Results 1 
9 of
9
BestEffort Parallel Execution Framework for Recognition and Mining Applications
"... Recognition and mining (RM) applications are an emerging class of computing workloads that will be commonly executed on future multicore and manycore computing platforms. The explosive growth of input data and the use of more sophisticated algorithms in RM applications will ensure, for the foresee ..."
Abstract

Cited by 12 (1 self)
 Add to MetaCart
(Show Context)
Recognition and mining (RM) applications are an emerging class of computing workloads that will be commonly executed on future multicore and manycore computing platforms. The explosive growth of input data and the use of more sophisticated algorithms in RM applications will ensure, for the foreseeable future, a significant gap between the computational needs of RM applications and the capabilities of rapidly evolving multi or manycore platforms. To address this gap, we propose a new parallel programming model that inherently embodies the notion of besteffort computing, wherein the underlying parallel computing environment is not expected to be perfect. The proposed besteffort computing model leverages three key characteristics of RM applications: (1) the input data is noisy and it often contains significant redundancy, (2) computations performed on the input data are statistical in nature, and (3) some degree of imprecision in the output is acceptable. As a specific instance of a besteffort parallel programming model, we describe an “iterativeconvergence” parallel template which is used by a significant class of RM applications. We show how the besteffort computing template can be used to not only reduce computational workload, but to also eliminate dependencies between computations and further increase parallelism. Our experiments on an 8core machine demonstrate a speedup of 3.5X and 4.3X for the Kmeans and GLVQ algorithms, respectively, over a conventional parallel implementation. We also show that there is almost no material impact on the accuracy of results obtained from besteffort implementations in the application context of image segmentation using Kmeans and eye detection in images using GLVQ. 1
Lockfree synchronization for dynamic embedded realtime software
 In ACM/IEEE Design, Automation, and Test in Europe (DATE
, 2006
"... We consider nonblocking synchronization for dynamic embedded realtime systems such as those that are subject to resource overloads and arbitrary activity arrivals. The multiwriter/multireader problem inherently occurs in such systems, when their activities must concurrently and mutually exclusiv ..."
Abstract

Cited by 10 (2 self)
 Add to MetaCart
(Show Context)
We consider nonblocking synchronization for dynamic embedded realtime systems such as those that are subject to resource overloads and arbitrary activity arrivals. The multiwriter/multireader problem inherently occurs in such systems, when their activities must concurrently and mutually exclusive share data objects. We consider lockfree synchronization for this problem under the unimodal arbitrary arrival model (or UAM). UAM embodies a “stronger” adversary than most traditional arrival models. We establish the fundamental tradeoffs between lockfree and lockbased object sharing under UAM — the first such result. Our tradeoffs include analytical conditions under which activities ’ accrued timeliness utility is greater under lockfree than lockbased, and the consequent upper bound on the increase in accrued utility. Our implementation experience on a POSIX RTOS reveals that the lockfree scheduling algorithm yields higher accrued utility, by as much as 65%, and critical time satisfactions, by as much as 80%, over lockbased. 1.
On multiprocessor utility accrual realtime scheduling with statistical timing assurances
 In IFIP Embedded and Ubiquitous Computing (EUC
, 2006
"... We present the first Utility Accrual (or UA) realtime scheduling algorithm for multiprocessors, called gMUA. The algorithm considers an application model where realtime activities are subject to time/utility function time constraints, variable execution time demands, and resource overloads where t ..."
Abstract

Cited by 9 (1 self)
 Add to MetaCart
(Show Context)
We present the first Utility Accrual (or UA) realtime scheduling algorithm for multiprocessors, called gMUA. The algorithm considers an application model where realtime activities are subject to time/utility function time constraints, variable execution time demands, and resource overloads where the total activity utilization demand exceeds the total capacity of all processors. We consider the scheduling objective of (1) probabilistically satisfying lower bounds on each activity’s maximum utility and (2) maximizing the systemwide, total accrued utility. We establish several properties of gMUA including optimal total utility (for a special case), conditions under which individual activity utility lower bounds are satisfied, a lower bound on systemwide total accrued utility, and bounded sensitivity for assurances to variations in execution time demand estimates. Our simulation experiments validate our analytical results and confirm the algorithm’s effectiveness and superiority.
On scheduling garbage collector in dynamic realtime systems with statistical timing assurances
 In IEEE ISORC
, 2006
"... We consider garbage collection (GC) in dynamic realtime systems. We consider the timebased GC approach of running the collector as a separate, concurrent thread, and focus on realtime scheduling to obtain assurances on mutator timing behavior, while ensuring that memory is never exhausted. We pre ..."
Abstract

Cited by 5 (1 self)
 Add to MetaCart
(Show Context)
We consider garbage collection (GC) in dynamic realtime systems. We consider the timebased GC approach of running the collector as a separate, concurrent thread, and focus on realtime scheduling to obtain assurances on mutator timing behavior, while ensuring that memory is never exhausted. We present a scheduling algorithm called GCUA. The algorithm considers mutator activities that are subject to time/utility function time constraints, variable execution time demands, the unimodal arbitrary arrival model that allows a strong adversary, and resource overloads. We establish several properties of GCUA including probabilisticallysatisfied utility lower bounds for each mutator activity, a lower bound on the systemwide total accrued utility, bounded sensitivity for the assurances to variations in mutator execution time demand estimates, and no memory exhaustion at all times. Our simulation experiments validate our analytical results and confirm the algorithm’s effectiveness and superiority. I.
Utility Accrual RealTime Scheduling and Synchronization on Single and Multiprocessors: Models, Algorithms, and Tradeoffs
, 2006
"... This dissertation presents a class of utility accrual scheduling and synchronization algorithms for dynamic, single and multiprocessor realtime systems. Dynamic realtime systems operate in environments with runtime uncertainties including those on activity execution times and arrival behaviors. ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
(Show Context)
This dissertation presents a class of utility accrual scheduling and synchronization algorithms for dynamic, single and multiprocessor realtime systems. Dynamic realtime systems operate in environments with runtime uncertainties including those on activity execution times and arrival behaviors. We consider the time/utility function (or TUF) timing model for specifying application time constraints, and the utility accrual (or UA) timeliness optimality criteria of satisfying lower bounds on accrued activity utility, and maximizing the total accrued utility. Efficient TUF/UA scheduling algorithms exist for single processors—e.g., the Resourceconstrained Utility Accrual scheduling algorithm (RUA), and the Dependent Activity Scheduling Algorithm (DASA). However, they all use lockbased synchronization. To overcome shortcomings of lockbased (e.g., serialized object access, increased runtime overhead, deadlocks), we consider nonblocking synchronization including waitfree and lockfree synchronization. We present a bufferoptimal, schedulerindependent waitfree synchronization protocol (the first such), and develop waitfree versions of RUA and DASA. We also develop their lockfree versions, and upper bound their retries under the unimodal arbitrary arrival model. The tradeoff between waitfree, lockfree, and lockbased is fundamentally about their space and
On BestEffort Utility Accrual RealTime Scheduling on Multiprocessors
, 2010
"... We consider the problem of scheduling realtime tasks on a multiprocessor system. Our primary focus is scheduling on multiprocessor systems where the total task utilization demand, U, is greater than m, the number of processors on a multiprocessor systemi.e., the total available processing capacity ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
We consider the problem of scheduling realtime tasks on a multiprocessor system. Our primary focus is scheduling on multiprocessor systems where the total task utilization demand, U, is greater than m, the number of processors on a multiprocessor systemi.e., the total available processing capacity of the system. When U> m, the system is said to be overloaded; otherwise, the system is said to be underloaded. While significant literature exists on multiprocessor realtime scheduling during underloads, little is known about scheduling during overloads, in particular, in the presence of task dependenciese.g., due to synchronization constraints. We consider realtime tasks that are subject to time/utility function (or TUF) time constraints, which allow task urgency to be expressed independently of task importancee.g., the most urgent task being the least important. The urgency/importance decoupling allowed by TUFs is especially important during overloads, when not all tasks can be optimally completed. We consider the timeliness optimization objective of maximizing the total accrued utility and the number of deadlines satis ed during overloads, while ensuring task mutual exclusion constraints and freedom from deadlocks. This problem is NPhard. We develop a class of polynomialtime heuristic
FixedPriority Scheduling Algorithms with Multiple Objectives in Hard Realtime Systems
, 2006
"... In the context of FixedPriority Scheduling in RealTime Systems, we investigate scheduling mechanisms for supporting systems where, in addition to timing constraints, their performance with respect to additional QoS requirements must be improved. This type of situation may occur when the worstcase ..."
Abstract
 Add to MetaCart
In the context of FixedPriority Scheduling in RealTime Systems, we investigate scheduling mechanisms for supporting systems where, in addition to timing constraints, their performance with respect to additional QoS requirements must be improved. This type of situation may occur when the worstcase resource requirements of all or some running tasks cannot be simultaneously met due to task contention. Solutions to these problems have been proposed in the context of both fixedpriority and dynamicpriority scheduling. In fixedpriority scheduling, the typical approach is to artificially modify the attributes or structure of tasks, and/or usually require nonstandard runtime support. In dynamicpriority scheduling approaches, utility functions are employed to make scheduling decisions with the objective of maximising the utility. The main difficulties with these approaches are the inability to formulate and model appropriately utility functions for each task, and the inability to guarantee hard deadlines without executing computationally costly algorithms. In this thesis we propose a different approach. Firstly, we introduce the concept of
Noname manuscript No. (will be inserted by the editor) On Scheduling Garbage Collector in Dynamic RealTime Systems With Statistical Timing Assurances
"... Abstract We consider garbage collection (GC) in dynamic realtime systems. We consider the timebased GC approach of running the collector as a separate, concurrent thread, and focus on realtime scheduling to obtain assurances on mutator timing behavior, while ensuring that memory is never exhaus ..."
Abstract
 Add to MetaCart
(Show Context)
Abstract We consider garbage collection (GC) in dynamic realtime systems. We consider the timebased GC approach of running the collector as a separate, concurrent thread, and focus on realtime scheduling to obtain assurances on mutator timing behavior, while ensuring that memory is never exhausted. We present a scheduling algorithm called GCUA. The algorithm considers mutator activities that are subject to time/utility function time constraints, variable execution time demands, the unimodal arbitrary arrival model that allows a strong adversary, and resource overloads. We establish several properties of GCUA including probabilisticallysatisfied utility lower bounds for each mutator activity, a lower bound on the systemwide total accrued utility, bounded sensitivity for the assurances to variations in mutator execution time demand estimates, and no memory exhaustion at all times. Our simulation experiments validate our analytical results and confirm the algorithm’s effectiveness and superiority. 1
A RealTime Task Scheduling Algorithm Based on Dynamic Priority
"... By studying the dynamic value density and urgency of a task, a preemptive scheduling strategy based on dynamic priority assignment is proposed. In the strategy, two parameters p and q are used to adjust the weight that the value density and urgency of a task impact on its priority, and a parameter ..."
Abstract
 Add to MetaCart
(Show Context)
By studying the dynamic value density and urgency of a task, a preemptive scheduling strategy based on dynamic priority assignment is proposed. In the strategy, two parameters p and q are used to adjust the weight that the value density and urgency of a task impact on its priority, and a parameter β is used to avoid the possible system thrashing. Finally, the simulations show that our algorithm is prior to the analogous algorithms, such as EDF, HVF and HVDF, on gainedvalue of the system, deadline miss ratio and preemptive number. 1