Use of data dependence graphs in the design of bit-level systolic arrays

The use of bit-level systolic array circuits as building blocks in the construction of larger word-level systolic systems is investigated. It is shown that the overall structure and detailed timing of such systems may be derived quite simply using the dependence graph and cut-set procedure developed by S. Y. Kung (1988). This provides an attractive and intuitive approach to the bit-level design of many VLSI signal processing components. The technique can be applied to ripple-through and partly pipelined circuits as well as fully systolic designs. It therefore provides a means of examining the relative tradeoff between levels of pipelining, chip area, power consumption, and throughput rate within a given VLSI design.

DESIGNING VLSI SIGNAL PROCESSING CHIPS USING DATA DEPENDENCE GRAPHS.

The highly structured nature of many digital signal processing operations allows these to be directly implemented as regular VLSI circuits. This feature has been successfully exploited in the design of a number of commercial chips, some examples of which are described. While many of the architectures on which such chips are based were originally derived on heuristic basis, there is an increasing interest in the development of systematic design techniques for the direct mapping of computations onto regular VLSI arrays. The purpose of this paper is to show how the the technique proposed by Kung can be readily extended to the design of VLSI signal processing chips where the organisation of computations at the level of individual data bits is of paramount importance. The technique in question allows architectures to be derived using the projection and retiming of data dependence graphs.

Analysis of Dependence Tracking Algorithms for Task Dataflow Execution

Processor architectures has taken a turn towards many-core processors, which integrate multiple processing cores on a single chip to increase overall performance, and there are no signs that this trend will stop in the near future. Many-core processors are harder to program than multi-core and single-core processors due to the need of writing parallel or concurrent programs with high degrees of parallelism. Moreover, many-cores have to operate in a mode of strong scaling because of memory bandwidth constraints. In strong scaling increasingly finer-grain parallelism must be extracted in order to keep all processing cores busy.

Task dataflow programming models have a high potential to simplify parallel program- ming because they alleviate the programmer from identifying precisely all inter-task de- pendences when writing programs. Instead, the task dataflow runtime system detects and enforces inter-task dependences during execution based on the description of memory each task accesses. The runtime constructs a task dataflow graph that captures all tasks and their dependences. Tasks are scheduled to execute in parallel taking into account dependences specified in the task graph.

Several papers report important overheads for task dataflow systems, which severely limits the scalability and usability of such systems. In this paper we study efficient schemes to manage task graphs and analyze their scalability. We assume a programming model that supports input, output and in/out annotations on task arguments, as well as commutative in/out and reductions. We analyze the structure of task graphs and identify versions and generations as key concepts for efficient management of task graphs. Then, we present three schemes to manage task graphs building on graph representations, hypergraphs and lists. We also consider a fourth edge-less scheme that synchronizes tasks using integers. Analysis using micro-benchmarks shows that the graph representation is not always scalable and that the edge-less scheme introduces least overhead in nearly all situations.

Effects of trauma on bipolar disorder: the mediational role of interpersonal difficulties and alcohol dependence

Objectives: This study examined: (i) the prevalence of trauma in a bipolar disorder (BD) sample, and (ii) how trauma histories mediated by interpersonal difficulties and alcohol dependence impact on the severity of BD. The prevalence of posttraumatic stress disorder (PTSD) and its relationship to outcomes in BD were also examined.

Methods: Sixty participants were recruited from a geographically well-defined mental health service in Northern Ireland. Self-reported trauma histories, PTSD, interpersonal difficulties and alcohol dependence and were examined in relation to illness severity.

Results: A high prevalence of trauma was found. Trauma predicted the frequency of hospital admissions (R-2 = 0.08), quality of life (R-2 = 0.23) and inter-episode depressive symptoms (R-2 = 0.13). Interpersonal difficulties, but not alcohol dependence, appeared to play an important role in mediating these adverse effects. While only 8% of the sample met criteria for active PTSD, this comorbid disorder was associated with BD severity.

Conclusions: This study indicates that awareness of trauma is important in understanding individual differences in bipolar presentations. The theoretical and clinical implications of evidence that trauma is related to more adverse outcomes in BD are discussed. The finding that interpersonal difficulties mediate the relationship between trauma and BD severity is novel. The need for adjunctive evidence-based treatments targeting interpersonal difficulties is considered.

Energy dependence for pair production with electron capture in relativistic heavy-ion collisions

In 'Charge transfer from the negative-energy continuum: alternative mechanism for pair production in relativistic atomic collisions', Eichler (1995 Phys. Rev. Lett. 75 3653) proposes an alternative mechanism for capture by pair production, and from it derives an analytic expression for the total cross section with a surprisingly strong energy dependence. We show that, in fact, there is no alternative mechanism; rather the above mechanism may be more transparently viewed as an ionization-like transition in one centre with inclusion of continuum distortion by the second centre. We further show that to Centre the initial and final states on the target and projectile leads to confusion in the momentum transfer vectors, and hence, respectively that the alleged high-energy behaviour is erroneous.

Geometrical dependence in photoionization of H-2(+) in high-intensity, high-frequency, ultrashort laser pulses

The ionization dynamics of H2 + exposed to high-intensity, high-frequency, ultrashort laser pulses is investigated with two theoretical approaches. The time-dependent Schrödinger equation is solved by a direct numerical method, and a simple two-center interference-diffraction model is studied. The energy and angular distributions of the photoelectron for various internuclear distances and relative orientations between the internuclear axis of the molecule and the polarization of the field are calculated. The main features of the photoelectron spectrum pattern are described well by the interference-diffraction model, and excellent quantitative agreement between the two methods is found. The effect of quantal vibration on the photoelectron spectrum is also calculated. We find that vibrational average produces some broadening of the main features, but that the patterns remain clearly distinguishable.

Time dependence of rotational state populations of excited hydrogen molecules in an RF excited plasma reactor

We report on time-dependent population distributions of excited rotational states of hydrogen in a capacitively coupled RF discharge. The common model to obtain the gas temperature from the rotational distribution is not applicable at all times during the discharge cycle due to the time dependence of the EEDF. The apparent temperature within a cycle assumes values between 350 K and 450 K for the discharge parameters of this experiment. We discuss the optimum time window within the discharge cycle that yields the best approximation to the actual temperature. Erroneous results can be obtained, in principle, with time-integrated measurements; we find, however, that in the present case the systematic error amounts to only approximately 20 K. This is due to the fact that the dominant contribution to the average intensity arises during that time window for which the assumptions underlying the analysis are best fulfilled. A similar analysis can be performed for N+2 rotational bands with a small amount of nitrogen added to the discharge gas. These populations do not exhibit the time variations found in the case of H2.