A real coded genetic algorithm for data partitioning and scheduling in networks with arbitrary processor release time

The problem of scheduling divisible loads in distributed computing systems, in presence of processor release time is considered. The objective is to find the optimal sequence of load distribution and the optimal load fractions assigned to each processor in the system such that the processing time of the entire processing load is a minimum. This is a difficult combinatorial optimization problem and hence genetic algorithms approach is presented for its solution.

An Accurate Model for EESM and its Application to Analysis of CQI Feedback Schemes and Scheduling in LTE

The Effective Exponential SNR Mapping (EESM) is an indispensable tool for analyzing and simulating next generation orthogonal frequency division multiplexing (OFDM) based wireless systems. It converts the different gains of multiple subchannels, over which a codeword is transmitted, into a single effective flat-fading gain with the same codeword error rate. It facilitates link adaptation by helping each user to compute an accurate channel quality indicator (CQI), which is fed back to the base station to enable downlink rate adaptation and scheduling. However, the highly non-linear nature of EESM makes a performance analysis of adaptation and scheduling difficult; even the probability distribution of EESM is not known in closed-form. This paper shows that EESM can be accurately modeled as a lognormal random variable when the subchannel gains are Rayleigh distributed. The model is also valid when the subchannel gains are correlated in frequency or space. With some simplifying assumptions, the paper then develops a novel analysis of the performance of LTE's two CQI feedback schemes that use EESM to generate CQI. The comprehensive model and analysis quantify the joint effect of several critical components such as scheduler, multiple antenna mode, CQI feedback scheme, and EESM-based feedback averaging on the overall system throughput.

Joint evaluation of channel feedback schemes, rate adaptation, and scheduling in OFDMA downlinks with feedback delays

Orthogonal frequency-division multiple access (OFDMA) systems divide the available bandwidth into orthogonal subchannels and exploit multiuser diversity and frequency selectivity to achieve high spectral efficiencies. However, they require a significant amount of channel state feedback for scheduling and rate adaptation and are sensitive to feedback delays. We develop a comprehensive analysis for OFDMA system throughput in the presence of feedback delays as a function of the feedback scheme, frequency-domain scheduler, and rate adaptation rule. Also derived are expressions for the outage probability, which captures the inability of a subchannel to successfully carry data due to the feedback scheme or feedback delays. Our model encompasses the popular best-n and threshold-based feedback schemes and the greedy, proportional fair, and round-robin schedulers that cover a wide range of throughput versus fairness tradeoff. It helps quantify the different robustness of the schedulers to feedback overhead and delays. Even at low vehicular speeds, it shows that small feedback delays markedly degrade the throughput and increase the outage probability. Further, given the feedback delay, the throughput degradation depends primarily on the feedback overhead and not on the feedback scheme itself. We also show how to optimize the rate adaptation thresholds as a function of feedback delay.

Low Complexity Power and Scheduling Policies for Wireless Networks to Provide Quality of Service

We consider optimal power allocation policies for a single server, multiuser system. The power is consumed in transmission of data only. The transmission channel may experience multipath fading. We obtain very efficient, low computational complexity algorithms which minimize power and ensure stability of the data queues. We also obtain policies when the users may have mean delay constraints. If the power required is a linear function of rate then we exploit linearity and obtain linear programs with low complexity.

Joint Power Control, Scheduling and Routing for Multihop Energy Harvesting Sensor Networks

We study wireless multihop energy harvesting sensor networks employed for random field estimation. The sensors sense the random field and generate data that is to be sent to a fusion node for estimation. Each sensor has an energy harvesting source and can operate in two modes: Wake and Sleep. We consider the problem of obtaining jointly optimal power control, routing and scheduling policies that ensure a fair utilization of network resources. This problem has a high computational complexity. Therefore, we develop a computationally efficient suboptimal approach to obtain good solutions to this problem. We study the optimal solution and performance of the suboptimal approach through some numerical examples.

Register Allocation and Optimal Spill code Scheduling in Software Pipelined Loops using 0-1 Integer Linear Programming Formulation

In achieving higher instruction level parallelism, software pipelining increases the register pressure in the loop. The usefulness of the generated schedule may be restricted to cases where the register pressure is less than the available number of registers. Spill instructions need to be introduced otherwise. But scheduling these spill instructions in the compact schedule is a difficult task. Several heuristics have been proposed to schedule spill code. These heuristics may generate more spill code than necessary, and scheduling them may necessitate increasing the initiation interval. We model the problem of register allocation with spill code generation and scheduling in software pipelined loops as a 0-1 integer linear program. The formulation minimizes the increase in initiation interval (II) by optimally placing spill code and simultaneously minimizes the amount of spill code produced. To the best of our knowledge, this is the first integrated formulation for register allocation, optimal spill code generation and scheduling for software pipelined loops. The proposed formulation performs better than the existing heuristics by preventing an increase in II in 11.11% of the loops and generating 18.48% less spill code on average among the loops extracted from Perfect Club and SPEC benchmarks with a moderate increase in compilation time.

Wireless scheduling with partial channel state information: large deviations and optimality

We consider a server serving a time-slotted queued system of multiple packet-based flows, where not more than one flow can be serviced in a single time slot. The flows have exogenous packet arrivals and time-varying service rates. At each time, the server can observe instantaneous service rates for only a subset of flows ( selected from a fixed collection of observable subsets) before scheduling a flow in the subset for service. We are interested in queue length aware scheduling to keep the queues short. The limited availability of instantaneous service rate information requires the scheduler to make a careful choice of which subset of service rates to sample. We develop scheduling algorithms that use only partial service rate information from subsets of channels, and that minimize the likelihood of queue overflow in the system. Specifically, we present a new joint subset-sampling and scheduling algorithm called Max-Exp that uses only the current queue lengths to pick a subset of flows, and subsequently schedules a flow using the Exponential rule. When the collection of observable subsets is disjoint, we show that Max-Exp achieves the best exponential decay rate, among all scheduling algorithms that base their decision on the current ( or any finite past history of) system state, of the tail of the longest queue. To accomplish this, we employ novel analytical techniques for studying the performance of scheduling algorithms using partial state, which may be of independent interest. These include new sample-path large deviations results for processes obtained by non-random, predictable sampling of sequences of independent and identically distributed random variables. A consequence of these results is that scheduling with partial state information yields a rate function significantly different from scheduling with full channel information. In the special case when the observable subsets are singleton flows, i.e., when there is effectively no a priori channel state information, Max-Exp reduces to simply serving the flow with the longest queue; thus, our results show that to always serve the longest queue in the absence of any channel state information is large deviations optimal.

Experiences With WM: A Centralised Scheduling Approach for Performance Management of IEEE 802.11 Wireless LANs

In our earlier work ([1]) we proposed WLAN Manager (or WM) a centralised controller for QoS management of infrastructure WLANs based on the IEEE 802.11 DCF standards. The WM approach is based on queueing and scheduling packets in a device that sits between all traffic flowing between the APs and the wireline LAN, requires no changes to the AP or the STAs, and can be viewed as implementing a "Split-MAC" architecture. The objectives of WM were to manage various TCP performance related issues (such as the throughput "anomaly" when STAs associate with an AP with mixed PHY rates, and upload-download unfairness induced by finite AP buffers), and also to serve as the controller for VoIP admission control and handovers, and for other QoS management measures. In this paper we report our experiences in implementing the proposals in [1]: the insights gained, new control techniques developed, and the effectiveness of the WM approach in managing TCP performance in an infrastructure WLAN. We report results from a hybrid experiment where a physical WM manages actual TCP controlled packet flows between a server and clients, with the WLAN being simulated, and also from a small physical testbed with an actual AP.

Providing QoS to Real and Data Applications in WiMAX Mesh Networks

We consider the problem of centralized routing and scheduling for IEEE 802.16 mesh networks so as to provide Quality of Service (QoS) to individual real and interactive data applications. We first obtain an optimal and fair routing and scheduling policy for aggregate demands for different source- destination pairs. We then present scheduling algorithms which provide per flow QoS guarantees while utilizing the network resources efficiently. Our algorithms are also scalable: they do not require per flow processing and queueing and the computational requirements are modest. We have verified our algorithms via extensive simulations.

Experiences with a centralized scheduling approach for performance management of IEEE 802.11 wireless LANs

We present a centralized integrated approach for: 1) enhancing the performance of an IEEE 802.11 infrastructure wireless local area network (WLAN), and 2) managing the access link that connects the WLAN to the Internet. Our approach, which is implemented on a standard Linux platform, and which we call ADvanced Wi-fi Internet Service EnhanceR (ADWISER), is an extension of our previous system WLAN Manager (WM). ADWISER addresses several infrastructure WLAN performance anomalies such as mixed-rate inefficiency, unfair medium sharing between simultaneous TCP uploads and downloads, and inefficient utilization of the Internet access bandwidth when Internet transfers compete with LAN-WLAN transfers, etc. The approach is via centralized queueing and scheduling, using a novel, configurable, cascaded packet queueing and scheduling architecture, with an adaptive service rate. In this paper, we describe the design of ADWISER and report results of extensive experimentation conducted on a hybrid testbed consisting of real end-systems and an emulated WLAN on Qualnet. We also present results from a physical testbed consisting of one access point (AP) and a few end-systems.

Automatic Data Allocation and Buffer Management for Multi-GPU Machines

Multi-GPU machines are being increasingly used in high-performance computing. Each GPU in such a machine has its own memory and does not share the address space either with the host CPU or other GPUs. Hence, applications utilizing multiple GPUs have to manually allocate and manage data on each GPU. Existing works that propose to automate data allocations for GPUs have limitations and inefficiencies in terms of allocation sizes, exploiting reuse, transfer costs, and scalability. We propose a scalable and fully automatic data allocation and buffer management scheme for affine loop nests on multi-GPU machines. We call it the Bounding-Box-based Memory Manager (BBMM). BBMM can perform at runtime, during standard set operations like union, intersection, and difference, finding subset and superset relations on hyperrectangular regions of array data (bounding boxes). It uses these operations along with some compiler assistance to identify, allocate, and manage data required by applications in terms of disjoint bounding boxes. This allows it to (1) allocate exactly or nearly as much data as is required by computations running on each GPU, (2) efficiently track buffer allocations and hence maximize data reuse across tiles and minimize data transfer overhead, and (3) and as a result, maximize utilization of the combined memory on multi-GPU machines. BBMM can work with any choice of parallelizing transformations, computation placement, and scheduling schemes, whether static or dynamic. Experiments run on a four-GPU machine with various scientific programs showed that BBMM reduces data allocations on each GPU by up to 75% compared to current allocation schemes, yields performance of at least 88% of manually written code, and allows excellent weak scaling.

6PANview: Application performance conscious network monitoring for 6LoWPAN based WSNs

6PANview[1] is a Wireless Sensor Network(WSN) monitoring system for 6LoWPAN/RPL networks which we developed as an overlay network for a WSN application. A monitoring system, while performing its operations for maintaining the health of the monitored network, must also be conscious of its impact on the application performance, and must strive to minimize this impact. To this end, we propose a centralized scheduling algorithm within 6PANview which non-intrusively analyzes application traffic arrival patterns at the base station, identifies network idle periods and schedules monitoring activities. The proposed algorithm finds those periodic sequences which are likely to have given rise to the pattern of arrivals seen at the base station. Parts of those sequences are then extended to coarsely predict future traffic and find epochs where low traffic is predicted, in order to schedule monitoring traffic or other activities at these times. We present simulation results for the proposed prediction and scheduling algorithm and its implementation as part of 6PANview. As an enhancement, we briefly talk about using 6PANview's overlay network architecture for distributed scheduling.

Using formal reasoning on a model of tasks for FreeRTOS

FreeRTOS is an open-source real-time microkernel that has a wide community of users. We present the formal specification of the behaviour of the task part of FreeRTOS that deals with the creation, management, and scheduling of tasks using priority-based preemption. Our model is written in the Z notation, and we verify its consistency using the Z/Eves theorem prover. This includes a precise statement of the preconditions for all API commands. This task model forms the basis for three dimensions of further work: (a) the modelling of the rest of the behaviour of queues, time, mutex, and interrupts in FreeRTOS; (b) refinement of the models to code to produce a verified implementation; and (c) extension of the behaviour of FreeRTOS to multi-core architectures. We propose all three dimensions as benchmark challenge problems for Hoare's Verified Software Initiative.

Adaptive executions of hyperbolic block-structured AMR applications on GPU systems

A block-structured adaptive mesh refinement (AMR) technique has been used to obtain numerical solutions for many scientific applications. Some block-structured AMR approaches have focused on forming patches of non-uniform sizes where the size of a patch can be tuned to the geometry of a region of interest. In this paper, we develop strategies for adaptive execution of block-structured AMR applications on GPUs, for hyperbolic directionally split solvers. While effective hybrid execution strategies exist for applications with uniform patches, our work considers efficient execution of non-uniform patches with different workloads. Our techniques include bin-packing work units to load balance GPU computations, adaptive asynchronism between CPU and GPU executions using a knapsack formulation, and scheduling communications for multi-GPU executions. Our experiments with synthetic and real data, for single-GPU and multi-GPU executions, on Tesla S1070 and Fermi C2070 clusters, show that our strategies result in up to a 3.23 speedup in performance over existing strategies.