The improved computation presented in this paper is aimed to optimize the neural networks learning process using Levenberg-Marquardt (LM) algorithm. Quasi-Hessian matrix and gradient vector are computed directly, without Jacobian multiplication storage. memory limitation problem for LM training solved. Considering symmetry of quasi-Hessian matrix, only elements its upper/lower triangular array need be calculated. Therefore, speed significantly, not because smaller stored memory, but also...

Radial basis function (RBF) networks have advantages of easy design, good generalization, strong tolerance to input noise, and online learning ability. The properties RBF make it very suitable design flexible control systems. This paper presents a review on different approaches designing training networks. recently developed algorithm is introduced for compact performing efficient process. At last, several problems are applied test the main networks, including their generalization ability,...

The existing ISFET-based DNA sequencing detects hydrogen ions released during the polymerization of strands on microbeads, which are scattered into microwell array above ISFET sensor with unknown distribution. However, false pH detection happens at empty microwells due to crosstalk from neighboring microbeads. In this paper, a dual-mode CMOS is proposed have accurate toward sequencing.Dual-mode sensing, optical and chemical modes, realized by integrating image (CIS) sensor, fabricated in...

<b>R</b> is a mature open-source programming language for statistical computing and graphics. Many areas of research are experiencing rapid growth in the size data sets. Methodological advances drive increased use simulations. A common approach to parallel computing. This paper presents an overview techniques with on computer clusters, multi-core systems, grid It reviews sixteen different packages, comparing them their state development, technology used, as well usability, acceptance,...

Design of hybrid circuits and systems based on CMOS nano-device requires rethinking fundamental circuit analysis to aid design exploration. Conventional with modified nodal (MNA) cannot consider new nano-devices such as memristor together the traditional devices. This paper has introduced a MNA method magnetic flux (Φ) state variable. New SPICE-like simulator is thereby developed for circuits. A number memristor-based designs are explored, oscillator, chaotic circuit, programmable logic,...

The data-oriented applications have introduced increased demands on memory capacity and bandwidth, which raises the need to rethink architecture of current computing platforms. logic-in-memory is highly promising as future logic-memory integration paradigm for high throughput data-driven applications. From technology aspect, one recently nonvolatile device, domain-wall nanowire (or race-track) not only shows potential power efficient memory, but also by its unique physics spintronics. This...

In this paper, we propose to adapt parallelizing transformations, more specifically, reduction parallelizations, the actual reference pattern executed by a loop, i.e., particular input data and dynamic phase of program. More precisely will show how, after validating at run-time (when is not possible compile time) can dynamically characterize its choose most appropriate method for it. For purpose, develop library parallel algorithms, including both previously known novel schemes, which...

There is great attention to develop hardware accelerator with better energy efficiency, as well throughput, than GPUs for convolutional neural network (CNN). The existing solutions have relatively limited parallelism large power consumption (including leakage power). In this paper, we present a resistive random access memory (ReRAM)-accelerated CNN that can achieve significantly higher throughput and efficiency when the trained binary constraints on both weights activations, further mapped...

Lensless microfluidic imaging with super-resolution processing has become a promising solution to miniaturize the conventional flow cytometer for point-of-care applications. The previous multi-frame system can improve resolution but limited cell rate and hence low throughput when capturing multiple subpixel-shifted images. This paper introduces single-frame on-line machine-learning contact images of cells. A corresponding contact-imaging based prototype is demonstrated recognition counting....

Mapping virtual processes onto physical processos is one of the most important issues in parallel computing. The problem mapping processes/tasks processors equivalent to graph embedding which has been studied extensively. Although many techniques have proposed for embeddings two-dimensional grids, hypercubes, etc., there are few efforts on three-dimensional grids and tori. Motivated better support task Blue Gene/L supercomputer, this paper, we present integration topology library that based...

NAMD (nanoscale molecular dynamics) is a production dynamics (MD) application for biomolecular simulations that include assemblages of proteins, cell membranes, and water molecules. In simulation, the problem size fixed large number iterations must be executed in order to understand interesting biological phenomena. Hence, we need MD applications scale thousands processors, even though individual timestep on one processor quite small. has demonstrated its performance several parallel...

Hybrid integration of CMOS and nonvolatile memory (NVM) devices has become the foundation for emerging memory-based computing. The primary challenge to validate hybrid system with both NVM is develop a SPICE-like simulator that can simulate dynamic behavior accurately efficiently. Since memristor, spin-transfer-toque magnetic-tunneling-junction (STT-MTJ) phase-change-memory (PCM) are most promising candidates next generation devices, it under great interest in including these new standard...

<?Pub Dtl=""?> To provide wide frequency tuning range (FTR) with compact implementation area, a new inductive method is introduced in this paper for CMOS 60 GHz voltage controlled oscillator (VCO). The based on switching inductor-loaded transformer by configuring different current return-paths the secondary coil of transformer. Different from previous methods, proposed VCO topology can achieve FTR multiple sub-bands at within area only one Two VCOs are demonstrated 65 nm design targets...

A 239–281 GHz sub-THz imager with circular- polarize d substrate-integrated-waveguide (SIW) antenna is demonstrated in CMOS 65 nm process high spectrum resolution and sensitivity for spectroscopy imaging system consideration of depolarization effects. The a receiver that consists circular polarized substrate integrated waveguide antenna, down-conversion mixer power gain amplifier (PGA). proposed measured <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML"...

A lensless blood cell counting system integrating microfluidic channel and a complementary metal oxide semiconductor (CMOS) image sensor is promising technique to miniaturize the conventional optical lens based imaging for point-of-care testing (POCT). However, such has limited resolution, making it imperative improve resolution from system-level using super-resolution (SR) processing. Yet, how towards better detection recognition with low cost of processing resources without degrading...

Traditional TEM-based I/O communication through either PCB traces or free space has significant path loss and electromagnetic (EM) interferences that are hardly employed toward low-power dense I/Os for data server. We propose a 0.1-1 THz sub-terahertz (THz) interface using surface plasmonic polariton (SPP) interconnects in CMOS to obtain high-energy efficiency low-crosstalk interconnect. The wave can be established with on-chip generation further deployed signal propagation by corrugating...

Big-data storage poses significant challenges to anonymization of sensitive information against data sniffing. Not only will the encryption bandwidth be limited by I/O traffic, transfer between processor and memory also expose input-output mapping intermediate computations on channels that are susceptible semi-invasive non-invasive attacks. Limited simplistic cell-level logic, existing logic-in-memory computing architectures incapable performing complete process within at reasonable...

Current parallelizing compilers cannot identify a significant fraction of parallelizable loops because they have complex or statically insufficiently defined access patterns. In our previously proposed framework we speculatively executed loop as doall, and applied fully parallel data dependence test to determine if it had any cross-processor dependences; the failed, then was re-executed serially. While this method exploits doall parallelism well, can cause slowdowns for with even one flow...

Based on a newly introduced zero phase shifter (ZPS), 2-D distributed power-combining network is developed in this paper to provide simultaneous amplification and power combining. One type of metamaterial, called the composite right/left-handed transmission line, deployed design ZPS with detailed considerations achieve low loss wideband performance. The proposed combining implemented for one 60-GHz amplifier (PA) by UMC standard 65-nm CMOS process. Measured results show that fabricated PA...

A robust physical design of 3-D IC requires investigation on through-silicon via (TSV). The large temperatures and stress gradients can severely affect TSV delay with variation. traditional model treats as a resistor linear electrical-thermal dependence, which ignores the fundamental device physics. In this paper, physics-based electrical-thermal–mechanical is developed for signal TSVs in IC. With consideration liner material also stress, nonlinear established between electrical temperature...

In this paper, low phase-noise, low-power, and compact oscillators are demonstrated at the millimeter-wave region based on differential transmission lines (DTLs) loaded with metamaterial resonators. There two types of resonators explored: split-ring (SRRs) complementary (CSRRs). By creating a sharp stopband resonance frequency from SRR or CSRR, backward electrical-magnetic (EM) wave is reflected to couple forward EM form standing in DTL host, which results high- <i...