Core Switching Research Articles

Energy Optimization for Large-Scale 3D Manycores in the Dark-Silicon Era

In this paper, we study the impact of the idle/dynamic power consumption ratio on the effectiveness of a multi-Vdd/frequency manycore design. We propose a new tool called LVSiM (a Low-Power and Variation-Aware Manycore Simulator) to carry out the experiments. It is a novel manycore simulator targeted towards low-power optimization methods including within-die process and workload variations. LVSiM provides a holistic platform for multi-Vdd/frequency voltage island analysis, optimization, and design. It provides a tool for the early design exploration stage to analyze large-scale manycores with a given number of cores on 3D-stacked layers, network-on-chip communication busses, technology parameters, voltage and frequency values, and power grid parameters, using a variety of different optimization methods. LVSiM has been calibrated with Sniper/McPAT at a nominal frequency, and then the energy-delay-product (EDP) numbers were compared after frequency scaling. The average error is shown to be 10% after frequency scaling, which is sufficient for our purposes. The experiments in this work are carried out for different Idle/Dynamic ratios considering 1260 benchmarks with task sizes ranging from 4000 to 16 000 executing on 3200 cores. The best configurations are shown to produce on average 20.7% to 24.6% EDP savings compared to the nominal configuration. Traditional scheduling methods are used in the nominal configuration with the unused cores switched off. In addition, we show that, as the Idle/Dynamic ratio increases, the multi-Vdd/frequency approach becomes less effective. In the case of a high Idle/Dynamic ratio, the minimum EDP can be achieved through switching off unused cores as opposed to using a multi-Vdd/frequency approach. This conclusion is important, especially in the dark-silicon era, where switching cores on and/or off as needed is a common practice.

Edge-soliton-meditated fast vortex core dynamic switching in a notched magnetic disk

Magnetic vortices are characterized by the sense of in-plane magnetization chirality and by the polarity of the vortex core. Searching for efficient ways to control the vortex polarity and chirality is of highly fundamental and practical significance towards high speed and high density non-volatile memory applications. Here we report a novel vortex core (VC) switching process in a notched disk, driven by a pulse field. Micromagnetic simulations demonstrate that the dynamics of edge solitons around the notch play a critical role in the switching process, that leads to a sub-nanosecond VC fast switching. The switching diagram as a function of the field strength and duration was presented to indicate the operational range. Further studies on the VC switching under an out-of-plane bias field demonstrate that such switching is deterministic and robust against magnetic environment. Finally, by applying a resonant alternative field, a periodic VC switching is observed with much reduced switching field. The presented results indicate the potential of notched disks for VC based memory applications.

Micromagnetics of FePt-L10 Media With Thermally Insulating Magnetic Grain Boundaries

In heat assisted magnetic recording (HAMR), grain-to-grain Curie temperature variation presents a practical limit to grain size reduction hence, the scaling of recording area density capability (ADC), for the advancement of granular FePt-L1 0 media [1] [2]. At small grain sizes, a reduction of grain size yields an increase of surface-to-volume ratio, thereby, a reduction of Curie temperature [3]. The resulting grain size dependence of Curie temperature gives rise to the cause of the Curie temperature variation due to unavoidable grain size distributions with today’s film fabrication technique. Aiming at this grand challenge, a recent experimental study has shown a potentially practical solution by successfully fabricating granular FePt-L1 0 media with thermally insulating magnetic grain boundaries [4]. In this paper, we present a systematic modeling study on the micromagnetic recording characteristics of such media. In the study, both the FePt-L1 0 grain core and a thin layer of ferromagnetic grain boundaries are meshed and modeled using dynamic Landau-Lifshitz-Bloch (LLB) equations. The room temperature anisotropy field of the FePt core is assumed to be $H_{k}=80$ kOe and zero magnetocrystalline anisotropy is assumed for the soft magnetic grain boundary material. Gaussian geometric thermal profile is assumed along with realistic media motion are used to simulate practical recording process. Figure 1(a) shows the $M_{s}(\mathrm {T})$ curves for both the FePt grain core and the insulating magnetic grain boundary material assumed. Note that the assumed magnetic grain boundary has a Curie temperature significantly higher than that of the FePt core. For all the results presented in this digest, the diameter of the FePt core is assumed to be 6nm with 1nm thick magnetic grain boundary. The modeling simulation shows that the ferromagnetic grain boundary provide significant assist in the magnetization process of the FePt grain core during recording. With the grain boundary coupled to the FePt core, the magnetization process of the core becomes significantly more deterministic. Figure 1(b) shows the switching probability of the FePt grain core as a function of recording field amplitude with the medium moving at a speed of 15 m/s and a spatial thermal gradient of 8 K/nm. The case “Not Coupled” indicating no exchange coupling between the grain boundary and the FePt core is plotted for comparison. With the magnetic grain boundary exchange coupled to the core, full saturation recording can be reached at lower field values in comparison with the case of non-magnetic grain boundaries. The effect is even more enhanced for smaller size of FePt grain cores, which will be presented in the full paper. Please note the “Not Coupled” case is essentially the same as the case with non-magnetic grain boundaries. Figures 2(a) and 2(b) show the writing of a transition when the head field direction is reversed (black dashed lines). The case with the exchange-coupled grain boundary and FePt core shows more rapid and better defined magnetization switching with the magnetizations of the core and grain boundary switching together. Whereas for the case that the core and grain boundary are not coupled, the switching of the FePt core is much more gradual with noticeably fluctuations. Also note that the grain boundary actually magnetizes a little faster than the FePt core does. In summary, the magnetic grain boundary exchange coupled to the FePt core provides significant assistance to the magnetization process of the FePt core during recording, both in heat assisted magnetization process and in transition magnetization process with field reversal. Therefore, it is important for the Curie temperature of the grain boundary material to be higher than that of the FePt core for the obvious reasons. The level of grain boundary magnetization at the recording temperature is important and the higher the magnetization level, the more enhanced assisting effect will be during recording. It is also important to note that the exchange coupling between the grain boundaries of adjacent grains needs to be minimized in practice. The experimental study has shown that this objective could be achieved in practice [4]. In conclusion, granular FePt media with a thin layer of soft magnetic grain boundaries that are exchange coupled to the FePt cores not only can suppress or eliminate grain-to-grain Curie temperature variation, but also can assist the magnetization process, especially for small size grains, during recording in HAMR. The more deterministic magnetization processes helps to suppress thermal-induced transition jitter during recording process, consequently enhancing the scalability of both the linear recording density capability and the overall ADC.

High Step-Up Boost-Fly-Back Converter with Soft Switching for Photovoltaic Applications

In this paper, a novel high step-up single switch DC–DC converter with soft switching is presented. The main application of this converter is the connection of photovoltaic (PV) system to a 400[Formula: see text]V DC-bus. The proposed converter achieves high step-up voltage gain with small duty cycle by a combined boost and fly-back topology. Also, its switch voltage stress is lower than the output voltage. Besides, in the proposed converter, any auxiliary switch or magnetic core has not been used — therefore, the number of converter components has not been increased much in comparison with the conventional boost-fly-back converter. The operation principles of the converter and its theoretical operation waveforms are presented. In order to evaluate the theoretical analysis, a prototype of the converter is designed and experimentally implemented. The practical results are presented for a 100[Formula: see text]W boost-fly-back converter with input voltage of 40[Formula: see text]V and output voltage of 400[Formula: see text]V. Also, the output capacitor is designed to have less than 1% ripple on output voltage.

Computer Network Simulation of Firewall and VoIP Performance Monitoring

Fast growing in communication technology has influenced global changes and challenges appear in the field of network security issues. Security solutions must be efficient and operate in a way to deal with threats, reject and stop the network intruders and Trojans. The simulated network of Salahaddin university new campus is planned to build on an area of (3000X3000) meter square. The network consists of many main and secondary devices. The network is mainly consisted of one core switch that provides a very high data transfer through connecting all the collected positions by a variety of cable medias to the entire network switches which are installed in each college location. The department of academic administration (presidency) of the university design is similar to the network designed in each college. The mentioned switch obtains the services from a router that isolates the network from the cloud which supports the services of internet to the network. The firewall is connected to the switch that connects the main server and cloud together. This paper focuses on undertaking a simulation to analyze and examine the performance of the whole network when two scenarios are implemented, the first is if firewall devices is used, the second is when firewall is not used, since the project of building Salahaddin University new campus is at the initial stage, therefore, the researchers think that, it is very important to figure out the drawbacks and deadlocks of using firewall upon each branch of the network and overall network performance before the submitting the final networks design that going to be implemented and installed, because this will indicate many differences on the construction, for example, the network panels ways, the cable collecting locations, network channels and many other device fixing things depending on the media types in addition of the demand of future expansion capabilities. The results show that the using or adding of firewall device to the university campus computer network, will improve the overall network performance though increasing the data stream on many network sections and sectors, also better results are obtained.

High isolation and high linearity double‐pole double‐throw CMOS switch matrix for Ka‐band wireless transceiver front‐end IC

AbstractThis letter proposes a double‐pole double‐throw (DPDT) switch with high isolation and high linearity using a 65‐nm CMOS process. To improve the isolation, resonant inductors are employed for the core switches, which are composed of a single‐stacked NMOS with gate‐body floating techniques. And, the switch is controlled by a cross‐biasing technique to increment the power linearity. The active circuits are designed separately from the passive elements to reduce the interference. The core area of the proposed switch excluding all of the pads is 380 × 300 μm2. The minimum insertion loss is 2.57 dB at 23 GHz and maximum isolation is 39.7 dB at 26 GHz, respectively. The third‐order intermodulation intercept point (IIP3) is 37 dBm at 28 GHz.

Optical Quenching of Magnetic Vortex Visualized In Situ by Lorentz Electron Microscopy

Understanding the fundamental dynamics of topological vortex and antivortex naturally formed in micro/nanoscale ferromagnetic building blocks under external perturbations is crucial to magnetic vortex based information processing and spintronic devices. All previous studies have focused on magnetic vortex-core switching via external magnetic fields, spin-polarized currents, or spin waves, which have largely prohibited the investigation of novel spin configurations that could emerge from the ground states in ferromagnetic disks and their underlying dynamics. Here, we report in situ visualization of femtosecond laser quenching induced magnetic vortex change in various symmetric ferromagnetic Permalloy disks by Lorentz phase imaging using 4D electron microscopy. Besides the switching of magnetic vortex chirality and polarity, we observed with distinct occurrence frequencies a plenitude of complex magnetic structures that have never been observed by magnetic field or current assisted switching. These complex magnetic structures consist of a number of newly created topological magnetic defects (vortex and antivortex) strictly conserving the topological winding number, demonstrating the direct impact of topological invariant on the magnetization dynamics in ferromagnetic disks. Their spin configurations show mirror or rotation symmetry due to the geometrical confinement of the disks. Combined micromagnetic simulations with the experimental observations reveal the underlying magnetization dynamics and formation mechanism of the optical quenching induced complex magnetic structures. Their distinct occurrence rates are pertinent to their formation-growth energetics and pinning effects at the disk edge. Based on these findings, we propose a paradigm of optical-quenching-assisted fast switching of vortex cores for the control of magnetic vortex based information recording and spintronic devices.

Design of SRIO Reconstruction and Switching Network Architecture

The image processing system of the parallel processor array will face the problems of complex system structure, long development period and the inability to flexibly adjust and upgrade the platform hardware. This paper proposes a design concept of dynamic reconfiguration hardware architecture and a compatible design. Dynamic reconstruction and dynamic reconstructed image processing system of image algorithm. At the same time, CPS1848 is used as the SRIO switch chip as the switching core of the high-speed data network and a switching network routing method is designed. Tests have proved that the transmission performance of the system's SRIO high-speed data internetwork meets the requirements. Compared with previous image processing devices, the system can flexibly meet the task requirements of multiple application scenarios; improve the versatility of the device and save hardware resources and equipment costs.

A scheduling based energy-aware core switching technique to avoid thermal threshold values in multi-core processing systems

The increasing on-chip temperature has become a critical issue in the multiprocessor system on chip (MPSoC). It does not only disturb the reliability and performance of the system but it also increases the power consumption. Load balancing is a common scheduling technique to address the thermal issues where the load is transferred to relatively less used cores. However, during the transfer of workload, the destination cores may be in sleep mode and thus take some time to switch to the running mode. It results in wastage of some processing time. This paper presents a thermal-aware load balancing technique that avoids thermal emergencies while eliminating the switching delays. The proposed technique estimates the time taken by a task set to reach the temperature threshold values with the help of offline recorded thermal profiles of datasets. Moreover, cores are selected in a round-robin manner to equilibrate thermal gradients. The technique is based on Global Earliest Deadline First (GEDF) scheduling algorithm and considers ambient temperature, thermal cycles and energy consumption. The proposed technique is tested in a simulation environment comprising of scheduling and thermal simulation model. The results show that the technique reduces the average temperature up to 17%, thermal cycles up to 15%, energy consumption up to 20% and lowers the temporal and spatial gradients as compared to the commonly used predictive thermal-aware and thermal balancing techniques.

Architecture and morphodynamics of subcritical sediment waves in an ancient channel–lobe transition zone

AbstractIn modern systems, submarine channel–lobe transition zones show a well‐documented assemblage of depositional and erosional bedforms. In contrast, the stratigraphic record of channel–lobe transition zones is poorly constrained, because preservation potential is low and criteria have not been established to identify depositional bedforms in these settings. Several locations from an exhumed fine‐grained base of slope system (Unit B, Laingsburg depocentre, Karoo Basin, South Africa) show exceptional preservation of sandstone beds with distinctive morphologies and internal facies distributions. The regional stratigraphic context, lack of a basal confining surface, wave‐like morphology in dip section, size and facies characteristics support an interpretation of subcritical sediment waves within a channel–lobe transition zone setting. Some sediment waves show steep (10 to 25°) unevenly spaced (10 to 100 m) internal truncation surfaces that are dominantly upstream‐facing, which suggests significant spatio‐temporal fluctuations in flow character. Their architecture indicates that individual sediment wave beds accrete upstream, in which each swell initiates individually. Lateral switching of the flow core is invoked to explain the sporadic upstream‐facing truncation surfaces, and complex facies distributions vertically within each sediment wave. Variations in bedform character are related to the axial to marginal positions within a channel–lobe transition zone. The depositional processes documented do not correspond with known bedform development under supercritical conditions. The proposed process model departs from established mechanisms of sediment wave formation by emphasising the evidence for subcritical rather than supercritical conditions, and highlights the significance of lateral and temporal variability in flow dynamics and resulting depositional architecture.

Hybrid Routing and Adaptive Spectrum Allocation for Flex-Grid Optical Interconnects

A hybrid routing scheme with an adaptive spectrum assignment is proposed for flex-grid all-optical core switch supporting multihop transparent paths in data center networks. Compared with conventional spectrum assignment algorithms (RSA) developed for a multihop network with optical-electric-optical (OEO) conversion in every hop (i.e., RSA for EO) and that devised for an alloptical multihop network (i.e., RSA for AO), the present RSA algorithm provides better utilization of network resources. Being aware of the all-optical bypass path in hopping, the proposed RSA reduces the blocking probability due to lack of bandwidth-tunable transceivers, which is the major reason for blocking for an RSA for the EO. Similar to the RSA for the AO, the proposed RSA is compatible with the number-of-hops adaptive spectrum assignment, which improves spectrum efficiency. On the other hand, the new algorithm enhances connectivity by eliminating the number-of-hops limitation, which severely constrains the performance of RSA for the AO. Simulations for the system are carried out to investigate the performance of the new algorithm. The impacts of various parameters, such as traffic load, ratio of connection requests with different data rates, and resource configuration on the link cost, are studied in terms of network blocking probability (BP). The achievable traffic load of the proposed RSA under varied connection degrees (i.e., the maximum number of ports that one rack has in order to connect to the core switch) and number of racks is also assessed to keep BP no more than 0.1. The results show that the proposed RSA with appropriate cost functions outperforms the EO and AO, which implies that it has the highest scalability.

Writing and storing information in an array of magnetic vortex nanodisks using their azimuthal modes

The switching of a vortex core of a single disk in an array of a multilayer system is investigated by micromagnetic simulation. We found that the perpendicular uniaxial anisotropy decreases the frequencies of the azimuthal mode in disks with magnetic vortex configuration. We obtained a phase diagram of magnetic field intensity vs. frequency of the azimuthal mode, as a function of the value of perpendicular uniaxial anisotropy. We demonstrated that rotating magnetic fields (CW and CCW) with frequency equal to azimuthal modes can be used to switch the vortex core of single disks in a disk array. This allows obtaining different memory states with a single array of nanodisks, and therefore writing information through the application of rotating fields.

25–34 GHz Single-Pole, Double-Throw CMOS Switches for a Ka-Band Phased-Array Transceiver

This paper presents two single-pole, double-throw (SPDT) mm-wave switches for Ka-band phased-array transceivers, fabricated with a 65-nm complementary metal oxide semiconductor (CMOS) process. One switch employs cross-biasing (CB) control with a single supply, while the other uses dual-supply biasing (DSB) control with positive and negative voltages. Negative voltages were generated internally, using a ring oscillator and a charge pump. Identical gate and body floated N-type metal oxide semiconductor field effect transistors (N-MOSFETs) in a triple well were used as the switch core transistors. Inductors were used to improve the isolation between the transmitter (TX) and receiver (RX), as well as insertion loss, by canceling the parasitic capacitance of the switch core transistors at resonance. The size of the proposed radio frequency (RF) switch is 260 μm × 230 μm, excluding all pads. The minimum insertion losses of the CB and DSB switches were 2.1 dB at 28 GHz and 1.93 dB at 24 GHz, respectively. Between 25 GHz and 34 GHz, the insertion losses were less than 2.3 dB and 2.5 dB, the return losses were less than 16.7 dB and 17.3 dB, and the isolation was over 18.4 dB and 15.3 dB, respectively. The third order input intercept points (IIP3) of the CB and DSB switches were 38.4 dBm and 39 dBm at 28 GHz, respectively.

Application and Analysis of Multicast Blocking Modelling in Fat-Tree Data Center Networks

Multicast can improve network performance by eliminating unnecessary duplicated flows in the data center networks (DCNs). Thus it can significantly save network bandwidth. However, the network multicast blocking may cause the retransmission of a large number of data packets and seriously influence the traffic efficiency in data center networks, especially in the fat-tree DCNs with multirooted tree structure. In this paper, we build a multicast blocking model and apply it to solve the problem of network blocking in the fat-tree DCNs. Furthermore, we propose a novel multicast scheduling strategy. In the scheduling strategy, we select the uplink connecting to available core switch whose remaining bandwidth is close to and greater than the three times of bandwidth multicast requests so as to reduce the operation time of the proposed algorithm. Then the blocking probability of downlink in the next time-slot is calculated in multicast subnetwork by using Markov chains theory. With the obtained probability, we select the optimal downlink based on the available core switch. In addition, theoretical analysis shows that the multicast scheduling algorithm has close to zero network blocking probability as well as lower time complexity. Simulation results verify the effectiveness of our proposed multicast scheduling algorithm.

ZVT Resonant Core Reset Forward Converter With a Simple Auxiliary Circuit

In this paper, a new zero-voltage switching and resonant core reset pulse width modulation forward converter with a simple auxiliary circuit is introduced. The proposed auxiliary circuit can be applied to both single-switch and two-switch forward converters. The main and the auxiliary mosfet s of the proposed converters operate under soft switching condition independent of line and load variation. Also, due to the capability of operating at duty cycles higher than 0.5, the proposed converters can be used for wide input voltage applications. In comparison to the existing zero-voltage transition (ZVT) forward converters, this auxiliary circuit has low number of components and also does not require any additional magnetic element which has improved the power density. The proposed converter is compared with exiting ZVT forward converters in details. The converter steady-state operation is analyzed and various operating modes are explained. Also, in addition to a design example, a comprehensive design procedure is presented. The experimental results of a laboratory prototype of the converter with 85–180 V input voltage and 24 V output voltage at 150 W are provided which verify the theoretical analysis and confirm the converter operation. The overall efficiency of 93.5% is achieved for the proposed converter at full load.

Selective pH-Responsive Core-Sheath Nanofiber Membranes for Chem/Bio/Med Applications: Targeted Delivery of Functional Molecules.

Core-sheath fibers using different Eudragit materials were successfully produced, and their controlled multi-pH responses have been demonstrated. Core-sheath fibers made of Eudragit L 100 (EL100) core and Eudragit S 100 (ES100) sheath provide protection and/or controlled release of core material at pH 6 by adjusting the sheath thickness (controlled by the flow rate of source polymer solution). The thickest sheath (∼250 nm) provides the least core release ∼1.25%/h, while the thinnest sheath (∼140 nm) provides much quicker release ∼16.75%/h. Furthermore, switching core and sheath material dramatically altered the pH response. Core-sheath fibers made of ES100 core and EL100 sheath can provide a consistent core release rate, while the sheath release rate becomes higher as the sheath layer becomes thinner. For example, the thinnest sheath (∼120 nm) provides a core and sheath release ratio of 1:2.5, while the thickest sheath (∼200 nm) shows only a ratio of 1:1.7. All core-sheath Eudragit fibers show no noticeable release at pH 5, while they are completely dissolved at pH 7. Extremely high surface area in the porous network of the fiber membranes provides much faster (>30 times) response to external pH changes as compared to that of equivalent cast films.

Loss estimation and control mechanism in bufferless optical packet-switched networks based on multilayer perceptron

Artificial neural networks (ANNs) are well-known estimators for the output of broad range of complex systems and functions. In this paper, a common ANN architecture called multilayer perceptron (MLP) is used as a fast optical packet loss rate (OPLR) estimator for bufferless optical packet-switched (OPS) networks. Considering average loads at the ingress switches of an OPS network, the proposed estimator estimates total OPLR as well as ingress OPLRs (the OPLR of optical packets sent from individual ingress switches). Moreover, a traffic policing algorithm called OPLRC is proposed to control ingress OPLRs in bufferless slotted OPS networks with asymmetric loads. OPLRC is a centralized greedy algorithm which uses estimated ingress OPLRs of a trained MLP to tag some optical packets at the ingress switches as eligible for drop at the core switches in case of contention. This will control ingress OPLRs of un-tagged optical packets within the specified limits while giving some chance for tagged optical packets to reach their destinations. Eventually, the accuracy of the proposed estimator along with the performance of the proposed algorithm is evaluated by extensive simulations. In terms of the algorithm, the results show that OPLRC is capable of controlling ingress OPLRs of un-tagged optical packets with an acceptable accuracy.

A simple and versatile system for the ATP-dependent assembly of chromatin

Chromatin is the natural form of DNA in the eukaryotic nucleus and is the substrate for diverse biological phenomena. The functional analysis of these processes ideally would be carried out with nucleosomal templates that are assembled with customized core histones, DNA sequences, and chromosomal proteins. Here we report a simple, reliable, and versatile method for the ATP-dependent assembly of evenly spaced nucleosome arrays. This minimal chromatin assembly system comprises the Drosophila nucleoplasmin-like protein (dNLP) histone chaperone, the imitation switch (ISWI) ATP-driven motor protein, core histones, template DNA, and ATP. The dNLP and ISWI components were synthesized in bacteria, and each protein could be purified in a single step by affinity chromatography. We show that the dNLP-ISWI system can be used with different DNA sequences, linear or circular DNA, bulk genomic DNA, recombinant or native Drosophila core histones, native human histones, the linker histone H1, the non-histone chromosomal protein HMGN2, and the core histone variants H3.3 and H2A.V. The dNLP-ISWI system should be accessible to a wide range of researchers and enable the assembly of customized chromatin with specifically desired DNA sequences, core histones, and other chromosomal proteins.

A Study of the Number of Wavelengths Impact in the Optical Burst Switching Core Node

A Study of the Number of Wavelengths Impact in the Optical Burst Switching Core Node

Packet Loss Rate Differentiation in slotted Optical Packet Switching OCDM/WDM

We propose a multi-class mechanism for Optical Code Division Multiplexing (OCDM), Wavelength Division Multiplexing (WDM) Optical Packet Switch (OPS) architecture capable of supporting Quality of Service (QoS) transmission. OCDM/WDM has been proposed as a competitive hybrid switching technology to support the next generation optical Internet. This paper addresses performance issues in the slotted OPS networks and proposed four differentiation schemes to support Quality of Service. In addition, we present a comparison between the proposed schemes as well as, a simulation scheduler design which can be suitable for the core switch node in OPS networks. Using software simulations the performance of our algorithm in terms of losing probability, the packet delay, and scalability is evaluated.