Large Number Of Cores Research Articles

BLAS Kütüphanelerinin GPU Mimarilerindeki Nicel Performans Analizi

Basic Linear Algebra Subprograms (BLAS) are a set of linear algebra routines commonly used by machine learning applications and scientific computing. BLAS libraries with optimized implementations of BLAS routines offer high performance by exploiting parallel execution units in target computing systems. With massively large number of cores, graphics processing units (GPUs) exhibit high performance for computationally-heavy workloads. Recent BLAS libraries utilize parallel cores of GPU architectures efficiently by employing inherent data parallelism. In this study, we analyze GPU-targeted functions from two BLAS libraries, cuBLAS and MAGMA, and evaluate their performance on a single-GPU NVIDIA architecture by considering architectural features and limitations. We collect architectural performance metrics and explore resource utilization characteristics. Our work aims to help researchers and programmers to understand the performance behavior and GPU resource utilization of the BLAS routines implemented by the libraries.

High-performance Deterministic Concurrency Using Lingua Franca

Actor frameworks and similar reactive programming techniques are widely used for building concurrent systems. They promise to be efficient and scale well to a large number of cores or nodes in a distributed system. However, they also expose programmers to nondeterminism, which often makes implementations hard to understand, debug, and test. The recently proposed reactor model is a promising alternative that enables deterministic concurrency. In this article, we present an efficient, parallel implementation of reactors and demonstrate that the determinacy of reactors does not imply a loss in performance. To show this, we evaluate Lingua Franca (LF), a reactor-oriented coordination language. LF equips mainstream programming languages with a deterministic concurrency model that automatically takes advantage of opportunities to exploit parallelism. Our implementation of the Savina benchmark suite demonstrates that, in terms of execution time, the runtime performance of LF programs even exceeds popular and highly optimized actor frameworks. We compare against Akka and CAF, which LF outperforms by 1.86× and 1.42×, respectively.

Hybridised multigrid preconditioners for a compatible finite‐element dynamical core

AbstractCompatible finite‐element discretisations for the atmospheric equations of motion have recently attracted considerable interest. Semi‐implicit timestepping methods require the repeated solution of a large saddle‐point system of linear equations. Preconditioning this system is challenging, since the velocity mass matrix is nondiagonal, leading to a dense Schur complement. Hybridisable discretisations overcome this issue: weakly enforcing continuity of the velocity field with Lagrange multipliers leads to a sparse system of equations, which has a similar structure to the pressure Schur complement in traditional approaches. We describe how the hybridised sparse system can be preconditioned with a non‐nested two‐level preconditioner. To solve the coarse system, we use the multigrid pressure solver that is employed in the approximate Schur complement method previously proposed by the some of the authors. Our approach significantly reduces the number of solver iterations. The method shows excellent performance and scales to large numbers of cores in the Met Office next‐generation climate and weather prediction model LFRic.

Life cycle assessment of injection moulding tools and multicomponent plastic cap production

Long-term product improvement requires detailed analysis that includes information from the entire product life cycle. Life cycle assessment according to the standards ISO 14040 and ISO 14044 provides information on the environmental impact of products throughout their life cycle. The aim of the study is to analyze the environmental impact of a multicomponent plastic product in two variants and the associated customized tools using life cycle assessment. The subject of this analysis is the previous and the improved design of a multicomponent plastic cap for 19 l water bottles and the associated custom-manufactured tools. The main improvements of the custom-made tools are in the larger number of cores, and the new cap design was improved with fewer components and mass. The results show that the production and packaging of the improved multicomponent plastic cap has more than two times lower environmental impacts in the categories of global warming potential, freshwater eutrophication, terrestrial acidification ozone formation, human health and non reneweable, fossil. The environmental impact of custom injection moulding tools is strongly influenced by the capacity of the injection moulding machine and the number of cores or number of products that can be produced in a batch. In addition, the results of the improved 19 l multicomponent plastic cap showed a lower environmental impact compared to previous studie of the 5 l plastic cap.

Development characteristics and controlling factors of dolomite reservoirs of permian Qixi a Formation in central Sichuan Basin

In recent years, with the fine exploration of carbonate reservoirs in the Qixia Formation, central Sichuan Basin, researchers have made important oil and gas discoveries in the dolomite reservoirs. However, the characteristics, genesis and controlling factors of dolomite reservoirs in the Qixia Formation in this area are still unclear. In this paper, the petrology, pore structures, physical properties and geochemical characteristics of the dolomite reservoirs in the Permian Qixia Formation in the central Sichuan Basin have been systematically studied based on a large number of cores, thin sections, physical property and geochemical tests. Furthermore, the genesis and main controlling factors of dolomite reservoirs are clarified. The study shows that the dolomites have the characteristics of small single-layer thickness and multi-layer development, and they are usually interbeded with the limestones. The reservoir types are mainly fine crystal and fine medium crystal dolomite, and the reservoir spaces include intercrystalline pores and dissolution pores (or caves). Moreover, the reservoir physical properties are characterized by medium porosity and medium to high permeability. Quasi-syngenetic dolomitization is the main origin of dolomite in the target layer, and the dolomite was slightly modified by hydrothermal solution at the end of the Maokou Formation period. The development of dolomite reservoirs in the Qixia Formation in the study area is affected by palaeogeomorphology, sedimentation and diagenesis. The granular shoal facies developed in the high parts of the paleogeomorphology provides the material basis for the formation of dolomite reservoirs. The high frequency sequence interface controls the development of dolomite reservoirs in the highstand systems tract. In addition, the quasi-syngenetic dissolution promots the development of the early secondary pores and provids an effective channel for the migration of the diagenetic fluids in the later stage. In the direction of orthogonal horizontal deposition, the dolomites have the characteristics of thin monolayer thickness, multiple sequences and are interbedded with limestone. The dolomite shoals in the study area are distributed as clumps in the plane, which has great potential for hydrocarbon exploration.

Logging identification of complex lithology of the Lower Jurassic Da'anzhai Member in the eastern slope of the western Sichuan Depression

In recent years, with the in-depth development of oil and gas exploration in the eastern slope of the western Sichuan Depression, breakthroughs have been made in the research on high-quality gas reservoirs in the Da'anzhai Member of the Ziliujing Formation in this area. Fine logging identification of lithology in the Da'anzhai Member in the eastern slope of the western Sichuan Depression is the core subject of reservoir sweet spot prediction. In this paper, taking the Da'anzhai Member in the western Sichuan Basin as an example, the work flow and comparisons of multi-lithology logging models have been systematically conducted, using a large number of cores, geological data and logging interpretation models. The research shows that there are four main lithologies developed in the Da'anzhai Member, namely shale, marl-bearing limestone, shell limestone and sandstone. Intersection and spider diagram methods can effectively screen out the logging parameters that are sensitive to lithology, including natural gamma (GR), neutron (CNL), deep lateral resistivity (RD), and acoustic wave time difference (AC). The lithology of the Da'anzhai Member in the study area can be identified with high precision via the "lithology probability factor", "BP neural network" and cluster analysis methods. Among them, the "lithologic probability factor" and "BP neural network" methods have a prediction accuracy of lithology exceeds 80%. Therefore, these two methods are optimized as the most effective methods for logging identification of lithology in the Da'anzhai Member. This study has certain reference value for the lithology identification of similar gas reservoirs worldwide. • The calculation and comparison of multi-lithology logging models have been systematically conducted, using a large number of cores, geological data and logging interpretation models. • The lithology of the Da'anzhai Member in the study area can be identified with high precision by the "lithology probability factor" method, the "BP neural network" method and the cluster analysis method. • A complex lithology identification model is established.

Massively parallelized interpolated factored Green function method

This paper presents the first parallel implementation of the novel “Interpolated Factored Green Function” (IFGF) method introduced recently for the accelerated evaluation of discrete integral operators arising in wave scattering and other areas (Bauinger and Bruno, Jour. Computat. Phys., 2021). On the basis of the hierarchical IFGF interpolation strategy, the proposed (hybrid MPI-OpenMP) parallel implementation results in efficient data communication, and it scales up to large numbers of cores—without any hard limitations on the number of cores efficiently employed. Moreover, on any given number of cores, the proposed parallel approach preserves the O(Nlog⁡N) computing cost inherent in the sequential version of the IFGF algorithm. Unlike other approaches, the IFGF method does not utilize the Fast Fourier Transform (FFT), and it is thus better suited for efficient parallelization in distributed-memory computer systems. In particular, the IFGF method relies on a “peer-to-peer” strategy wherein, at every level, field propagation is directly enacted via “exchanges” between “peer” polynomials of constant degree, without data accumulation in large-scale “telephone-central” mathematical constructs such as those used in the Fast Multipole Method (FMM) and pure FFT-based approaches. A variety of numerical results presented in this paper illustrate the character of the proposed parallel algorithm, in particular demonstrating scaling from 1 to all 1,680 cores available in the High Performance Computing cluster used, and for problems of up to 4,096 wavelengths in acoustic size.

Combining reduction with synchronization barrier on multi‐core processors

SummaryWith the rise of multi‐core processors with a large number of cores, the need for shared memory reduction that performs efficiently on a large number of cores is more pressing. Efficient shared memory reduction on these multi‐core processors will help share memory programs be more efficient. In this article, we propose a reduction method combined with a barrier method that uses SIMD read/write instructions to combine barrier signaling and reduction value to minimize memory/cache traffic between cores, thereby reducing barrier latency. We compare different barriers and reduction methods on three multi‐core processors and show that the proposed combining barrier/reduction methods are 4 and 3.5 times faster than respectively GCC 11.1 and Intel 21.2 OpenMP 4.5 reduction.

Improving Cache Utilization of Nested Parallel Programs by Almost Deterministic Work Stealing

Nested (fork-join) parallelism eases parallel programming by enabling high-level expression of parallelism and leaving the mapping between parallel tasks and hardware to the runtime scheduler. A challenge in dynamic scheduling of nested parallelism is how to exploit data locality, which has become more demanding in the deep cache hierarchies of modern processors with a large number of cores. This paper introduces <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">almost deterministic work stealing (ADWS)</i> , which efficiently exploits data locality by deterministically planning a cache-hierarchy-aware schedule, while allowing a little scheduling variety to facilitate dynamic load balancing. Furthermore, we propose an extension of our prior work on ADWS to achieve better shared cache utilization. The improved version of the scheduler is called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">multi-level ADWS</i> . The idea is that only part of a computation whose working set size is small enough to fit into a shared cache is scheduled by ADWS within the cache recursively, thus avoiding excessive capacity misses. Our evaluation on a benchmark of parallel decision tree construction demonstrated that multi-level ADWS outperformed the conventional random work stealing of Cilk Plus by 61% and it showed a 40% performance improvement over the previous ADWS design.

CALLIOPE: Pseudospectral Shearing Magnetohydrodynamics Code with a Pencil Decomposition

Abstract The pseudospectral method is a highly accurate numerical scheme suitable for turbulence simulations. We have developed an open-source pseudospectral code, calliope, which adopts the P3DFFT library to perform a fast Fourier transform with the two-dimensional (pencil) decomposition of numerical grids. calliope can solve incompressible magnetohydrodynamics (MHD), isothermal compressible MHD, and rotational reduced MHD with parallel computation using very large numbers of cores (&gt;105 cores for 20483 grids). The code can also solve for local magnetorotational turbulence in a shearing frame using the remapping method. calliope is currently the only pseudospectral code that can compute magnetorotational turbulence using pencil-domain decomposition. This paper presents the numerical scheme of calliope and the results of linear and nonlinear numerical tests, including compressible local magnetorotational turbulence with the largest grid number reported to date.

An efficient multi-level cache system for geometrically interconnected many-core chip multiprocessor

&lt;p&gt;&lt;span lang="EN-US"&gt;Many-core chip multiprocessor offers high parallel processing power for big data analytics; however, they require efficient multi-level cache and interconnection to achieve high system throughput. Using on-chip first level L1 and second level L2 per core fast private caches is expensive for large number of cores. In this paper, for moderate number of cores from 16 to 64, we present a cost and performance efficient multi-level cache system with per core L1 and last level shared bus cache on each bus line of a cost-efficient geometrically bus-based interconnection. In our approach, we extracted cache hit and miss concurrencies and applied concurrent average memory access time to more accurately determine the cache system performance. We conducted least recently used cache policy-based simulation for cache system with L1, with L1/L2, and with L1/shared bus cache. Our simulation results show that an average system throughput improvement of 2.5x can be achieved by using system with L1/shared bus cache system compared to using only first level L1 or L1/L2. Further, we show that the throughput degradation for the proposed cache system is only within 5% for a single bus fault, suggesting a good bus fault tolerance.&lt;/span&gt;&lt;/p&gt;

Fine Logging Evaluation of Fractures in Continental Tight Oil Sandstone Reservoirs of Yanchang Formation in the Ordos Basin

Fractures are the main seepage channels in tight reservoirs, and they affect the distribution of high-quality reservoirs and the enrichment of hydrocarbons. Fine logging identification of fractures in the strongly heterogeneous continental tight oil reservoirs of the Upper Triassic Yanchang Formation in the Ordos Basin is a hot and difficult point in the field of petroleum geology. In this paper, taking the Yanchang Formation as an example, a fractal model was constructed to identify fractures in tight oil reservoirs using a large number of cores, conventional and micro-resistivity imaging logging data. The results show that high-angle, vertical and bedding fractures are mainly developed in the Yanchang Formation tight sandstones. There is a negative correlation between sand body thickness and fracture development degree. The fracture sensitivity parameters were used to construct a coupled fractal fracture index. The fractal model incorporates logging information from natural gamma, acoustic wave time difference, rock density, and shallow lateral resistivity. In addition, the constructed fracture fractal index realizes the functions of multi-conventional logging information fusion, which can effectively identify fracture development segments in sandstone. According to the statistics, the fracture identification rate is 83.3%. The study also found that with the increase of sandstone brittleness index, the fracture index has a “S” shape increasing trend. Therefore, the content of brittle mineral components in sandstone is an important factor affecting the development degree of natural fractures, and fractures are more likely to occur in high brittle, thin sand bodies. The highly brittle framework minerals have strong stress-supporting capacity, which can keep fractures open by resisting high overlying loads.

NxTF: An API and Compiler for Deep Spiking Neural Networks on Intel Loihi

Spiking Neural Networks (SNNs) is a promising paradigm for efficient event-driven processing of spatio-temporally sparse data streams. Spiking Neural Networks (SNNs) have inspired the design of and can take advantage of the emerging class of neuromorphic processors like Intel Loihi. These novel hardware architectures expose a variety of constraints that affect firmware, compiler, and algorithm development alike. To enable rapid and flexible development of SNN algorithms on Loihi, we developed NxTF: a programming interface derived from Keras and compiler optimized for mapping deep convolutional SNNs to the multi-core Intel Loihi architecture. We evaluate NxTF on Deep Neural Networks (DNNs) trained directly on spikes as well as models converted from traditional DNNs, processing both sparse event-based and dense frame-based datasets. Further, we assess the effectiveness of the compiler to distribute models across a large number of cores and to compress models by exploiting Loihi’s weight-sharing features. Finally, we evaluate model accuracy, energy, and time-to-solution compared to other architectures. The compiler achieves near-optimal resource utilization of 80% across 16 Loihi chips for a 28-layer, 4M parameter MobileNet model with input size 128×128. In addition, we report the lowest error rate of 8.52% for the CIFAR-10 dataset on neuromorphic hardware, using an off-the-shelf MobileNet.

Development Characteristics and Finite Element Simulation of Fractures in Tight Oil Sandstone Reservoirs of Yanchang Formation in Western Ordos Basin

Structural fractures have a significant control effect on the large-scale accumulation of hydrocarbons in the Yanchang Formation. Previous studies have affirmed the important role of fractures in hydrocarbon accumulations in strongly deformed zones. However, for low-amplitude structural areas, the degree of fracture development is relatively low, and their control on sweet spots of hydrocarbons has not yet formed a unified understanding. In this paper, taking the Upper Triassic Yanchang Formation in the western Ordos Basin as an example, the development characteristics, prediction method, and the distribution of fractures in tight sandstone reservoirs in low-amplitude structural areas have been systematically studied using a large number of cores, thin sections, paleomagnetism, FMI logging, acoustic emission, productivity data, and finite element method. The research results showed that the Yanchang Formation in the study area mainly develop high-angle and vertical fractures, which were formed by regional tectonic shearing. Fractures are mainly developed in the fine-grained and ultra-fine-grained sandstones of the distributary channel and estuary bar microfacies, while the fractures in the medium-grained sandstones of the distributary channel and the mudstones of the distributary bay are relatively underdeveloped. The core fractures and micro-fractures of the Yanchang Formation all have the regional distribution characteristics, and the fracture strikes are mainly between NE50° and NE 70°. Moreover, the finite element method was used to predict the fractures in the target layer, and the prediction results are consistent with the actual distribution results of the fractures. The coupling analysis of fractures and tight oil sandstone distribution showed that the existence of fractures provided conditions for the accumulation of hydrocarbons in the Yanchang Formation. The confluence and turning areas of the river channels were repeatedly scoured by river water, and the rocks were brittle and easy to form fractures. The thickness of the fractured sandstone in these areas is usually greater than 0.4 m. Moderately developed fracture zones are prone to form hydrocarbon accumulation “sweet spots,” and the fracture indexes of these areas are usually distributed between 0.8 and 1.2. However, when the fracture index exceeds 1.2, over-developed fractures are unfavorable for the accumulation of hydrocarbons.

Lenslet array-free efficient coherent combining of broadband pulses at the output of a multicore fiber with a square core grid.

An efficient optical scheme for coherent combining of radiation from the output of a multicore fiber (MCF) with a square array of cores in the out-of-phase supermode is proposed. The scheme uses only simple optical elements and is suitable for an arbitrary number of MCF cores. In a proof-of-concept experiment broadband pulses transmitted through a 25-core fiber were combined with 81% efficiency and good beam quality. In numerical modeling a close to unity efficiency is obtained for a large number of cores. The proposed scheme can be used in a reverse direction for efficient beam splitting and launching the out-of-phase supermode into the MCF.

Influence of mid-term base level cycle changes on the development characteristics of shallow water delta sand bodies: A case study of the Fuyu oil layer in the Mu 17 block in the southern Songliao Basin

The Fuyu oil layer in the southern Songliao Basin is a shallow-water delta sedimentary system. Frequent changes in the base level lead to obvious differences in the superimposed style, scale and geometry of the sand bodies, which seriously affects the distribution of remaining oil. Taking the Fuyu oil layer in the Mu 17 block in the southern Songliao Basin as an example, using a large number of cores and logging data in the dense well pattern development area, based on the relationship between the accommodated space (A) and the sediment supply (S), it is clear that The Fuyu oil layer develops the superimposed styles of cut-stack, superimposed and isolated sand bodies. In the depositional period of MSC1 and MSC2, the superimposed style of the cut-and-stacked sand body was the main one; in the depositional period of MSC3, the superimposed style of the superimposed sand body was dominant; in the depositional period of MSC4, the sand body was the least developed, and the superimposed style of the isolated sand body was the main one.This study has important guiding significance for clarifying the influence of medium-term base level cycle changes on the superposition law of shallow water delta sand bodies.

Data‐aware and simulation‐driven planning of scientific workflows on IaaS clouds

AbstractThe promise of an easy access to a virtually unlimited number of resources makes Infrastructure as a Service Clouds a good candidate for the execution of data‐intensive workflow applications composed of hundreds of computational tasks. Thanks to a careful execution planning, workflow management systems can build a tailored compute infrastructure by combining a set of virtual machine instances. However, these applications usually rely on files to handle dependencies between tasks. A storage space shared by all virtual machines may become a bottleneck and badly impact the application execution time. In this article, we propose an original data‐aware planning algorithm that leverages two characteristics of a family of virtual machines instances, that is, a large number of cores and a dedicated storage space on fast SSD drives, to improve data locality, hence reducing the amount of data transfers over the network during the execution of a workflow. We also propose a simulation‐driven approach to solve a cost‐performance optimization problem and correctly dimension the virtual infrastructure onto which execute a given workflow. Experiments conducted with real application workflows show the benefits of the presented algorithms. The data‐aware planning leads to a clear reduction of both execution time and volume of data transferred over the network while the simulation‐driven approach allows us to dimension the infrastructure in a reasonable time.

3D NoC Low-Power Mapping Optimization Based on Improved Genetic Algorithm.

Power optimization is an important part of network-on-chip(NoC) design. This paper proposes an improved algorithm based on genetic algorithm on how to properly map IP (Intellectual Property) cores to 3D NoC. First, in view of the randomness of the traditional genetic algorithm in individual selection, an improved greedy algorithm is used in the initial population generation stage to make the generated individuals reach the optimal. Secondly, in view of the weak local optimization ability of the traditional genetic algorithm and prone to premature problems, the simulated annealing algorithm is added in the crossover operation stage to make the offspring reach the global optimum. The experimental results show that compared with the traditional genetic algorithm, the algorithm has better convergence and low power consumption performance, which can quickly search for a better solution, in the case of a large number of cores (124 IP cores), the average power consumption can be reduced by 42.2%.

Resolution improvement in real-time and video mosaicing for fiber bundle imaging

A fiber bundle allows easy access to a wide range of human tissue. It has smaller diameters, which limits its effective field of view (FOV), and consists of a large number of cores surrounded by a cladding layer, which reduces its spatial resolution. In this paper, we develop an algorithm that processes successively captured raw fiber bundle images in an online fashion. Our algorithm tackles the tasks of super-resolution (SR) and video mosaicing jointly. The natural movement of the fiber tip in successive frames produces offsets that are random in the pixel domain to apply multi-frame SR imaging. Meanwhile, the associated FOV can be extended by mosaicing reconstructed SR images with obtained shifted information. Our approach has low computational complexity that allows for processing in real-time. The performance of resolution improvement in real-time and video mosaicing is demonstrated on the resolution target and biological samples.

Discovery of Bozhong 13-2 Archean large monoblock volatile buried hill oilfield and its oil and gas exploration significance

The hydrocarbon accumulation pattern of the Mesozoic covered Archean buried hill in the Bohai Oilfield was not clarified before, which restricts the oil and gas exploration progress in this area. In recent years, reservoir development modes and reservoir prediction technologies have been researched after the structural evolution and stratigraphic distribution characteristics of the buried hill in the Bozhong sag were investigated based on new 3D seismic data and a large number of cores and logging data of newly drilling wells. Then, the hydrocarbon accumulation pattern of covered type Archean buried hill was analyzed. Finally, this pattern was applied to guide the oil and gas exploration deployment in Bozhong 13-2 structure, and a significant discovery of Bozhong 13-2 large monoblock volatile buried hill oilfield was realized. And the following research results were obtained. First, the structure in this area mainly experienced multi-period tectonic movements, such as Indosinian, Yanshanian and Himalayan, and it is characterized by compression and thrusting to form a hill during Indosinian-early Yanshanian, extension and inversion transformation during middle Yanshanian, and burial and finalization during Himalayan, forming Bozhong 13-2 Mesozoic covered Archean large buried hill trap. Second, under the effect of Mesozoic cover and the control of multi-stage stress superposition, the reservoir development in this area has the distribution characteristics of “different vertical top, internal lateral continuity”. Third, the fracture prediction technology based on smooth reflection strength filter is developed. And by means of this technology, multi-scale fractured reservoirs inside buried hills can be predicted effectively. Fourth, the hydrocarbon accumulation pattern of “overpressure injection-relay migration” for covered type buried hills is established, which provides a basis for long-distance migration and large-scale accumulation of oil and gas in covered type buried hills. In conclusion, these understandings guide the discovery of Bozhong 13-2 Oilfield, achieve a breakthrough in the oil and exploration of Mesozoic covered buried hill in the Bohai Oilfield and can be used as the reference for the oil and gas exploration in the similar covered type buried hills of the Bohai Bay Basin and other areas.