Part I Architecture and Algorithms of High Performance Computing High-speed ultra-low-power simulation methodology for material science
نویسندگان
چکیده
Hybrid mathematical principles were introduced to large-scale electronic structure theories andrealized one-hundred-million atom (100-nm-scale) calculations on the K computer. Novellinear-algebraic theories were constructed as Krylov-subspace solvers for generalized shiftedlinear equations ((zS-H)x=b) and were implemented in our order-N calculation code ‘ELSES’(http://www.elses.jp/) with modeled (tight-binding-form) systems. A high parallel efficiencywas obtained with up to the full core calculations on the K computer. The linear-algebraictheory is general and, recently, was applied to ab initio quantum transport calculations on thereal space grid formalism. Moreover, hybrid mathematical priciples are constructed on iInternaleigen-pair calculation with Sylvester's theorem of inertia and hybrid direct solver forgeneralized eigen-value problem. Several application studies were carried out: The nano domainof sp2-sp3 nano-composite carbon solid was calculated for the nano-polycrystalline diamond, aultrahard material and typical domain shapes and boundary structures were obtained. An ionicliquid (N-Methyl-N-propylpiperidinium bis trifluoromethanesulfonyl imide), a battery-relatedmaterial, was calculated and the diffusion mechanism was revealed. Several amourphous-likestructures and bundles of conjugated polymers were calculated and characteristic localized πstates were obtained. The present work was carried out, with several members of ComputicsProject: T. Sogabe, S.-L. Zhang, and T. Miyata (Nagoya U), T. Ono (Tsukuba U), asinter-disciplinary collaborations among physics, mathematics and the high-performancecomputation field or ‘Application-Algorithm-Architecture co-design'. Fig. 1 (a) Concept of Application-Algorithm-Architecture co-design. (b) Theschematic workflow of the order-N algorithm based on the shifted Krylov subspacetheory.
منابع مشابه
A Novel Very High Performance CMOS Current Mirror with extremely low input and ultra high output resistance
In this paper a novel very high performance current mirror is presented. It favorably benefits from such excellent parameters as: Ultra high output resistance (36.9GΩ), extremely low input resistance (0.0058Ω), low output (~0.18V) and low input voltage (~0.18V) operation, very low power consumption (20μW), very low offset current (1pA), ultra wide current dynamic range (150dB), and ultra high a...
متن کاملA Novel ±0.5V Ultra High Current Drive and Output Voltage Headroom Current Output Stage with Very High Output Impedance
A novel ultra-high compliance, low power, very accurate and high output impedance current output stage (COS) with extremely high output current drive capability is proposed in this paper. The principle of operation of this unique structure is discussed, its most important formulas are derived and its outstanding performance is verified by HSPICE simulation in TSMC 0.18µm CMOS, BSIM3, and Level4...
متن کاملHigh Speed Delay-Locked Loop for Multiple Clock Phase Generation
In this paper, a high speed delay-locked loop (DLL) architecture ispresented which can be employed in high frequency applications. In order to design the new architecture, a new mixed structure is presented for phase detector (PD) and charge pump (CP) which canbe triggered by double edges of the input signals. In addition, the blind zone is removed due to the elimination of reset signal. Theref...
متن کاملDesign and Implementation of Digital Demodulator for Frequency Modulated CW Radar (RESEARCH NOTE)
Radar Signal Processing has been an interesting area of research for realization of programmable digital signal processor using VLSI design techniques. Digital Signal Processing (DSP) algorithms have been an integral design methodology for implementation of high speed application specific real-time systems especially for high resolution radar. CORDIC algorithm, in recent times, is turned out to...
متن کاملEnergy Aware Resource Management of Cloud Data Centers
Cloud Computing, the long-held dream of computing as a utility, has the potential to transform a large part of the IT industry, making software even more attractive as a service and shaping the way IT hardware is designed and purchased. Virtualization technology forms a key concept for new cloud computing architectures. The data centers are used to provide cloud services burdening a significant...
متن کاملSymmetrical, Low-Power, and High-Speed 1-Bit Full Adder Cells Using 32nm Carbon Nanotube Field-effect Transistors Technology (TECHNICAL NOTE)
Carbon nanotube field-effect transistors (CNFETs) are a promising candidate to replace conventional metal oxide field-effect transistors (MOSFETs) in the time to come. They have considerable characteristics such as low power consumption and high switching speed. Full adder cell is the main part of the most digital systems as it is building block of subtracter, multiplier, compressor, and other ...
متن کامل