Solving global optimization problems on GPU cluster

Barkalov, Konstantin; Gergel, Victor; Lebedev, Ilya

doi:10.1063/1.4952194

Cited by 22 publications

(16 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where ε max is large enough (in our experiments ε max was set equal to 20). The function Er(ε) can be multiextremal, non-differentiable and hard to evaluate even at one value of ε, since each its computation requires to reconstruct the interpolant (1). It is supposed that Er(ε) satisfies the Lipschitz condition over the interval [0, ε max ]:…”

Section: Statement Of the Optimization Problemmentioning

confidence: 99%

An Experimental Study of Univariate Global Optimization Algorithms for Finding the Shape Parameter in Radial Basis Functions

Mukhametzhanov

Cavoretto²,

Rossi³

2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

In this contribution, an interpolation problem using radial basis functions is considered. A recently proposed approach for the search of the optimal value of the shape parameter is studied. The approach consists of using global optimization algorithms to minimize the error function obtained using a leave-one-out cross validation (LOOCV) technique, which is commonly used for solving machine learning problems. In this paper, the proposed approach is studied experimentally on classes of randomly generated test problems using the GKLS-generator, which is widely used for testing global optimization algorithms. The experimental study on classes of randomly generated test problems is very important from the practical point of view, since results show the behavior of the algorithms for solving not a single test problem, but the whole class with controllable difficulty, which is the main property of the GKLS-generator. The obtained results are relevant, since the experiments have been carried out on 200 randomized test problems, and show that the algorithms are efficient for solving difficult real-life problems demonstrating a promising behavior.

show abstract

Section: Statement Of the Optimization Problemmentioning

confidence: 99%

An Experimental Study of Univariate Global Optimization Algorithms for Finding the Shape Parameter in Radial Basis Functions

Mukhametzhanov

Cavoretto²,

Rossi³

2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

show abstract

“…Additional information on such parallel computation schemes can be found in [22]. 4. Globalizer System Architecture.…”

Section: Parallel Computations For Systems With Distributed Memorymentioning

confidence: 99%

“…One of the first developments was the SYMOP multiextremal optimization system [17], which has been applied for solving many optimization problems. The ExaMin system [2], [3], [4], [16] was developed and used to investigate various parallel algorithms for solving the global optimization problems on the high-performance computational systems.…”

Section: Parallel Computations For Systems With Distributed Memorymentioning

confidence: 99%

Globalizer: A novel supercomputer software system for solving time-consuming global optimization problems

Gergel¹,

Barkalov²,

Sysoyev³

2018

Numerical Algebra, Control &Amp; Optimization

Self Cite

View full text Add to dashboard Cite

In this paper, we describe the Globalizer software system for solving the global optimization problems. The system is designed to maximize the use of computational potential of the modern high-performance computational systems in order to solve the most time-consuming optimization problems. The highly parallel computations are facilitated using various distinctive computational schemes: processing several optimization iterations simultaneously, reducing multidimensional optimization problems using multiple Peano spacefilling curves, and multi-stage computing based on the nested block reduction schemes. These novelties provide for the use of the supercomputer system capabilities with shared and distributed memory and with large numbers of processors to solve the global optimization problems efficiently.

show abstract

“…-And finally, once again, nobody knows how to make efficient fully Bayesian use of a cluster of GPUs in the (n, p) = (Big, Big) setting, where such a cluster would be most needed if it could be effectively utilized. If it's sufficient in your problem to settle for MAP estimates, see Barkalov, Gergel and Lebedev (2016) for one approach to computing them on a GPU cluster.…”

Section: Many Of the Sharding Papers Concentrate On Examples In Which...mentioning

confidence: 99%

Comment: A brief survey of the current state of play for Bayesian computation in data science at big-data scale

Draper¹,

Terenin²

2017

Braz. J. Probab. Stat.

View full text Add to dashboard Cite

Solving global optimization problems on GPU cluster

Cited by 22 publications

References 20 publications

An Experimental Study of Univariate Global Optimization Algorithms for Finding the Shape Parameter in Radial Basis Functions

An Experimental Study of Univariate Global Optimization Algorithms for Finding the Shape Parameter in Radial Basis Functions

Globalizer: A novel supercomputer software system for solving time-consuming global optimization problems

Comment: A brief survey of the current state of play for Bayesian computation in data science at big-data scale

Contact Info

Product

Resources

About