The needed reformulation of algorithms and applications from different areas of research toward their usage for ultrascale systems and platforms has to address different challenges that arise from the different application areas, algorithms and programs. The challenges include scalability of the applications using a large number of system resources efficiently, the usage of resilience methods to include mechanisms to enable application programs to react to system failures, as well as the inclusion of energy-awareness features into the application programs to be able to obtain an energy-efficient execution. The programming models should enable to concentrate on the algorithmic aspects and problem-specific issues of the specific application area such that program development is supported as far as possible.

Chapter Contents:

• 6.1 Application-specific analytical energy models
• 6.1.1 Motivation
• 6.1.2 Measuring power and energy
• 6.1.2.1 Energy measurement techniques
• 6.1.2.2 Metrics to capture energy efficiency
• 6.1.3 Application-specific energy models
• 6.1.3.1 General power and energy models for frequency scaling
• 6.1.3.2 Theoretical and application-specific models
• 6.1.3.3 Power modeling for the SPLASH-2 benchmarks
• 6.1.3.4 Power modeling for multithreaded RK methods
• 6.1.3.5 Power modeling for MPI communications
• 6.1.3.6 Power modeling for the high-performance Linpack
• 6.1.4 Summary
• 6.2 On parallel algorithms for the numerical solution of problems with fractional powers of elliptic operators
• 6.2.1 Definitions of fractional power of elliptic operators
• 6.2.2 State of the art in numerical algorithms
• 6.2.3 Parallel algorithms
• 6.2.4 Extension to a mixed boundary value problem in the semi-infinite cylinder C= Ω x [0,∞) ⊂ Rd+1
• 6.2.5 Reduction to a pseudo-parabolic PDE problem
• 6.2.6 Integral representation of the solution of problem (6.21)
• 6.2.7 Approximation of the solution of problem (6.21) using rational approximations
• 6.2.8 Comparison of accuracy
• 6.2.9 Conclusions
• 6.3 Scalability of 3D relaxational gravity inversion in HPC Avitohol
• 6.3.1 Methodology of work
• 6.3.2 Inversion quality results and discussions
• 6.3.3 Scalability of inversion
• 6.3.4 Conclusions
• 6.4 Massive parallelization of the k-clique problem using subchromatic functions
• 6.4.1 Preliminaries
• 6.4.2 Subchromatic upper bounds
• 6.4.3 Disturbing structures
• 6.4.4 Partitioning the k-clique problem for parallel architectures
• 6.5 Summary

Inspec keywords: parallel processing; computational complexity; distributed algorithms

Other keywords: algorithmic aspects; system failures; system resources; energy-efficient execution; ultrascale systems; applications scalability; program development; programming models

Subjects: Computational complexity; Parallel software; Parallel programming and algorithm theory