Third Party Funds Group - Sub project
ExaStencils - Advanced Stencil-Code Engineering



Project Details
Acronym: ExaStencils
Start date: 01/01/2013
End date: 31/12/2018

Abstract (technical / expert description):

Future exascale computing systems with 107 processing units and supporting up to 1018 FLOPS peak performance will require a tight co-design of application, algorithm, and architecture aware program development to sustain this performance for many applications of interest, mainly for two reasons. First, the node structure inside an exascale cluster will become increasingly heterogeneous, always exploiting the most recent available on-chip manycore/GPU/HWassist technology. Second, the clusters themselves will be composed of heterogeneous subsystems and interconnects. As a result, new software techniques and tools supporting the joint algorithm and architecture-aware program development will become indispensable not only (a) to ease application and program development, but also (b) for performance analysis and tuning, (c) to ensure short turn-around times, and (d) for reasons of portability.

Project ExaStencils will investigate and provide a unique, tool-assisted, domain-specific codesign approach for the important class of stencil codes, which play a central role in high performance simulation on structured or block-structured grids. Stencils are regular access patterns on (usually multidimensional) data grids. Multigrid methods involve a hierarchy of very fine to successively coarser grids. The challenge of exascale is that, for the coarser grids, less processing power is required and communication dominates. From the computational algorithm perspective, domain-specific investigations include the extraction and development of suitable stencils, the analysis of performance-relevant algorithmic tradeoffs (e.g., the number of grid levels) and the analysis and reduction of synchronization requirements guided by a template model of the targeted cluster architecture. Based on this analysis, sophisticated programming and software tool support shall be developed by capturing the relevant data structures and program segments for stencil computations in a domain-specific language and applying a generator-based product-line technology to generate and optimize automatically stencil codes tailored to each application–platform pair. A central distinguishing mark of ExaStencils is that domain knowledge is being pursued in a coordinated manner across all abstraction levels, from the formulation of the application scenario down to the generation of highly-optimized stencil code.

For the developed unique and first-time seamless cross-level design flow, the three objectives of (1) a substantial gain in productivity, (2) high flexibility in the choice of algorithm and execution platform, and (3) the provision of the ExaFLOPS performance for stencil code shall be demonstrated in a detailed, final evaluation phase.


Overall project details
Overall project: SPP 1648: Software for Exascale Computing

Publications

Journal article
Köstler H, Schmitt C, Kuckuk S, et al. (2017)
A Scala Prototype to Generate Multigrid Solver Implementations for Different Problems and Target Multi-Core Platforms
International Journal of Computational Science and Engineering

Unpublished / Preprint
Schmitt C, Schmid M, Kuckuk S, et al. (2017)
Reconfigurable Hardware Generation of Multigrid Solvers with Conjugate Gradient Coarse-Grid Solution
Parallel Processing Letters -

Journal article
Unat D, Dubey A, Hoefler T, et al. (2017)
Trends in Data Locality Abstractions for HPC Systems
IEEE Transactions on Parallel and Distributed Systems

Article in Edited Volumes
Schmid M, Schmitt C, Hannig F, et al. - Ed.: Dirk Koch, Frank Hannig, and Daniel Ziener (2016)
Big Data and HPC Acceleration with Vivado HLS
FPGAs for Software Programmers

Article in Edited Volumes
Schmitt C, Kuckuk S, Hannig F, et al. (2016)
Systems of Partial Differential Equations in ExaSlang
Software for Exascale Computing - SPPEXA 2013-2015 - Lecture Notes in Computational Science and Engineering

Conference contribution
Schmitt C, Schmid M, Hannig F, et al. (2015)
Generation of Multigrid-based Numerical Solvers for FPGA Accelerators
Proceedings of the 2nd International Workshop on High-Performance Stencil Computations (HiStencils)

Conference contribution
Schmitt C, Kuckuk S, Köstler H, et al. (2014)
An Evaluation of Domain-Specific Language Technologies for Code Generation
Proc. of the 14th International Conference on Computational Science and its Applications (ICCSA)

Other publication type (Technical Report)
Lengauer C, Apel S, Bolten M, et al. - Ed.: Universität Passau (2014)
ExaStencils: Advanced Stencil-Code Engineering - First Project Report

Journal article
Grebhahn A, Kuckuk S, Schmitt C, et al. (2014)
Experiments on Optimizing the Performance of Stencil Codes with SPL Conqueror
Parallel Processing Letters

Conference contribution
Grebhahn A, Siegmund N, Apel S, et al. (2014)
Optimizing Performance of Stencil Code with SPL Conqueror
Proceedings of the 1st International Workshop on High-Performance Stencil Computations (HiStencils)

Journal article
Membarth R, Reiche O, Schmitt C, et al. (2014)
Towards a Performance-portable Description of Geometric Multigrid Algorithms using a Domain-specific Language
Journal of Parallel and Distributed Computing

Conference contribution
Lengauer C, Apel S, Größlinger A, et al. (2014)
ExaStencils: Advanced Stencil-Code Engineering
Proc. of Euro-Par 2014: Parallel Processing Workshops -

Conference contribution
Schmitt C, Kuckuk S, Hannig F, et al. (2014)
ExaSlang: A Domain-Specific Language for Highly Scalable Multigrid Solvers
Proc. of the 4th International Workshop on Domain-Specific Languages and High-Level Frameworks for High Performance Computing (WOLFHPC)

Conference contribution
Schmid M, Reiche O, Schmitt C, et al. (2014)
Code Generation for High-Level Synthesis of Multiresolution Applications on FPGAs
Proc. of the First International Workshop on FPGAs for Software Programmers (FSP)

Share link
Last updated on 2017-04-21 at 10:02
PDF downloaded successfully