Search

Sparse matrix-vector multiplication (SpMV) plays a critical role in a wide range of linear algebra computations, particularly in scientific and engineering disciplines. However, the irregular memory access patterns, extensive memory usage, high bandwidth requirements, and underutilization of parallelism hinder the computational efficiency of SpMV on GPUs. In this paper, we propose a novel approach called block-wise dynamic mixed-precision (BDMP) to address these challenges. Our methodology involves partitioning the original matrix into uniformly sized blocks, with each block's size determined by considering architectural characteristics and accuracy requirements. Additionally, we dynamically assign precision to each block using a precision selection method that takes into account the value distribution of the original sparse matrix. We develop two distinct SpMV computation algorithms for BDMP: BDMP-PBP (Precision-based partitioning) and BDMP-TCKI (Tailored compression and kernel implementation). BDMP-PBP partitions the matrix into two independent matrices for separate computations based on block precision, offering flexibility for integration with other optimization techniques. Meanwhile, BDMP-TCKI focuses on achieving significant thread-level parallelism and memory utilization by tailoring an appropriate compressed storage format and kernel implementation for each block. We compare BDMP with NVIDIA's cuSPARSE library and three state-of-the-art SpMV methods, including SELLP, MergeBase, and BalanceCSR, using matrices from the University of Florida's SuiteSparse dataset collection. BDMP-PBP and BDMP-TCKI show average speedups up to 2.64 × and 2.91 × on Turing RTX 2080Ti, and up to 2.99 × and 3.22 × on Ampere A100. The results demonstrate that BDMP enables the optimization of computation speed without compromising the precision necessary for reliable results. [ABSTRACT FROM AUTHOR]

In recent years, mobile medical services have gradually become the common focus of the communications and healthcare industries. People can get medical information and services efficiently and conveniently anytime, anywhere. However, due to the limitations of the mobile terminal's own computing and storage capabilities, mobile therapy is greatly challenged. Therefore, we propose a mobile medical health system based on cloud computing. Firstly, principal component analysis was used to obtain representative features. Then, a simplified feature subset was applied to support vector machine (SVM) based on Sigmoid kernel function. The data set was categorized by SVM as cancer patient and normal object experimental results show that the method is improved in accuracy, sensitivity, and specificity. [ABSTRACT FROM AUTHOR]

With the rapid development of the mobile Internet and the Internet of Things, the number of personal wireless devices and wireless services have grown exponentially resulting in the increase in massive data which requires massive spectrum to transmission. To improve the utilization of scarce spectrum resource, cognitive radio (CR) is considered as the most promising technology. Spectrum sensing is the premise and key technology in CR. However, current spectrum sensing strategies have some issues to sense multiple heterogeneous channels in heterogeneous network, such as neglect of the different performances of each secondary user on heterogeneous channels and sensing task assignment without consideration of the selection of sensing time. To address these issues, two parallel cooperative spectrum sensing strategies based on heuristic method are proposed, which are heuristic method with sensing time exhaustion and heuristic method with sensing time heuristic. These strategies consider both the sensing task assignment and sensing time selection to maximize the available throughput under the restrictions of detection probability and false alarm probability. Monte Carlo simulations show that the proposed strategies can obtain higher available throughputs than iterative Hungarian-based strategy, improved iterative Hungarian-based strategy and greedy-based strategy. [ABSTRACT FROM AUTHOR]

To address the issue of spectrum scarcity in future communication for smart world, cognitive radio (CR) has emerged as a promising solution. In CR, spectrum sensing is the premise and key technology. To defense against the PUE attack in spectrum sensing, two spectrum sensing methods against PUE attack based on fractal dimension are proposed, which are spectrum sensing defending against PUE attack based on normalized Sevcik fractal dimension in frequency domain (SSMS) and spectrum sensing defending against PUE attack based on normalized Sevcik fractal dimension in frequency domain and normalized Petrosian fractal dimension (SSMSP). They identify the modulation type of the received signal by SVM classifier according to the fractal dimension of the received signal to detect PUE attack. The simulation results show that the two proposed methods can effectively identify the modulation type and detect the PUE attack. When SNR is larger than 10 dB, the PUE detection probabilities of both SSMS and SSMSP can reach 100%. Even when SNR is - 20 dB, the PUE detection probabilities are still larger than 97%. Therefore, SSMS and SSMSP can effectively defend against PUE attack. [ABSTRACT FROM AUTHOR]

To address the problem of spectrum scarcity in future communication for smart world, cognitive radio (CR) is viewed as an effective way and has been widely studied in recent years. Spectrum sensing is the key to deploy Cognitive radio (CR) network. Parallel cooperative spectrum sensing provides an effective solution in multi-channel scenario. In this paper, joint optimal sensing duration selection and sensing task allocation is studied in parallel cooperative spectrum sensing for heterogeneous multi-channel CR systems. Two cross-layer parallel cooperative spectrum sensing methods based on iterative KM algorithm are proposed. Sensing duration selection and sensing task allocation are designed jointly to maximize the available throughput. A two-step method is implemented. Firstly it determines the optimal sensing task allocation for fixed sensing duration by iterative KM algorithm, and then selects the optimal sensing duration by exhaustion method and quartering method, respectively. The simulation results show that the proposed method can optimally select the sensing duration and obtain higher available throughput than other compared methods. [ABSTRACT FROM AUTHOR]

The recently explored Tianzuo hydrothermal field in serpentinized ultramafic rocks of the amagmatic segment of the ultraslow-spreading Southwest Indian Ridge displays high-temperature sulfide mineralization (isocubanite, sphalerite, and minor pyrrhotite) and low-temperature (pyrite and covellite) phases. Pyrite can be subdivided into pyrite-I and -II, with the former generally having a pseudomorphic texture after pyrrhotite and the latter typically growing around isocubanite, sphalerite, and pyrite-I or occurring as individual grains in quartz veinlets. The sulfide minerals have the greatest range of δ34S values (− 23.8 to 14.1‰), found so far among modern sediment-starved ridges, with distinct δ34S values for low- and high-temperature mineral phases. The high δ34S values of isocubanite (9.6 to 12.2‰) and sphalerite (9.1 to 14.1‰) suggest that sulfate, which precipitated from seawater during an early low-temperature phase of hydrothermal circulation, was the main sulfur source for these sulfides. Pyrite-II has the lowest and most variable δ34S values (− 23.8 to − 3.6‰), suggesting microbial sulfate reduction. Pyrite-I has variable and generally positive δ34S values (− 0.1 to 12.0‰), with sulfur being inherited from pyrrhotite from the original thermochemical reduction of sulfate, mixed with volcanogenic sulfur. Intermittent magmatism represented by gabbroic intrusions, and high permeability caused by well-developed fractures associated with detachment faults, contributed to the formation of sulfides in the Tianzuo hydrothermal field. These factors possibly control sulfide mineralization in amagmatic segments of ultraslow-spreading ridges. [ABSTRACT FROM AUTHOR]

To give full play to the high efficiency and parallelism of multi-processor systems, the fireworks algorithm (FWA) is improved, and a multi-processor task scheduling algorithm based on improved FWA, named IMFWA, is proposed. IMFWA maps continuous space to discrete space by designing the fireworks location coding method, improves the Gaussian mutation process, and sets adaptive dimensions to accelerate the convergence speed of the algorithm. At the same time, in order to reduce the time complexity of the algorithm and shorten the time finding the optimal task scheduling sequence, the fitness-based tournament selection strategy is used instead of the rule based on Euclidean distance. Finally, IMFWA is compared with the basic fireworks algorithm and the genetic algorithms on the Matlab platform for performance analysis. The results show that the IMFWA has advantages in the convergence speed, and the negative impact of the number of tasks is also lower than the fireworks algorithm and genetic algorithm. [ABSTRACT FROM AUTHOR]

Seafloor hydrothermal polymetallic sulfide deposits are a new type of resource, with great potential economic value and good prospect development. This paper discusses turbidity, oxidation-reduction potential, and temperature anomalies of hydrothermal plumes from the Zouyu-1 and Zouyu-2 hydrothermal fields on the southern Mid-Atlantic Ridge. We use the known location of these vent fields and plume data collected in multiple years (2009, 2011, 2013) to demonstrate how real-time plume exploration can be used to locate active vent fields, and thus associated sulfide deposits. Turbidity anomalies can be detected 10 s of km from an active source, but the location precision is no better than a few kilometers because fine-grained particles are quasi-conservative over periods of many days. Temperature and oxidation-reduction potential anomalies provide location precision of a few hundred meters. Temperature anomalies are generally weak and difficult to reliably detect, except by chance encounters of a buoyant plume. Oxidation-reduction potential is highly sensitive (nmol concentrations of reduced hydrothermal chemicals) to discharges of all temperatures and responds immediately to a plume encounter. Real-time surveys using continuous tows of turbidity and oxidation-reduction potential sensors offer the most efficient and precise surface ship exploration presently possible. [ABSTRACT FROM AUTHOR]

Polymetallic sulfides have attracted more attentions, which are regarded as one of the potential seabed mineral resources for their high grade of precious metal elements such as Cu, Zn, Pb, Au, and Ag. Since 2007, there are four China dayang cruises (CDCs) with eight legs totally, which have been carried out to investigate polymetallic sulfides on the southwest Indian ridge (SWIR), and eight hydrothermal vents have been found. In 2011, the China Ocean Mineral Resources Research and Development Association (COMRA) and International Seabed Authority (ISA) signed a 15-year contract for exploration for polymetallic sulfides located on the SWIR between 46°S and 53°S. The area allocated to the COMRA is approximately 10,000 km. In this paper, we review the recent research advances about hydrothermal activity of the SWIR by the CDCs. As for the sulfides exploration in the future, some aspects should be emphasized, such as the controlling factors of sulfide mineralization on the SWIR, prospecting methods of inactive and buried sulfide deposits, sulfides resources assessment method, and the near-bottom sulfides exploration technology systems. [ABSTRACT FROM AUTHOR]