Conventional Supercomputers Research Articles

A quantum computing concept for 1-D elastic wave simulation with exponential speedup

SUMMARY Quantum computing has attracted considerable attention in recent years because it promises speedups that conventional supercomputers cannot offer, at least for some applications. Though existing quantum computers (QCs) are, in most cases, still too small to solve significant problems, their future impact on domain sciences is already being explored now. Within this context, we present a quantum computing concept for 1-D elastic wave propagation in heterogeneous media with two components: a theoretical formulation and an implementation on a real QC. The method rests on a finite-difference approximation, followed by a sparsity-preserving transformation of the discrete elastic wave equation to a Schrödinger equation, which can be simulated directly on a gate-based QC. An implementation on an error-free quantum simulator verifies our approach and forms the basis of numerical experiments with small problems on the real QC IBM Brisbane. The latter produce simulation results that qualitatively agree with the error-free version but are contaminated by quantum decoherence and noise effects. Complementing the discrete transformation to the Schrödinger equation by a continuous version allows the replacement of finite differences by other spatial discretization schemes, such as the spectral-element method. Anticipating the emergence of error-corrected quantum chips, we analyse the computational complexity of the best quantum simulation algorithms for future QCs. This analysis suggests that our quantum computing approach may lead to wavefield simulations that run exponentially faster than simulations on classical computers.

Sampling a Rare Protein Transition Using Quantum Annealing.

Simulating spontaneous structural rearrangements in macromolecules with classical molecular dynamics is an outstanding challenge. Conventional supercomputers can access time intervals of up to tens of μs, while many key events occur on exponentially longer time scales. Path sampling techniques have the advantage of focusing the computational power on barrier-crossing trajectories, but generating uncorrelated transition paths that explore diverse conformational regions remains a problem. We employ a hybrid path-sampling paradigm that addresses this issue by generating trial transition paths using a quantum annealing (QA) machine. We first employ a classical computer to perform an uncharted exploration of the conformational space. The data set generated in this exploration is then postprocessed using a path integral-based method to yield a coarse-grained network representation of the reactive kinetics. By resorting to a quantum annealer, quantum superposition can be exploited to encode all of the transition pathways in the initial quantum state, thus potentially solving the path exploration problem. Furthermore, each QA cycle yields a completely uncorrelated trial trajectory. We previously validated this scheme on a prototypically simple transition, which could be extensively characterized on a desktop computer. Here, we scale up in complexity and perform an all-atom simulation of a protein conformational transition that occurs on the millisecond time scale, obtaining results that match those of the Anton special-purpose supercomputer. Despite limitations due to the available quantum annealers, our study highlights how realistic biomolecular simulations provide potentially impactful new ground for applying, testing, and advancing quantum technologies.

Variational Quantum Algorithms for Computational Fluid Dynamics

Quantum computing uses the physical principles of very small systems to develop computing platforms which can solve problems that are intractable on conventional supercomputers. There are challenges not only in building the required hardware but also in identifying the most promising application areas and developing the corresponding quantum algorithms. The availability of intermediate-scale noisy quantum computers is now propelling the developments of novel algorithms, with applications across a variety of domains, including in aeroscience. Variational quantum algorithms are particularly promising because they are comparatively noise tolerant and aim to achieve a quantum advantage with only a few hundred qubits. Furthermore, they are applicable to a wide range of optimization problems arising throughout the natural sciences and industry. To demonstrate the possibilities for the aeroscience community, we give a perspective on how variational quantum algorithms can be used in computational fluid dynamics. We discuss how classical problems are translated into quantum algorithms and their logarithmic scaling with problem size. For an explicit example, we apply this method to Burgers’s equation in one spatial dimension. We argue that a quantum advantage over classical computing methods could be achieved by the end of this decade if quantum hardware progresses as currently envisaged and emphasize the importance of joining up development of quantum algorithms with application-specific expertise to achieve a real-world impact.

Photonic Quantum Computers Enlighten the World: A review of their development, types, and applications

It has been demonstrated that the photonic quantum computer is significantly faster than conventional supercomputers and that the practical quantum computer is one of the most promising ways to solve real-life problems. In this article, the development of photonic quantum computers and their potential applications are summarized, and three types of photonic quantum computing machines are detailed, including photonic quantum machines, coherent Ising machines (CIMs), and programmable photonic quantum computers. The photonic quantum industry, together with some start-up companies in associated application fields, are profiled. Compared with superconducting and ion traps, photonic quantum computing has its own advantages and disadvantages. It will be seen whether photonic quantum computer companies can capture a share of the future quantum computing market. The major challenges lie in the scalability of photonic systems and their adaptation and integration with silicon (Si) systems.

A high-performance compilation strategy for multiplexing quantum control architecture

Quantum computers have already shown significant potential to solve specific problems more efficiently than conventional supercomputers. A major challenge towards noisy intermediate-scale quantum computing is characterizing and reducing the various control costs. Quantum programming describes the process of quantum computation as a sequence, whose elements are selected from a finite set of universal quantum gates. Quantum compilation translates quantum programs to ordered pulses to the quantum control devices subsequently and quantum compilation optimization provides a high-level solution to reduce the control cost efficiently. Here, we propose a high-performance compilation strategy for multiplexing quantum control architecture. For representative benchmarks, the utilization efficiency of control devices increased by 49.44% on average in our work, with an acceptable circuit depth expansion executing on several real superconducting quantum computers of IBM.

How to plan for a post-quantum security ecosystem

Sooner or later, companies will have to adapt to security in an age of quantum computing. Even the slowest of these new machines can perform calculations millions of times faster than the fastest conventional supercomputers, meaning that many of the encryption techniques on which digital security is built could become insecure. This would result in confidential information being revealed and digitally signed documents rendered untrustworthy.

MapReduce SVM Game

Abstract Despite technological advances making computing devices faster, smaller, and more prevalent in today's age, data generation and collection has outpaced data processing capabilities. Simply having more compute platforms does not provide a means of addressing challenging problems in the big data era. Rather, alternative processing approaches are needed and the application of machine learning to big data is hugely important. The MapReduce programming paradigm is an alternative to conventional supercomputing approaches, and requires less stringent data passing constrained problem decompositions. Rather, MapReduce relies upon defining a means of partitioning the desired problem so that subsets may be computed independently and recom- bined to yield the net desired result. However, not all machine learning algorithms are amenable to such an approach. Game-theoretic algorithms are often innately distributed, consisting of local interactions between players without requiring a central authority and are iterative by nature rather than requiring extensive retraining. Effectively, a game-theoretic approach to machine learning is well suited for the MapReduce paradigm and provides a novel, alternative new perspective to addressing the big data problem. In this paper we present a variant of our Support Vector Machine (SVM) Game classifier which may be used in a distributed manner, and show an illustrative example of applying this algorithm.

MODYLAS: A highly parallelized general‐purpose molecular dynamics simulation program

A large number of molecular dynamics (MD) simulations have been carried out so far on personal computer clusters and conventional supercomputers using general‐purpose MD software such as NAMD, GROMACS, CHARMM, and AMBER. The development of MD simulation program is closely related to the architecture of the computers. Recent trend of MD simulation is headed to large‐scale calculation, long‐time calculation of small systems, and calculation for large statistics. The most challenging one is the large‐scale calculation. To realize the large‐scale calculation, we have to efficiently control the massively parallel supercomputers with several 10 thousands of nodes. There, it is essential to develop software having a compatibility of such computer architecture. Recently, we have developed general‐purpose MD software MODYLAS (MOlecular DYnamics simulation software for LArge Systems) for massively parallel supercomputers such as K‐computer. Here, we describe the outline, important features, and computation of MODYLAS. An example result of the 200 ns‐MD simulation of the viral capsid system consisting of about 6,500,000 atoms is also presented. © 2014 Wiley Periodicals, Inc.

Adaptation of fluid model EULAG to graphics processing unit architecture

SummaryThe goal of this study is to adapt the multiscale fluid solver EULerian or LAGrangian framewrok (EULAG) to future graphics processing units (GPU) platforms. The EULAG model has the proven record of successful applications, and excellent efficiency and scalability on conventional supercomputer architectures. Currently, the model is being implemented as the new dynamical core of the COSMO weather prediction framework. Within this study, two main modules of EULAG, namely the multidimensional positive definite advection transport algorithm (MPDATA) and the variational generalized conjugate residual, elliptic pressure solver Generalized Conjugate Residual (GCR) are analyzed and optimized. In this paper, a method is proposed, which ensures a comprehensive analysis of the resource consumption including registers, shared, and global memories. This method allows us to identify bottlenecks of the algorithm, including data transfers between host and global memory, global and shared memories, as well as GPU occupancy. We put the emphasis on providing a fixed memory access pattern, padding as well as organizing computation in the MPDATA algorithm. The testing and validation of the new GPU implementation have been carried out based on modeling decaying turbulence of a homogeneous incompressible fluid in a triply‐periodic cube. Simulations performed using the standard version of EULAG and its new GPU implementation give similar solutions. Preliminary results show a promising increase in terms of computational efficiency. Copyright © 2014 John Wiley & Sons, Ltd.

A Review of Hardware Acceleration for Computational Genomics

Hardware accelerators are becoming increasingly commonplace in delivering high performance computing solutions at a fraction of the cost of conventional supercomputers and standalone CPU clusters, despite the additional programming effort required to utilize them. This paper provides a survey on the use of hardware accelerators such as FPGAs and GPUs in the area of biological sequence analysis, particularly in the domain of computational genomics. We also survey research on hardware acceleration in response to emerging trends in high-throughput sequencing, and applications enabled by it. We conclude the survey with remarks on how these trends influence the use of hardware acceleration in bioinformatics, and the role of recently developed or soon to be released accelerator technologies.

Improved DNA-sticker arithmetic: tube-encoded-carry, Logarithmic Number System and Monte-Carlo methods

The sticker model of computation, implemented using robotic processing of DNA, manipulates in parallel many bitstrings, called strands, that are contained in a limited number of tubes. Prior sticker arithmetic algorithms, patterned on digital-electronics, generate carry bits in the strand, either wasting bits or using a clear operation (with problematic biochemical implementation). The novel addition algorithm here does not need to record the carry. Instead, which tube holds a particular strand implicitly encodes the carry. The speed and number of tubes are half that of prior approaches. Further encoding data in the Logarithmic Number System (LNS) allows such integer operations to perform cost-effective real multiplications, divisions and roots. An example LNS Euclidian norm is more efficient than prior methods, assuming perfect operations. Unfortunately, DNA-stickers are unreliable. This paper uses sticker unreliability as a source of randomness to implement Monte-Carlo (MC) arithmetic (previously fabricated in silicon at the cost of pseudo-random generators). With stickers, the randomness is free. MC engineering mimics natural systems using unreliable but redundant components. Here, MC randomness is only useful in low-order bits. Multiple re-testing of the same bit ("refinement") trades improved reliability for slower operation using more tubes. Simulations (with different sizes, probabilities and refinement) show that increasing refinement as a function of bit position allows imperfect implementations to achieve suitable MC strand-error distributions, predicting 1000x speed-mass advantage of sticker-MCLNS over conventional supercomputers.

Distributed computing as a virtual supercomputer: Tools to run and manage large-scale BOINC simulations

Distributed computing (DC) projects tackle large computational problems by exploiting the donated processing power of thousands of volunteered computers, connected through the Internet. To efficiently employ the computational resources of one of world's largest DC efforts, GPUGRID, the project scientists require tools that handle hundreds of thousands of tasks which run asynchronously and generate gigabytes of data every day. We describe RBoinc, an interface that allows computational scientists to embed the DC methodology into the daily work-flow of high-throughput experiments. By extending the Berkeley Open Infrastructure for Network Computing (BOINC), the leading open-source middleware for current DC projects, with mechanisms to submit and manage large-scale distributed computations from individual workstations, RBoinc turns distributed grids into cost-effective virtual resources that can be employed by researchers in work-flows similar to conventional supercomputers. The GPUGRID project is currently using RBoinc for all of its in silico experiments based on molecular dynamics methods, including the determination of binding free energies and free energy profiles in all-atom models of biomolecules.

A directive-based MPI code generator for Linux PC clusters

Computation requirements in scientific fields are getting heavier and heavier. The advent of clustering systems provides an affordable alternative to expensive conventional supercomputers. However,...

Scientific Application Performance On Leading Scalar and Vector Supercomputering Platforms

The last decade has witnessed a rapid proliferation of superscalar cache-based microprocessors to build high-end computing (HEC) platforms, primarily because of their generality, scalability, and cost effectiveness. However, the growing gap between sustained and peak performance for full-scale scientific applications on conventional supercomputers has become a major concern in high performance computing, requiring significantly larger systems and application scalability than implied by peak performance in order to achieve desired performance. The latest generation of custom-built parallel vector systems have the potential to address this issue for numerical algorithms with sufficient regularity in their computational structure. In this work we explore applications drawn from four areas: magnetic fusion (GTC), plasma physics (LB-MHD-3D), astrophysics (Cactus), and material science (PARATEC). We compare performance of the vector-based Cray X1, X1E, Earth Simulator, NEC SX-8, with performance of three leading commodity-based super-scalar platforms utilizing the IBM Power3, Intel Itanium2, and AMD Opteron processors. Our work makes several significant contributions: a new data-decomposition scheme for GTC that (for the first time) enables a breakthrough of the teraflop barrier; the introduction of a new three-dimensional lattice Boltzmann magneto-hydrodynamic implementation used to study the onset evolution of plasma turbulence that achieves over 26 Tflop/s on 4800 ES processors; the highest per processor performance (by far) achieved by the full-production version of the Cactus ADM-BSSN; and the largest PARATEC cell size atomistic simulation to date. Overall, results show that the vector architectures attain unprecedented aggregate performance across our application suite, demonstrating the tremendous potential of modern parallel vector systems.

Supercomputing moves to universities and makes possible new ways to organize computational research

Too cheap to meter? Surely not. But microprocessors have come a long way in delivering inexpensive computing power.The development is eroding the distinction between computers made from low‐cost microprocessors, and more conventional supercomputers operated—for example—at the U.S. National Science Foundation's Supercomputer Centers. It is also shifting the balance in high‐performance computing (HPC) away from the NSF's Supercomputer Centers, to powerful, local computer facilities run by academic departments and universities.

Heterogeneous parallel and distributed computing

Heterogeneous network-based distributed and parallel computing is gaining increasing acceptance as an alternative or complementary paradigm to multiprocessor-based parallel processing as well as to conventional supercomputing. While algorithmic and programming aspects of heterogeneous concurrent computing are similar to their parallel processing counterparts, system issues, partitioning and scheduling, and performance aspects are significantly different. In this paper, we discuss the evolution of heterogeneous concurrent computing, in the context of the parallel virtual machine (PVM) system, a widely adopted software system for network computing. In particular, we highlight the system level infrastructures that are required, aspects of parallel algorithm development that most affect performance, system capabilities and limitations, and tools and methodologies for effective computing in heterogeneous networked environments. We also present recent developments and experiences in the PVM project, and comment on ongoing and future work.

Programming Environment for a High-Performance Parallel Supercomputer with Intelligent Communication

At the Electronics Laboratory of the Swiss Federal Institute of Technology (ETH) in Zürich, the high-performance parallel supercomputer MUSIC (MUlti processor System with Intelligent Communication) has been developed. As applications like neural network simulation and molecular dynamics show, the Electronics Laboratory supercomputer is absolutely on par with those of conventional supercomputers, but electric power requirements are reduced by a factor of 1,000, weight is reduced by a factor of 400, and price is reduced by a factor of 100. Software development is a key issue of such parallel systems. This article focuses on the programming environment of the MUSIC system and on its applications.

The Supercomputer Supernet: A Scalable Distributed Terabit Network

Conventional supercomputer interconnection networks consist of crossbar modules, which are connected by point-to-point copper or fiber links to create distributed mesh topologies (e.g., CP*, Nectar). This type of “physical networking” topology creates cable layout problems, dealing with bundles of cables/fibers between various pairs of modules. It also introduces several routing hops, increasing the probability of interference between connections and making it difficult to guarantee quality of service to real time applications. We describe a new network called the Supercomputer Supernet (SSN) that attempts to overcome these problems by replacing the point-to-point links with a fiber optic interconnect system. The novel scheme employs asynchronous pipeline crossbar switches (APCS) used in parallel supercomputers to interconnect multi-channel wavelength division multiplexed (WDM) fiber optic links to an optical star (or tree) “physical” topology. WDM will be used to subdivide the very large fiber bandwidth into several channels each of Gb/s bandwidth. WDM channels (supporting also time division multiplexing) will be established between modules, thus defining a dense “virtual” interconnection topology, which is dynamically reconfigurable, responding to changing traffic patterns. A pool of channels will be set aside for direct, end-to-end connections between crossbars, providing circuit-switched service for real-time traffic applications.

Achieving supercomputer performane for neural net simulation with an array of digital signal processors

Music, a digital signal processor (DSP)-based system with a parallel distributed-memory architecture that provides enormous computing power yet retains the flexibility of a general-purpose computer, is discussed. It is shown that Music reaches a peak performance of 2.7 Gflops at a significantly lower cost, power consumption, and space requirement than conventional supercomputers. The Music system hardware, programming, and backpropagation implementation are described.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Large-scale ab initio total energy calculations on parallel computers

An implementation of a set of total energy plane-wave pseudopotential codes on a parallel computer is described which allows calculations to be performed for systems containing many hundreds of atoms in the unit cell. Possible parallelisation strategies are discussed and it is shown that assigning parts of real and Fourier space across the processors is the least restricted approach. The performance of our parallel codes is demonstrated by timing tests carried out on several i860-based parallel machines and these are compared with tests performed on conventional sequential supercomputers. Ab initio computations on systems which are beyond the power of conventional supercomputers as well as ew perspectives for first-principles molecular dynamics are discussed.