We consider the problem of statically scheduling a task graph moldable streaming tasks (i.e., actor network) to multicore or many-core CPU with discrete dynamic voltage and frequency scaling (DVFS). employ an integer linear programming (ILP) approach that combines allocating cores tasks, mapping core subsets, selecting DVFS level for each task, considering all options fusion as provided by cost model, given data throughput latency requirements targeting low energy consumption. also propose...

We extend static scheduling of parallelizable tasks to machines with multiple core types, taking differences in performance and power consumption due task type into account. Next energy minimization for given deadline, i.e. throughput requirement, we consider makespan or average budgets. evaluate our approach by comparing schedules synthetic sets big.LITTLE other schedulers from literature. achieve an improvement up 33%.

A special type of social networks is the so-called affiliation network, consisting two modes vertices: actors and events. Up to now, in undirected case, closeness such has been measured by their jointly-attended Indirect contacts attenuated directed links are minor interest networks. These flaws make a veritable estimation of, e.g., possible message transfers amongst questionable. In this contribution, first, we discuss these matters from graph-theoretical point view. Second, so as avoid...

Tasks with stochastic runtimes and dependencies are frequently met in multicore applications, but static schedulers need deterministic task as input. We first demonstrate by scheduling experiments that both for binomially geometrically distributed runtimes, which often found taskgraphs, choice of average runtime scheduler input is sufficient to obtain schedules good makespan, i.e. inserting buffers depending on the standard deviation not helpful majority cases. Furthermore, we compute...

Summary Streaming task graphs are high‐level specifications for parallel applications operating on streams of data. For a static graph structure, schedulers can be used to map the tasks onto platform minimize energy consumption given throughput. We introduce dynamic elements into thus specifying which adapt behavior at runtime, example, switching from check‐only active mode. This in turn necessitates runtime system that remap and potentially their degree parallelism case change structure....

Abstract Design space exploration of a configurable, heterogeneous system for given application with required throughput searches combination cores or softcores different architectures that can be accommodated within the ASIC FPGA area and achieves optimizes power consumption. For soft real-time streaming application, modeled as task graph internally parallelizable tasks, this requires assigning core type quantity DVFS frequency level to each task, which implies runtime energy consumption,...

We introduce a novel approach for energy-efficient scheduling of data-dependent stream programs with packet types on multicore architectures voltage and frequency scaling. To have the given application meet specific throughput demands while minimizing energy consumption, we enhance existing crown by packet-type dependent parameters. Formulation as an integer linear program generates parametric, multi-scenario schedule table-driven execution. By inspecting at runtime, our online scheduler...

With the advent of Internet-of-Things, sensor data streams are widespread and thus a promising target for network steganography including covert channels. Therefore, we present an approach where sender receiver first build archive values that occur in stream over time interval, then encode bits secret message via belonging to class seen or not. If value’s category does not match required bit value, it is adapted switch its category. The chosen value minimizes distance original...

Static scheduling of independent, moldable tasks on parallel machines with frequency scaling comprises decisions core allocation, assignment, and ordering, to meet a deadline minimize energy consumption. Constraining some these reduces the solution space, i.e. may increase consumption, but also open path new, near-optimal approaches. We investigate how constraints different steps influence starting an unrestricted scheduler for tasks. The are partly derived from existing schedulers,...

Croivn scheduling is a static approach for sets of parallelizable tasks with common deadline, aiming to minimize energy consumption on parallel processors frequency scaling. We demonstrate that crown schedules are robust, i.e. the runtime prolongation one task by moderate percentage does not cause deadline transgression same fraction. In addition, speeding up some scheduled after prolonged task, can still be met at additional consumption. present heuristic perform this re-scaling online....

Crown scheduling is a static approach for sets of parallelizable tasks with common deadline, aiming to minimize energy consumption on parallel processors frequency scaling. We demonstrate that crown schedules are robust, i. e. the runtime prolongation one task by moderate percentage does not cause deadline transgression same fraction. In addition, speeding up some scheduled after prolonged task, can still be met at additional consumption. present heuristic perform this re-scaling online and...

Streaming task graphs are a high-level construct used to specify parallel applications operating on streams of data, and energy-efficient code meeting data throughput requirements can be generated automatically. Tasks themselves, so that two levels parallelism present. Normally, the graph structure is assumed static, static schedulers map (parallel) tasks onto platform minimize energy consumption for given throughput. We propose formulation introduces dynamic elements into structure, thus...

Hash-based message authentication code (HMAC) is a form of cryptographic checksum which frequently used in SSL/TLS. The IETF standard permits several hash function families with possible lengths. Thus, if checksums must be computed for number messages, multi-dimensional optimization problem arises, especially heterogeneous platform like ARM big.LITTLE used. We formulate the as an energy minimization constraints given by deadlines and different power performance profiles functions on core...

Efficiency is the main issue in any data envelopment analysis. Realizing output by a minimum of input or reaching maximum given credo, scale effects often are only sort accessory. In modern economics play prominent role, however. What right size decision making unit (DMU) and how to proceed there. Returns inform DMU about its hitherto disregard show way towards ideal activity size. Combining efficiency aspects scaling leads new DEA-best practice criterion DMUs gives them profound orientation...

Higher-radix floating-point representations have the potential for higher performance, lower energy footprint, and reduced gate count in embedded systems when compared to traditional binary numbers. Thus, they might also appear transmission of sensor data values. However, these number formats introduce redundancies, which can be exploited steganographic message transfer. We present a covert channel that exploits this redundancy trade bandwidth against introduced error thus detectability. In...

Abstract Terrorism is a scourge of humanity. Thousands people were killed by terrorists in the last years. To understand structure, information flow and leadership terrorist cells must for all societies to prevent future attacks. Network theory can help model such identify leading members therein. This contribution uses graph theoretical approach as basis enriches it an entropy‐driven knowledge processing. After presenting concept two cells, one Bali attack 9/11 attack, are analysed. The...

A social network (SN) is a group of actors and their mutual relations. Sociologists try to answer the question why networked in our society are more successful than others how this networking works. Directed or undirected graphs, hyper- multigraphs suitable means visualize Social networks with directed weighted links among need sophisticated instruments for analyses. We model these as probabilistic conditioned propositions. Then any actor i permits estimation transfer probabilities all j,...

For fixed-application scenarios in embedded soft-realtime computing, the ideal (w.r.t. energy consumption) heterogeneous multi-core CPU design within given chip dimensions can be configured by composing it from pre-layouted, rectangular submodules for each of a number <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">K</i> > 1 core types, where practice is small constant. example, =2 traditional ARM big.LITTLE designs. Nevertheless, even better...