Combining speculative execution and conditional resource sharing to efficiently schedule conditional behaviors

Speculative execution is an optimization technique used in modern processors by which predicted instructions are executed in advance with an objective of overlapping the latencies of slow operations. Branch prediction and load value speculation are examples of speculative execution used in modern pipelined processors to avoid execution stalls. However, speculative executions incur a performance penalty as an execution rollback when there is a misprediction. In this work, we propose to aid speculative execution with approximate computing by relaxing the execution rollback penalty associated with a misprediction. We propose a sensitivity analysis method for data and branches in a program to identify the data load and branch instructions that can be executed without any rollback in the pipeline and yet can ensure a certain user-specified quality of service of the application with a probabilistic reliability. Our analysis is based on statistical methods, particularly hypothesis testing and Bayesian analysis. We perform an architectural simulation of our proposed approximate execution and report the benefits in terms of CPU cycles and energy utilization on selected applications from the AxBench, ACCEPT, and Parsec 3.0 benchmarks suite.

Preventing execution of unauthorized software on a given computer plays a pivotal role in system security. The key problem is that although a program at the beginning of its execution can be verified as authentic, its execution flow can be redirected to externally injected malicious code us-ing, for example, a buffer overflow exploit. We introduce a novel, simplified, hardware-assisted in-trusion prevention platform. Our platform introduces over-lapping of program execution and MAC verification. It par-titions a program binary into blocks of instructions. Each block is signed using a keyed MAC that is attached as a footer to the block. When the control flow reaches a partic-ular block, its instructions are speculatively executed, while dedicated hardware verifies the attached MAC at run-time. The computation state is preserved during speculative exe-cution using a mediating buffer placed between the proces-sor and L1 data cache. Upon MAC verification, the results from this buffer are propagated externally. Central to this paper is the proposal of a novel optimization technique that initially identifies instructions that are likely to stall exe-cution, and reorders basic blocks within a given instruc-tion block to minimize the execution overhead. While the presented optimization technique is problem specific, it is flexible such that it can be adjusted for different optimiza-tion goals. Preliminary results showed that our optimiza-tion methods produced an average overhead reduction of 60 % on the SPEC2000 benchmark suite and Microsoft Vi-sual FoxPro. 1.

Real-time scheduling on heterogeneous system-on-chip architectures using an optimised artificial neural network

This work studies the interplay between multithreaded cores and speculative parallelism (e.g., transactional memory or thread-level speculation). These techniques are often used together, yet they have been developed independently. This disconnect causes major performance pathologies: increasing the number of threads per core adds conflicts and wasted work, and puts pressure on speculative execution resources. These pathologies often squander the benefits of multithreading.We present speculation-aware multithreading (SAM), a simple policy that addresses these pathologies. By coordinating instruction dispatch and conflict resolution priorities, SAM focuses execution resources on work that is more likely to commit, avoiding aborts and using speculation resources more efficiently.We design SAM variants for in-order and out-of-order cores. SAM is cheap to implement and makes multithreaded cores much more beneficial on speculative parallel programs. We evaluate SAM on systems with up to 64 SMT cores. With SAM, 8-threaded cores outperform single-threaded cores by 2.33x on average, while a speculation-oblivious policy yields a 1.85x speedup. SAM also reduces wasted work by 52%.

Situation-dependent remote AR collaborations: Image-based collaboration using a 3D perspective map and live video-based collaboration with a synchronized VR mode

Magnetic random access memory (MRAM) consisting of magnetic tunnel junctions (MTJs) is a fascinating candidate for high-density nonvolatile memory. An MTJ has two ferromagnetic metals called free and pinned layers separated by nonmagnetic spacer. The free layer in the MTJ shows two energetically stable states corresponding to the parallel and antiparallel alignments of the magnetizations. These magnetization alignments are distinguished by tunnel magnetoresistance (TMR) effect, i.e., the resistance at the antiparallel alignment is higher than that at the parallel alignment. The free layer is switched between these two stable states by spin torque effect [1], where the transfer of the spin angular momentum from the conducting electrons to the free layer excites the spin torque and induces the magnetization dynamics. The observation of the high TMR ratio more than 100% in CoFeB/MgO-based MTJ [2] has motivated both experimental and theoretical researchers to develop practical MRAM. The free layer should stay one of the stable states to keep information in MRAM. Unfortunately, however, the thermal fluctuation induces probabilistic switching of MTJ. Therefore, high thermal stability is required to guarantee the MRAM performance. Thermal stability of the MTJ is defined by the energy barrier of the free layer separating the two stable states divided by the thermal energy, k B T. Since the retention time of MRAM is exponentially proportional to the thermal stability, a small error in the evaluation of the thermal stability affects the estimation of the retention time significantly. Thus, an accurate evaluation of the thermal stability is necessary. The thermal stability has been evaluated by analyzing the switching probability of the free layer’s magnetization in the thermally activated region assisted by the spin torque. The fitting function of the switching probability is P=1-exp[-ft exp[-Δ(1-I/I c) b ]], where f, Δ, I, and I c are the attempt frequency, thermal stability, current, and switching current at zero temperature, respectively. The switching exponent, b, is assumed to be one [3]. In this study, we study the thermally assisted spin torque switching in an in-plane magnetized MTJ theoretically. The theoretical formula of the switching probability is derived from the Fokker-Planck equation. We find that the switching exponent is larger than one and depends on the applied electric current. The assumption that the switching exponent is one used in previous works is approximately valid only near the low current region, while relatively large current is applied in experiments to obtain the switching probability efficiently. In the high current region, the switching exponent is almost two. We also perform the numerical simulation of the Landau-Lifshitz-Gilbert (LLG) equation with stochastic torque. Figure 1(a) shows the dependence of the probability density, dP/dt, on the current pulse time for several temperatures. As shown, the probability density shows a peak at certain time. We call this time the switching time. Figure 1(b) shows the temperature dependence of the switching time on the temperature. It should be noted that the curvature of the switching time reflects the switching exponent [4]. The nonlinear dependence of the switching time is another evidence that the switching exponent is larger than one. The fact that the switching exponent is larger than one indicates that the previous works have underestimated the thermal stability. The presented formula enables us to evaluate the thermal stability with high accuracy. We also notice that the switching current at zero temperature, I c, can also be evaluated from the switching probability and our formula. We will talk the details of the theoretical analysis of the switching probability and numerical simulation. A comparison with latest experiments will also be presented. (Figure caption) Figure 1: (a) The dependence of the switching probability density on the current pulse time for several temperatures. (b) The temperature dependence of the switching time. [1] J. C. Slonczewski, J. Magn. Magn. Mater. 159, L1 (1996). [2] S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, and K. Ando, Nat. Mater. 3, 868 (2004). S. S. P. Parkin, C. Kaiser, A. Panchula, P. M. Rice, B. Hughes, M. Samant, and S.-H. Yang, ibid 3, 862 (2004). [3] R. H. Koch, J. A. Katine, and J. Z. Sun, Phys. Rev. Lett. 92, 088302 (2004). [4] T. Taniguchi and H. Imamura, J. Nanosci. Nanotechnol. 12, 7520 (2012). T. Taniguchi, M. Shibata, M. Marthaler, Y. Utsumi, and H. Imamura, Appl. Phys. Express 5, 063009 (2012). Figure 1

Our objective is to develop multi-functional nanotechnology-based anti-tumor drug delivery systems for improving the efficacy of treatments and reducing undesirable side effects. The essential part of this process is the development and validation of un-biased, quantitative analytical techniques. We employed a newly developed holographic imaging cytometry system HoloMonitor® M4 for label-free time-lapse cellular analysis (Phase Holographic Imaging, Sweden).In previous work we have demonstrated our ability to obtain quantitative, high content cellular feature data congruent to data obtained from traditional label-based systems. We applied a modified version of the Kolmogorov-Smirnov 2-sample test. The classic test takes control and test frequency distributions (histograms), and converts them to probability functions. The maximum vertical displacement between the two is reported as the D-value, to determine if the two distributions are significantly different. In modified versions of the test, it is a comparator; instead of a single value, we obtain histograms of the difference between the control and test normalized probability functions. These histograms are termed Brownian Bridges, where the end points are fixed at a value of zero, and the function is free to vary in between, in either up-going, down-going or mixed. Now, we report the development of a 4-sample Kolmogorov based method. Two modified K.S. tests are performed to compare distributions A to Ac and B to Bc. It is assumed that all the distributions have the same time basis, and each time point is processed sequentially. The resulting time point pairs are plotted as a vector in a two parameter scattergram, tracing the outlines of the probability distribution of the cell populations through time. Unlike, the Brownian Bridge histograms, here, the starting and the ending points are both at the same location (0,0), within the two dimensional scattergram, but are free to move in both positive and negative directions away from the origin. HeLa cells in a 12-well MatTek plate were adhered and treated with Paclitaxel (PCT) in concentrations from 0 to 40 nM in 5-fold increments, either pulsed for 4 hrs. followed by 48 hours of imaging or with continuous exposure for 48 hours of imaging. We obtained a variety of different plot types, including continuous vs. pulsed comparisons, inter dosage comparisons, as well as the ability to compare multiple features to each other. Of particular significance is the fact that in comparative dosage experiments over 10 nM, the pulsed and continuous probability zones are completely separated, reflecting live and dead cells, and allowing complex pharmacodynamic tracking. Citation Format: Ed Luther, Livia Mendes, Jiayi Pan, Elena Holden, Vladimir Torchilin. Development of a 4-sample version of the Kolmogorov-Smirnov test for evaluating the temporal physiology of cells treated with test compounds in a label-free, high content, platform for quantitative analysis of adherent cell-culture models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4553. doi:10.1158/1538-7445.AM2017-4553

Apomixis, the asexual reproduction via seeds, is associated to polyploidy and hybridization. To identify possible signatures of apomixis, and possible candidate genes underlying the shift from sex to apomixis, microarray-based gene expression patterns of live microdissected ovules at four different developmental stages were compared between apomictic and sexual individuals of the Ranunculus auricomus complex. Following predictions from previous work on mechanisms underlying apomixis penetrance and expressivity in the genus, gene expression patterns were classified into three categories based on their relative expression in apomicts compared to their sexual parental ancestors. We found evidence of misregulation and differential gene expression between apomicts and sexuals, with the highest number of differences detected during meiosis progression and emergence of aposporous initial (AI) cells, a key developmental stage in the ovule of apomicts where a decision between divergent reproductive pathways takes place. While most of the differentially expressed genes (DEGs) could not be annotated, gene expression was classified into transgressive, parent of origin and ploidy effects. Genes related to gametogenesis and meiosis demonstrated patterns reflective of transgressive and genome dosage effects, which support the hypothesis of a dominant factor controlling apomixis in Ranunculus and modulated by secondary modifiers. Three genes with probable functions in sporogenesis and gametogenesis development are identified and characterized for future studies.

New list of mosses in the Cusian woods (Orta Lake. Piedmont). — The author draws up a list of new entities and new places in the Cusian region, in addition to the entities already listed in his previous works, about wood mosses in the region, mentioned in the bibliography. The distinctive characters with reference to the known species and formes are emphasized. The boundaries of the considered region, the different kinds of woods as well as some geomorphologic, geologic and climatic notes have been explained in the previous works, together with a short account about the previous bryological researches that have been carried out in the region, from DE NOTARIS (1805-1877) onwards. New entities listed in the present and in the previous author's works: Pleuridium subulatum (Hedw.) Lindb. var. nurautianum; Oreoweisia Bruntoni (Smith) Milde var. cusiana; Anisothecium rufescens (Smith) Lindb. var. Vertacomorum; Dicranella subulata (Hedw.) Schimp, var. curvata (Hedw.) Rabenh. f. sculerensis; Gyroweisia t...

One of the important airline services that can influence the passengers’ loyalty is the in-flight meal service. In this study, the conceptual design process of new in-flight food delivery and waste collection system is carried out using the standard engineering design method to improve the current system. A critical step in this process is the design concept evaluation and selection where the best alternative design concept solution is chosen for further development. The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method is applied to facilitate the selection process of the best design concept for the new improved in-flight food delivery and waste collection system. The design evaluation criteria for the TOPSIS assessment procedures are taken from previous work done on design requirements analysis for the new in-flight food delivery and waste collection system. Similarly, five alternative design concepts for the new system are taken from the results of the previous research work done. Furthermore, an online public survey is conducted to acquire the assessment rating of all alternative design concepts for each design evaluation criterion. The assessment rating is assigned using a simple Likert rating scale. From the resultant TOPSIS ranking, Concept 3 has emerged as the best design concept with a closeness rating of 0.9589, which is very close to 1. For future research work, this selected final design concept for the new in-flight food delivery and waste collection system will be forwarded to the next engineering design stage for further development.

Recently, there has been significant work in the integration of probabilistic reasoning with first order logic representations. Learning algorithms for these models have been developed and they all considered modifications in the entire structure. In a previous work we argued that when the theory is approximately correct the use of techniques from theory revision to just modify the structure in places that failed in classification can be a more adequate choice. To score these modifications and choose the best one the log likelihood was used. However, this function was shown not to be well-suited in the propositional Bayesian classification task and instead the conditional log likelihood should be used. In the present paper, we extend this revision system showing the necessity of using specialization operators even when there are no negative examples. Moreover, the results of a theory modified only in places that are responsible for the misclassification of some examples are compared with the one that was modified in the entire structure using three databases and considering four probabilistic score functions, including conditional log likelihood.

Indoor localization systems based on existent radio communication networks often make use of received signal strength (RSS) as measured feature. In order to achieve a good accuracy such systems have a huge payload in the called calibration phase, where many labeled measurements are collected and used to build a representative feature map. This paper presents a new algorithm based on previous works from the same authors, where explicit calibration efforts are avoided by unsupervised online learning, while the system is already operational. Using probabilistic localization and non-parametric density estimation, the new approach uses unlabeled measurements to learn a feature map with the probabilistic mass function of the measurements, having as start only a rough initial model based on plausible physical properties. Simulations with artificial generated data and with real measurements validate the introduced algorithm, covering discontinuities on the feature map and multimodal distributions, imposed by a structured indoor environment.

Advanced micro-architectural features such as caches and branch prediction mechanisms supporting speculative execution are becoming commonplace within modern microprocessors. For developers of real-time systems, these mechanisms present predictability problems. Previous work has demonstrated accurate analysis for instruction caches, data caches, and branch prediction mechanisms are possible. However, the integration of these individual analysis methods is difficult to do without large increases in computational complexity or the introduction of pessimism regarding the worst-case execution time (WCET) estimate. In this paper, we discuss how a previously published analysis method for branch predictors can be integrated with instruction pipeline analysis.

Securing cloud-assisted Wireless Body Area Network (WBAN) environment by applying security mechanism that consumes less resources is still a challenging task. This research makes an attempt to address the same. One of the most prominent attacks in cloud-assisted WBAN is Distributed Denial of Service (DDoS) attack that not only disrupts the communication but also diminishes the network bandwidth and capacity. This work is an extension of our previous research work in which an Enhanced Very Fast Decision Tree (EVFDT) was proposed which could detect DDoS attack successfully. However, in our previous work, the proposed algorithm is evaluated on the dataset generated by implementing LEACH protocol in NS-2. In this paper, a real-time cloud-assisted WBAN test bed is deployed to investigate the efficiency and accuracy of proposed EVFDT algorithm for real-time sensor network traffic. To evaluate the performance of proposed algorithm on real-time WBAN, four metrics are used including classification accuracy, time, memory, and computational cost. It was observed that EVFDT outperforms the existing algorithms by maintaining better results for these metrics even in the presence of extreme noise. Experimental results show that the EVFDT algorithm attains significantly high detection accuracy with less false alarm rate.

Conventional parameter or tolerance designs focus on developing exact methods to minimize quality loss or manufacturing cost. The inherent assumption is that the response functions which represent the link between controllable variables and response values of quality characteristics are known before a design is developed. Moreover, parameter and tolerance values are assumed to be independent controllable variables in previous works; namely, they are determined separately in design activities. Currently, advanced computer software, such as computer-aided engineering, can help engineers to handle design problems with unknown response functions, at the stage of product design and process planning. Therefore, in this study, the software ANSYS was employed to obtain simulation data which represent the response values of quality characteristics. These response values will be used to fit a set of response functions for later analysis. However, previous works in computer simulation for design and planning usually lack consideration of the noise impact from an external design system. To approximate a realistic design environment, various levels of controllable variables, in conjunction with artificial noises created from uncontrollable variables, are used to generate simulated data for statistical analysis via Response Surface Methodology (RSM). Then, an optimization technique, such as mathematical programming, is adopted to integrate these response functions into one formulation so that optimal parameter and tolerance values are concurrently determined, with multiple quality characteristics taken into consideration. A bike-frame design was used to demonstrate the presented approach, followed by multiple quality characteristics of interest: material cost, bike-frame weight, structure reliability, and rigidity dependability. The goal is to minimize material cost and bike frame weight and to maximize structure reliability and rigidity dependability. This approach is useful for solving any complex design problems in the early stages, while providing enhanced functionality, quality, economic benefits, and a shorter design cycle.