The explosion of digital data and the ever-growing need for fast data analysis have made in-memory big-data processing in computer systems increasingly important. In particular, large-scale graph processing is gaining attention due to its broad applicability from social science to machine learning. However, scalable hardware design that can efficiently process large graphs in main memory is still an open problem. Ideally, cost-effective and scalable graph processing systems can be realized by building a system whose performance increases proportionally with the sizes of graphs that can be stored in the system, which is extremely challenging in conventional systems due to severe memory bandwidth limitations.In this work, we argue that the conventional concept of processing-in-memory (PIM) can be a viable solution to achieve such an objective. The key modern enabler for PIM is the recent advancement of the 3D integration technology that facilitates stacking logic and memory dies in a single package, which was not available when the PIM concept was originally examined. In order to take advantage of such a new technology to enable memory-capacity-proportional performance, we design a programmable PIM accelerator for large-scale graph processing called Tesseract. Tesseract is composed of (1) a new hardware architecture that fully utilizes the available memory bandwidth, (2) an efficient method of communication between different memory partitions, and (3) a programming interface that reflects and exploits the unique hardware design. It also includes two hardware prefetchers specialized for memory access patterns of graph processing, which operate based on the hints provided by our programming model. Our comprehensive evaluations using five state-of-the-art graph processing workloads with large real-world graphs show that the proposed architecture improves average system performance by a factor of ten and achieves 87% average energy reduction over conventional systems.

a b s t r a c tIncreasingly stringent limits from LHC searches for new physics, coupled with lack of convincing signals of weakly interacting massive particle (WIMP) in dark matter searches, have tightly constrained many realizations of the standard paradigm of thermally produced WIMPs as cold dark matter. In this article, we review more generally both thermally and non-thermally produced dark matter (DM). One may classify DM models into two broad categories: one involving bosonic coherent motion (BCM) and the other involving WIMPs. BCM and WIMP candidates need, respectively, some approximate global symmetries and almost exact discrete symmetries. Supersymmetric axion models are highly motivated since they emerge from compelling and elegant solutions to the two fine-tuning problems of the Standard Model: the strong CP problem and the gauge hierarchy problem. We review here non-thermal relics in a general setup, but we also pay particular attention to the rich cosmological properties of various aspects of mixed SUSY/axion dark matter candidates which can involve both WIMPs and BCM in an interwoven manner. We also review briefly a panoply of alternative thermal and non-thermal DM candidates.

Power efficient design of real-time systems based on programmable processors becomes more important as system functionality is increasingly realized through software. This paper presents a powerefficient version of a widely used fixed priority scheduling method. The method yields a power reduction by exploiting slack times, both those inherent in the system schedule and those arising from variations of execution times. The proposed run-time mechanism is simple enough to be implemented in most kernels. Experimental results show that the proposed scheduling method obtains a significant power reduction across several kinds of applications.

No abstract

In the framework of the CMSSM we study the gravitino as the lightest supersymmetric particle and the dominant component of cold dark matter in the Universe. We include both a thermal contribution to its relic abundance from scatterings in the plasma and a non-thermal one from neutralino or stau decays after freeze-out. In general both contributions can be important, although in different regions of the parameter space. We further include constraints from BBN on electromagnetic and hadronic showers, from the CMB blackbody spectrum and from collider and non-collider SUSY searches. The region where the neutralino is the next-to-lightest superpartner is severely constrained by a conservative bound from excessive electromagnetic showers and probably basically excluded by the bound from hadronic showers, while the stau case remains mostly allowed. In both regions the constraint from CMB is often important or even dominant. In the stau case, for the assumed reasonable ranges of soft SUSY breaking parameters, we find regions where the gravitino abundance is in agreement with the range inferred from CMB studies, provided that, in many cases, a reheating temperature T R is large, T R ∼ 10 9 GeV. On the other side, we find an upper bound T R ∼ < 5 × 10 9 GeV. Less conservative bounds from BBN or an improvement in measuring the CMB spectrum would provide a dramatic squeeze on the whole scenario, in particular it would strongly disfavor the largest values of T R ∼ 10 9 GeV. The regions favored by the gravitino dark matter scenario are very different from standard regions corresponding to the neutralino dark matter, and will be partly probed at the LHC.

We consider cosmological consequences of a heavy axino, decaying to the neutralino in R-parity conserving models. The importance and influence of the axino decay on the resultant abundance of neutralino dark matter depends on the lifetime and the energy density of axino. For a high reheating temperature after inflation, copiously produced axinos dominate the energy density of the universe and its decay produces a large amount of entropy. As a bonus, we obtain that the upper bound on the reheating temperature after inflation via gravitino decay can be moderated, because the entropy production by the axino decay more or less dilutes the gravitinos. I. AXINONeutralino, if it is the lightest supersymmetric particle (LSP) in R-parity conserving models, is a natural candidate for dark matter. Because of the TeV scale sparticle interactions, the thermal history of neutralinos allows the neutralino dark matter possibility. But, imposing a solution of the strong CP problem, the thermal history involves contributions from the additional sector.The strong CP problem is naturally solved by introducing a very light axion a. Most probably, it appears when the Peccei-Quinn (PQ) symmetry is broken at a scale of f a . Below the PQ scale, the effective axion interaction with gluons iswhere g s is the strong coupling constant [1]. The PQ scale is constrained by the astrophysical and cosmological considerations in the narrow window 10 10 GeV f a 10 12 GeV [2].TeV scale supersymmetry (SUSY) suggests axinoã, the superpartner of axion, around the electroweak scale in the gravity mediation scenario. Here, we consider the effects of heavy axinos in cosmology. The axino cosmology depends crucially on the axino decoupling temperature [3], The axion supermultiplet includes axion, saxion (the scalar partner) and axino. Both saxion and axino masses are split from the almost vanishing axion mass if SUSY is broken. The precise value of the axino mass depends on the model, specified by the SUSY breaking sector and the mediation sector to the axion supermultiplet [4]. In principle, the axion supermultiplet is independent from the observable sector in which case we may take the axino mass as a free parameter of order from keV to a value much larger than the gravitino mass [5,6]. Light axinos can be a dark matter (DM) candidate, which has been studied extensively [7,8,9]. Heavy axinos, however, cannot be the LSP and can decay to the LSP plus light particles. This heavy axino decay to neutralino was considered in the literature [6] where the neutralino relic density was not considered seriously. Some considered the axino as the next LSP decaying to the gravitino LSP in the gauge mediated SUSY breaking scenario [10]. Recently, supersymmetric axion models were studied with an emphasis on saxion [11], where the heavy axino possibility was also considered briefly [12].In this paper, we present a more or less complete cosmological analysis of a heavy axino with mass in the TeV region so that it is heavier than the LSP neutralino. Compared with the saxi...

We study the level of primordial non-Gaussianity in slow-roll two-field inflation. Using an analytic formula for the nonlinear parameter fNL in the case of a sum or product separable potential, we find that it is possible to generate significant non-Gaussianity even during slow-roll inflation with Gaussian perturbations at Hubble exit. In this paper we give the general conditions to obtain large non-Gaussianity and calculate the level of fine-tuning required to obtain this. We present explicit models in which the non-Gaussianity at the end of inflation can exceed the current observational bound of |fNL| 100.

We review models which generate a large non-Gaussianity of the local form. We first briefly consider three models which generate the non-Gaussianity either at or after the end of inflation; the curvaton scenario, modulated (p)reheating and an inhomogeneous end of inflation. We then focus on ways of generating the non-Gaussianity during inflation. We derive general conditions which a product or sum separable potential must satisfy in order to generate a large local bispectrum during slow-roll inflation. As an application we consider two-field hybrid inflation. We then derive a formalism not based on slow roll which can be applied to models in which the slow-roll parameters become large before inflation ends. An exactly soluble two-field model is given in which this happens. Finally we also consider further non-Gaussian observables; a scale dependence of fNL and the trispectrum. * Electronic address: Byrnes@physik.uni-bielefeld.de † Electronic address: kiyoung.choi@pusan.ac.kr II. SUMMARY OF POPULAR MODELS GENERATING A LARGE LOCAL NON-GAUSSIANITYMulti-field inflationary models [27] can lead to an observable signature of non-Gaussianity. These include models in which the large non-Gaussianity is generated either by the means of ending inflation, or after inflation. We review the three most popular of these models in this section, for a discussion on how these models are related see [30]. In the next section we review the possibility of generating large non-Gaussianity during multi-field slow-roll inflation. Later in Sec. V we discuss multi-field models of inflation without assuming the slow-roll conditions. This does not exhaust all of the possible ways of generating a large local non-Gaussianity from inflation, see also the reviews [31][32][33][34] and the papers [35,36]. Non-Guassianity of the local form can also be generated in the ekpyrotic scenario [37,38], although in the simplest case f N L is large and negative [39], which is observationally ruled out.