmittance (~ 60 -93 %), low absorbance and low reflectance in the visible/ near infrared region from ~ 500 nm to 1100 nm. The optical band gap energy was found to be in the range 2.28 -2.53 eV. All the films annealed at 300°C for 4 hours in vacuum (~ 10 -2 torr) showed a decrease in the optical transmittance with its absorption edge shifted towards the longer wavelength, leading to the result that the optical band gap decreases on annealing the films. Also, on annealing crystallinity of the films improves, resulting in decrease in the optical transmittance.

The bioelectret state has been proposed to be a universal property of enzymes and to play an important role for the catalytic action of enzymes. In the present investigation the bioelectret state in an enzyme amylase (EC 3.2.2.1) has been studied using the thermally stimulated discharge current (TSDC) technique. The enzyme has been found to be able to manifest the bioelectret state; the TSDC spectrum is characterized by four peaks at temperatures around 170 K, 220 K, 265 K and 300 K. The 265 K peak is attributed to rotation of some polar amino acid residues, whereas bound water molecules are identi®ed as the source of polarization storage for all other peaks. The 170 K peak is ascribed to bound water molecules directly attached to the polypeptide backbone; the 220 K peak is ascribed to a second layer of multiple hydrogen-bonded water molecules, proposed to be a chain-like structure around the polypeptide backbone, and the 300 K peak is attributed to the bound water molecules adsorbed on the surface of the enzyme matrix. The existence of several relaxation polarizations supports the Fro È hlich model of longitudinal polarization waves that are proposed to play an important role for the catalytic action of the enzyme.

We report synthesis of ZnO quantum dot embedded in polyvinylpyrrolidone (PVP) matrix and its functioning as acetone sensor. The specimen is prepared via quenching technique where bulk ZnO powder is calcined at very high temperature of 1200 • C and then quenched into ice cold polyvinylpyrrolidone solution. The samples have been characterized by using UV/vis spectroscopy, X-ray diffraction study, high resolution transmission electron microscopy (HRTEM) and scanning electron microscopy (SEM). These studies infer the sizes of quantum dots to be within 10 nm. The prepared quantum dot samples have been tested for sensing acetone vapour by exploring the variation of their resistance with time at different operating temperatures. It has been revealed that ZnO quantum dots can sense acetone at low operating (230 • C) temperature with less response time.

The rising number of cores in manycore archi tectures, along with technology scaling, results in high power densities and thermal issues on the die. To explore innovative thermal management techniques in such processors, we need an accurate online estimate of the power consumption. In this paper, we present the first ever power model for Intel many integrated core processors, which we use to show the benefit of a novel manycore specific thermal management technique called workload intermixing. Our proposed model leverages per formance monitoring events and accounts for operating voltage and clock frequency. We validate our model for Intel Knights Ferry (K N F) design and show that we have an average of 4.73% prediction error vs. measurements. We provide the breakdown of total power into three main components: compute, memory, and interconnect and use it as an input for thermal management.

The tile QR factorization provides an efficient and scalable way for factoring a dense matrix in parallel on multicore processors. This article presents a way of efficiently implementing the algorithm on a system with a powerful GPU and many multicore CPUs.