The Experts below are selected from a list of 14826 Experts worldwide ranked by ideXlab platform
Alexey Lastovetsky - One of the best experts on this subject based on the ideXlab platform.
-
bi objective optimization of data parallel applications on heterogeneous hpc platforms for performance and energy through Workload Distribution
IEEE Transactions on Parallel and Distributed Systems, 2021Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Arsalan Shahid, Alexey LastovetskyAbstract:Performance and energy are the two most important objectives for optimization on modern parallel platforms. In this article, we show that moving from single-objective optimization for performance or energy to their bi-objective optimization on heterogeneous processors results in a tremendous increase in the number of optimal solutions (Workload Distributions) even for the simple case of linear performance and energy profiles. We then study full performance and energy profiles of two real-life data-parallel applications and find that they exhibit shapes that are non-linear and complex enough to prevent good approximation of them as analytical functions for input to exact algorithms or optimization software for determining the Pareto front. We, therefore, propose a solution method solving the bi-objective optimization problem on heterogeneous processors. The method's novel component is an efficient and exact global optimization algorithm that takes as an input performance and energy profiles as arbitrary discrete functions of Workload size, which accurately and realistically take into account resource contention and NUMA inherent in modern parallel platforms, and returns the Pareto-optimal solutions (generally speaking, load imbalanced). To construct the input discrete energy functions, the method employs a methodology that accurately models the energy consumption by a hybrid data-parallel application executing on a heterogeneous HPC platform containing different computing devices using system-level power measurements provided by power meters. We experimentally analyse the proposed solution method using three data-parallel applications, matrix multiplication, 2D fast Fourier transform (2D-FFT), and gene sequencing, on two connected heterogeneous servers consisting of multicore CPUs, GPUs, and Intel Xeon Phi. We show that it determines a superior Pareto front containing the best load balanced solutions and all the load imbalanced solutions that are ignored by load balancing methods.
-
optimization of data parallel applications on heterogeneous hpc platforms for dynamic energy through Workload Distribution
European Conference on Parallel Processing, 2019Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Alexey LastovetskyAbstract:Energy is one of the most important objectives for optimization on modern heterogeneous high performance computing (HPC) platforms. The tight integration of multicore CPUs with accelerators in these platforms present several challenges to optimization of multithreaded data-parallel applications for dynamic energy.
Hamidreza Khaleghzadeh - One of the best experts on this subject based on the ideXlab platform.
-
bi objective optimization of data parallel applications on heterogeneous hpc platforms for performance and energy through Workload Distribution
IEEE Transactions on Parallel and Distributed Systems, 2021Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Arsalan Shahid, Alexey LastovetskyAbstract:Performance and energy are the two most important objectives for optimization on modern parallel platforms. In this article, we show that moving from single-objective optimization for performance or energy to their bi-objective optimization on heterogeneous processors results in a tremendous increase in the number of optimal solutions (Workload Distributions) even for the simple case of linear performance and energy profiles. We then study full performance and energy profiles of two real-life data-parallel applications and find that they exhibit shapes that are non-linear and complex enough to prevent good approximation of them as analytical functions for input to exact algorithms or optimization software for determining the Pareto front. We, therefore, propose a solution method solving the bi-objective optimization problem on heterogeneous processors. The method's novel component is an efficient and exact global optimization algorithm that takes as an input performance and energy profiles as arbitrary discrete functions of Workload size, which accurately and realistically take into account resource contention and NUMA inherent in modern parallel platforms, and returns the Pareto-optimal solutions (generally speaking, load imbalanced). To construct the input discrete energy functions, the method employs a methodology that accurately models the energy consumption by a hybrid data-parallel application executing on a heterogeneous HPC platform containing different computing devices using system-level power measurements provided by power meters. We experimentally analyse the proposed solution method using three data-parallel applications, matrix multiplication, 2D fast Fourier transform (2D-FFT), and gene sequencing, on two connected heterogeneous servers consisting of multicore CPUs, GPUs, and Intel Xeon Phi. We show that it determines a superior Pareto front containing the best load balanced solutions and all the load imbalanced solutions that are ignored by load balancing methods.
-
optimization of data parallel applications on heterogeneous hpc platforms for dynamic energy through Workload Distribution
European Conference on Parallel Processing, 2019Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Alexey LastovetskyAbstract:Energy is one of the most important objectives for optimization on modern heterogeneous high performance computing (HPC) platforms. The tight integration of multicore CPUs with accelerators in these platforms present several challenges to optimization of multithreaded data-parallel applications for dynamic energy.
Muhammad Fahad - One of the best experts on this subject based on the ideXlab platform.
-
bi objective optimization of data parallel applications on heterogeneous hpc platforms for performance and energy through Workload Distribution
IEEE Transactions on Parallel and Distributed Systems, 2021Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Arsalan Shahid, Alexey LastovetskyAbstract:Performance and energy are the two most important objectives for optimization on modern parallel platforms. In this article, we show that moving from single-objective optimization for performance or energy to their bi-objective optimization on heterogeneous processors results in a tremendous increase in the number of optimal solutions (Workload Distributions) even for the simple case of linear performance and energy profiles. We then study full performance and energy profiles of two real-life data-parallel applications and find that they exhibit shapes that are non-linear and complex enough to prevent good approximation of them as analytical functions for input to exact algorithms or optimization software for determining the Pareto front. We, therefore, propose a solution method solving the bi-objective optimization problem on heterogeneous processors. The method's novel component is an efficient and exact global optimization algorithm that takes as an input performance and energy profiles as arbitrary discrete functions of Workload size, which accurately and realistically take into account resource contention and NUMA inherent in modern parallel platforms, and returns the Pareto-optimal solutions (generally speaking, load imbalanced). To construct the input discrete energy functions, the method employs a methodology that accurately models the energy consumption by a hybrid data-parallel application executing on a heterogeneous HPC platform containing different computing devices using system-level power measurements provided by power meters. We experimentally analyse the proposed solution method using three data-parallel applications, matrix multiplication, 2D fast Fourier transform (2D-FFT), and gene sequencing, on two connected heterogeneous servers consisting of multicore CPUs, GPUs, and Intel Xeon Phi. We show that it determines a superior Pareto front containing the best load balanced solutions and all the load imbalanced solutions that are ignored by load balancing methods.
-
optimization of data parallel applications on heterogeneous hpc platforms for dynamic energy through Workload Distribution
European Conference on Parallel Processing, 2019Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Alexey LastovetskyAbstract:Energy is one of the most important objectives for optimization on modern heterogeneous high performance computing (HPC) platforms. The tight integration of multicore CPUs with accelerators in these platforms present several challenges to optimization of multithreaded data-parallel applications for dynamic energy.
Ravi Reddy Manumachu - One of the best experts on this subject based on the ideXlab platform.
-
bi objective optimization of data parallel applications on heterogeneous hpc platforms for performance and energy through Workload Distribution
IEEE Transactions on Parallel and Distributed Systems, 2021Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Arsalan Shahid, Alexey LastovetskyAbstract:Performance and energy are the two most important objectives for optimization on modern parallel platforms. In this article, we show that moving from single-objective optimization for performance or energy to their bi-objective optimization on heterogeneous processors results in a tremendous increase in the number of optimal solutions (Workload Distributions) even for the simple case of linear performance and energy profiles. We then study full performance and energy profiles of two real-life data-parallel applications and find that they exhibit shapes that are non-linear and complex enough to prevent good approximation of them as analytical functions for input to exact algorithms or optimization software for determining the Pareto front. We, therefore, propose a solution method solving the bi-objective optimization problem on heterogeneous processors. The method's novel component is an efficient and exact global optimization algorithm that takes as an input performance and energy profiles as arbitrary discrete functions of Workload size, which accurately and realistically take into account resource contention and NUMA inherent in modern parallel platforms, and returns the Pareto-optimal solutions (generally speaking, load imbalanced). To construct the input discrete energy functions, the method employs a methodology that accurately models the energy consumption by a hybrid data-parallel application executing on a heterogeneous HPC platform containing different computing devices using system-level power measurements provided by power meters. We experimentally analyse the proposed solution method using three data-parallel applications, matrix multiplication, 2D fast Fourier transform (2D-FFT), and gene sequencing, on two connected heterogeneous servers consisting of multicore CPUs, GPUs, and Intel Xeon Phi. We show that it determines a superior Pareto front containing the best load balanced solutions and all the load imbalanced solutions that are ignored by load balancing methods.
-
optimization of data parallel applications on heterogeneous hpc platforms for dynamic energy through Workload Distribution
European Conference on Parallel Processing, 2019Co-Authors: Hamidreza Khaleghzadeh, Muhammad Fahad, Ravi Reddy Manumachu, Alexey LastovetskyAbstract:Energy is one of the most important objectives for optimization on modern heterogeneous high performance computing (HPC) platforms. The tight integration of multicore CPUs with accelerators in these platforms present several challenges to optimization of multithreaded data-parallel applications for dynamic energy.
Richard J Linn - One of the best experts on this subject based on the ideXlab platform.
-
a heuristic for dynamic yard crane deployment in a container terminal
Iie Transactions, 2003Co-Authors: Richard J Linn, Chuqian ZhangAbstract:Rubber Tired Gantry Cranes (RTGCs) are the most widely used pieces of equipment in the Hong Kong sea-freight container yards. Workload Distribution in the yard changes continuously over time. The dynamic deployment of RTGCs is an important issue in yard operation management. This paper investigates the dynamic crane deployment problem with the objective of determining the crane deployment frequency and routes over a planning horizon to minimize the total Workload overflow. The problem is formulated as a mixed integer programming model. A heuristic algorithm is then developed to solve problems of practical sizes. The heuristic quickly finds a near optimal solution for crane deployment operation.
-
dynamic crane deployment in container storage yards
Transportation Research Part B-methodological, 2002Co-Authors: Chuqian Zhang, Yatwah Wan, Jiyin Liu, Richard J LinnAbstract:Storage yards at container terminals serve as temporary buffers for inbound and outbound containers. Rubber tyred gantry cranes (RTGCs) are the most frequently used equipment in yards for container handling. The efficiency of yard operations heavily depends on the productivity of these RTGCs. As the Workload Distribution in the yard changes over time, dynamic deployment of RTGCs among storage blocks (container stacking areas) is an important issue of terminal operations management. This paper addresses the crane deployment problem. Given the forecasted Workload of each block in each period of a day, the objective is to find the times and routes of crane movements among blocks so that the total delayed Workload in the yard is minimized. The problem is formulated as a mixed integer programming (MIP) model and solved by Lagrangean relaxation. To improve the performance of this approach, we augment the Lagrangean relaxation model by adding additional constraints and modify the solution procedure accordingly. Computational experiments show that the modified Lagrangean relaxation method generates excellent solutions in short time.
