Evaluating fault tolerance on asymmetric multicore systems-on-chip using iso-metrics

The end of Dennard scaling has promoted low power consumption into a firstorder concern for computing systems. However, conventional power conservation schemes such as voltage and frequency scaling are reaching their limits when used in performance-constrained environments. New technologies are required to break the power wall while sustaining performance on future processors. Low-power embedded processors and near-threshold voltage computing (NTVC) have been proposed as viable solutions to tackle the power wall in future computing systems. Unfortunately, these technologies may also compromise per-core performance and, in the case of NTVC, xreliability. These limitations would make them unsuitable for HPC systems and datacenters. In order to demonstrate that emerging low-power processing technologies can effectively replace conventional technologies, this study relies on ARM’s big.LITTLE processors as both an actual and emulation platform, and state-of-the-art implementations of the CG solver. For NTVC in particular, the paper describes how efficient algorithm-based fault tolerance schemes preserve the power and energy benefits of very low voltage operation.

Inexact-aware architecture design for ultra-low power bio-signal analysis

This study introduces an inexact, but ultra-low power, computing architecture devoted to the embedded analysis of bio-signals. The platform operates at extremely low voltage supply levels to minimise energy consumption. In this scenario, the reliability of static RAM (SRAM) memories cannot be guaranteed when using conventional 6-transistor implementations. While error correction codes and dedicated SRAM implementations can ensure correct operations in this near-threshold regime, they incur in significant area and energy overheads, and should therefore be employed judiciously. Herein, the authors propose a novel scheme to design inexact computing architectures that selectively protects memory regions based on their significance, i.e. their impact on the end-to-end quality of service, as dictated by the bio-signal application characteristics. The authors illustrate their scheme on an industrial benchmark application performing the power spectrum analysis of electrocardiograms. Experimental evidence showcases that a significance-based memory protection approach leads to a small degradation in the output quality with respect to an exact implementation, while resulting in substantial energy gains, both in the memory and the processing subsystem.

Fine-structure photoionization cross sections of Fe II

We present the first calculation of fine-structure photoionization cross sections for the ground state of singly ionized Fe. These large-scale ab initio calculations, limited to the near-threshold region, were performed in the close-coupling approximation using a Dirac–Coulomb R -matrix method implemented within a modified version of the DARC package. Our calculated cross sections reproduce in detail the resonance structures observed in previous experimental determinations.

Dielectronic recombination of W20 + (4d104f8): Addressing the half-open f shell

A recent measurement of the dielectronic recombination (DR) of W20+ [Schippers et al., Phys.Rev.A 83, 012711 (2011)] found an exceptionally large contribution from near-threshold resonances (1 eV). This still affected the Maxwellian rate coefficient at much higher temperatures. The experimental result was found to be higher by a factor of 4 or more than that currently in use in the 100- to 300-eV range, which is of relevance for modeling magnetic fusion plasmas. We have carried out DR calculations with AUTOSTRUCTURE which include all significant single-electron promotions. Our intermediate-coupling (IC) results are more than a factor of 4 larger than our LS-coupling ones at 1 eV but still lie a factor of 3 below experiment here. If we assume complete (chaotic)mixing of near-threshold autoionizing states, then our results come into agreement (to within 20%)with experiment below 2 eV. Our total IC Maxwellian rate coefficients are 50%–30% smaller than those based on experiment over 100–300 eV.

Electron impact fragmentation of thymine:partial ionization cross sections for positive fragments

We have measured mass spectra for positive ions for low-energy electron impact on thymine using a reflectron time-of-flight mass spectrometer. Using computer controlled data acquisition, mass spectra have been acquired for electron impact energies up to 100 eV in steps of 0.5 eV. Ion yield curves for most of the fragment ions have been determined by fitting groups of adjacent peaks in the mass spectra with sequences of normalized Gaussians. The ion yield curves have been normalized by comparing the sum of the ion yields to the average of calculated total ionization cross sections. Appearance energies have been determined. The nearly equal appearance energies of 83 u and 55 u observed in the present work strongly indicate that near threshold the 55 u ion is formed directly by the breakage of two bonds in the ring, rather than from a successive loss of HNCO and CO from the parent ion. Likewise 54 u is not formed by CO loss from 82 u. The appearance energies are in a number of cases consistent with the loss of one or more hydrogen atoms from a heavier fragment, but 70 u is not formed by hydrogen loss from 71 u.

Visual motion interferes with lexical decision on motion words

Embodied theories of cognition propose that neural substrates used in experiencing the referent of a word, for example perceiving upward motion, should be engaged in weaker form when that word, for example ‘rise’, is comprehended. Motivated by the finding that the perception of irrelevant background motion at near-threshold, but not supra-threshold, levels interferes with task execution, we assessed whether interference from near-threshold background motion was modulated by its congruence with the meaning of words (semantic content) when participants completed a lexical decision task (deciding if a string of letters is a real word or not). Reaction times for motion words, such as ‘rise’ or ‘fall’, were slower when the direction of visual motion and the ‘motion’ of the word were incongruent — but only when the visual motion was at nearthreshold levels. When motion was supra-threshold, the distribution of error rates, not reaction times, implicated low-level motion processing in the semantic processing of motion words. As the perception of near-threshold signals is not likely to be influenced by strategies, our results support a close contact between semantic information and perceptual systems.

Projectile breakup dynamics for (6)Li+(59)Co: Kinematical analysis of alpha-d coincidences

A study of the kinematics of the alpha-d coincidences in the (6)Li + (59)Co system at a bombarding energy of E(lab) = 29.6MeV is presented. With exclusive measurements performed over different angular intervals it is possible to identify the respective contributions of the sequential and direct projectile breakup components. The angular distributions of both breakup components are fairly well described by the Continuum-Discretized Coupled-Channels framework (CDCC). Furthermore, a careful analysis of these processes using a semiclassical approach provides information on both their lifetime and their distance of occurrence with respect to the target. Breakup to the low-lying (near-threshold) continuum is delayed, and happens at large internuclear distances. This suggests that the influence of the projectile breakup on the complete fusion process can be related essentially to the direct breakup to the (6)Li high-lying continuum spectrum.

Pion interactions in the X(3872)

We consider pion interactions in an effective field theory of the narrow resonance X(3872), assuming it is a weakly bound molecule of the charm mesons D-0(D) over bar (*0) and D-*0(D) over bar (0). Since the hyperfine splitting of the D-0 and D-*0 is only 7 MeV greater than the neutral pion mass, pions can be produced near threshold and are nonrelativistic. We show that pion exchange can be treated in perturbation theory and calculate the next-to-leading-order correction to the partial decay width Gamma[X -> D-0(D) over bar (0)pi(0)].

Correlation times in stochastic equations with delayed feedback and multiplicative noise

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

The Gerasimov-Drell-Hearn Sum Rule and the Spin Structure of the Proton

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Observation of a peaking structure in the J/psi phi mass spectrum from B-+/- -> J/psi phi K-+/- decays

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Parallel Architectures for Many-Core Systems-On-Chip in Deep Sub-Micron Technology

Despite the several issues faced in the past, the evolutionary trend of silicon has kept its constant pace. Today an ever increasing number of cores is integrated onto the same die. Unfortunately, the extraordinary performance achievable by the many-core paradigm is limited by several factors. Memory bandwidth limitation, combined with inefficient synchronization mechanisms, can severely overcome the potential computation capabilities. Moreover, the huge HW/SW design space requires accurate and flexible tools to perform architectural explorations and validation of design choices. In this thesis we focus on the aforementioned aspects: a flexible and accurate Virtual Platform has been developed, targeting a reference many-core architecture. Such tool has been used to perform architectural explorations, focusing on instruction caching architecture and hybrid HW/SW synchronization mechanism. Beside architectural implications, another issue of embedded systems is considered: energy efficiency. Near Threshold Computing is a key research area in the Ultra-Low-Power domain, as it promises a tenfold improvement in energy efficiency compared to super-threshold operation and it mitigates thermal bottlenecks. The physical implications of modern deep sub-micron technology are severely limiting performance and reliability of modern designs. Reliability becomes a major obstacle when operating in NTC, especially memory operation becomes unreliable and can compromise system correctness. In the present work a novel hybrid memory architecture is devised to overcome reliability issues and at the same time improve energy efficiency by means of aggressive voltage scaling when allowed by workload requirements. Variability is another great drawback of near-threshold operation. The greatly increased sensitivity to threshold voltage variations in today a major concern for electronic devices. We introduce a variation-tolerant extension of the baseline many-core architecture. By means of micro-architectural knobs and a lightweight runtime control unit, the baseline architecture becomes dynamically tolerant to variations.

Dendritic spikes in apical dendrites of neocortical layer 2/3 pyramidal neurons

Layer 2/3 (L2/3) pyramidal neurons are the most abundant cells of the neocortex. Despite their key position in the cortical microcircuit, synaptic integration in dendrites of L2/3 neurons is far less understood than in L5 pyramidal cell dendrites, mainly because of the difficulties in obtaining electrical recordings from thin dendrites. Here we directly measured passive and active properties of the apical dendrites of L2/3 neurons in rat brain slices using dual dendritic-somatic patch-clamp recordings and calcium imaging. Unlike L5 cells, L2/3 dendrites displayed little sag in response to long current pulses, which suggests a low density of I(h) in the dendrites and soma. This was also consistent with a slight increase in input resistance with distance from the soma. Brief current injections into the apical dendrite evoked relatively short (half-width 2-4 ms) dendritic spikes that were isolated from the soma for near-threshold currents at sites beyond the middle of the apical dendrite. Regenerative dendritic potentials and large concomitant calcium transients were also elicited by trains of somatic action potentials (APs) above a critical frequency (130 Hz), which was slightly higher than in L5 neurons. Initiation of dendritic spikes was facilitated by backpropagating somatic APs and could cause an additional AP at the soma. As in L5 neurons, we found that distal dendritic calcium transients are sensitive to a long-lasting block by GABAergic inhibition. We conclude that L2/3 pyramidal neurons can generate dendritic spikes, sharing with L5 pyramidal neurons fundamental properties of dendritic excitability and control by inhibition.

A perceptual memory for low-contrast visual signals

Detection of a visual signal can be facilitated by simultaneous presentation of a similar subthreshold signal. Here we show that the facilitatory effect of a subthreshold signal can persist for more than 16 s. Presenting a near-threshold Gabor signal (prime) produced a phase-independent increase in contrast sensitivity (40%) to similar successive signals (target) for a period of up to 16 s. This effect was obtained only when both prime and target were presented to the same eye. We further show that the memory trace is inactivated by presenting high-contrast signals before the target. These results suggest that activated neurons in the primary visual cortex retain a near-threshold memory trace that persists until reactivated.

Mechanical bases of frequency tuning and neural excitation at the base of the cochlea: Comparison of basilar-membrane vibrations and auditory-nerve-fiber responses in chinchilla

We review the mechanical origin of auditory-nerve excitation, focusing on comparisons of the magnitudes and phases of basilar-membrane (BM) vibrations and auditory-nerve fiber responses to tones at a basal site of the chinchilla cochlea with characteristic frequency ≈ 9 kHz located 3.5 mm from the oval window. At this location, characteristic frequency thresholds of fibers with high spontaneous activity correspond to magnitudes of BM displacement or velocity in the order of 1 nm or 50 μm/s. Over a wide range of stimulus frequencies, neural thresholds are not determined solely by BM displacement but rather by a function of both displacement and velocity. Near-threshold, auditory-nerve responses to low-frequency tones are synchronous with peak BM velocity toward scala tympani but at 80–90 dB sound pressure level (in decibels relative to 20 microPascals) and at 100–110 dB sound pressure level responses undergo two large phase shifts approaching 180°. These drastic phase changes have no counterparts in BM vibrations. Thus, although at threshold levels the encoding of BM vibrations into spike trains appears to involve only relatively minor signal transformations, the polarity of auditory-nerve responses does not conform with traditional views of how BM vibrations are transmitted to the inner hair cells. The response polarity at threshold levels, as well as the intensity-dependent phase changes, apparently reflect micromechanical interactions between the organ of Corti, the tectorial membrane and the subtectorial fluid, and/or electrical and synaptic processes at the inner hair cells.