On the arithmetic intensity of high-order finite-volume discretizations for hyperbolic systems of conservation laws
Abstract
It has been conjectured that higher-order discretizations for partial differential equations will have advantages over the lower-order counterparts commonly used today. The reasoning is that the increase in arithmetic operations will be more than offset by the reduction in data transfers and the increase in concurrent floating-point units. To evaluate this conjecture, the arithmetic intensity of a class of high-order finite-volume discretizations for hyperbolic systems of conservation laws is theoretically analyzed for spatial discretizations from orders three through eight in arbitrary dimensions. Additionally, three cache models are considered: the limiting cases of no cache and infinite cache as well as a finite-sized cache model. Models are validated experimentally by measuring floating-point operations and data transfers on an IBM Blue Gene/Q node. Theory and experiments demonstrate that high-order finite-volume methods will be able to provide increases in arithmetic intensity that will be necessary to make better utilization of on-node floating-point capability.
- Authors:
-
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Publication Date:
- Research Org.:
- Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
- Sponsoring Org.:
- USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- OSTI Identifier:
- 1860892
- Report Number(s):
- LLNL-JRNL-716535
Journal ID: ISSN 1094-3420; 859004
- Grant/Contract Number:
- AC52-07NA27344
- Resource Type:
- Accepted Manuscript
- Journal Name:
- International Journal of High Performance Computing Applications
- Additional Journal Information:
- Journal Volume: 33; Journal Issue: 1; Journal ID: ISSN 1094-3420
- Publisher:
- SAGE
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 97 MATHEMATICS AND COMPUTING; arithmetic intensity; high-order finite-volume methods; hyperbolic systems of conservation laws; processor-memory performance gap; algorithmic balance
Citation Formats
Loffeld, J., and Hittinger, J. A. F. On the arithmetic intensity of high-order finite-volume discretizations for hyperbolic systems of conservation laws. United States: N. p., 2017.
Web. doi:10.1177/1094342017691876.
Loffeld, J., & Hittinger, J. A. F. On the arithmetic intensity of high-order finite-volume discretizations for hyperbolic systems of conservation laws. United States. https://doi.org/10.1177/1094342017691876
Loffeld, J., and Hittinger, J. A. F. Wed .
"On the arithmetic intensity of high-order finite-volume discretizations for hyperbolic systems of conservation laws". United States. https://doi.org/10.1177/1094342017691876. https://www.osti.gov/servlets/purl/1860892.
@article{osti_1860892,
title = {On the arithmetic intensity of high-order finite-volume discretizations for hyperbolic systems of conservation laws},
author = {Loffeld, J. and Hittinger, J. A. F.},
abstractNote = {It has been conjectured that higher-order discretizations for partial differential equations will have advantages over the lower-order counterparts commonly used today. The reasoning is that the increase in arithmetic operations will be more than offset by the reduction in data transfers and the increase in concurrent floating-point units. To evaluate this conjecture, the arithmetic intensity of a class of high-order finite-volume discretizations for hyperbolic systems of conservation laws is theoretically analyzed for spatial discretizations from orders three through eight in arbitrary dimensions. Additionally, three cache models are considered: the limiting cases of no cache and infinite cache as well as a finite-sized cache model. Models are validated experimentally by measuring floating-point operations and data transfers on an IBM Blue Gene/Q node. Theory and experiments demonstrate that high-order finite-volume methods will be able to provide increases in arithmetic intensity that will be necessary to make better utilization of on-node floating-point capability.},
doi = {10.1177/1094342017691876},
journal = {International Journal of High Performance Computing Applications},
number = 1,
volume = 33,
place = {United States},
year = {Wed Feb 01 00:00:00 EST 2017},
month = {Wed Feb 01 00:00:00 EST 2017}
}
Works referenced in this record:
Improving the ratio of memory operations to floating-point operations in loops
journal, November 1994
- Carr, Steve; Kennedy, Ken
- ACM Transactions on Programming Languages and Systems, Vol. 16, Issue 6
A framework for hybrid parallel flow simulations with a trillion cells in complex geometries
conference, January 2013
- Godenschwager, Christian; Schornbaum, Florian; Bauer, Martin
- Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC '13
11 PFLOP/s simulations of cloud cavitation collapse
conference, January 2013
- Rossinelli, Diego; Koumoutsakos, Petros; Hejazialhosseini, Babak
- Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC '13
Understanding Application Performance via Micro-benchmarks on Three Large Supercomputers: Intrepid, Ranger and Jaguar
journal, May 2010
- Bhatelé, Abhinav; Wesolowski, Lukasz; Bohm, Eric
- The International Journal of High Performance Computing Applications, Vol. 24, Issue 4
Estimating interlock and improving balance for pipelined architectures
journal, August 1988
- Callahan, David; Cocke, John; Kennedy, Ken
- Journal of Parallel and Distributed Computing, Vol. 5, Issue 4
Multicore/Multi-GPU Accelerated Simulations of Multiphase Compressible Flows Using Wavelet Adapted Grids
journal, January 2011
- Rossinelli, Diego; Hejazialhosseini, Babak; Spampinato, Daniele G.
- SIAM Journal on Scientific Computing, Vol. 33, Issue 2
I/O complexity: The red-blue pebble game
conference, January 1981
- Jia-Wei, Hong; Kung, H. T.
- Proceedings of the thirteenth annual ACM symposium on Theory of computing - STOC '81
A Study on Balancing Parallelism, Data Locality, and Recomputation in Existing PDE Solvers
conference, November 2014
- Olschanowsky, Catherine; Strout, Michelle Mills; Guzik, Stephen
- SC14: International Conference for High Performance Computing, Networking, Storage and Analysis
Performance evaluations of gyrokinetic Eulerian code GT5D on massively parallel multi-core platforms
conference, January 2011
- Idomura, Yasuhiro; Jolliet, Sébastien
- State of the Practice Reports on - SC '11
An Optimizing Code Generator for a Class of Lattice-Boltzmann Computations
journal, July 2015
- Pananilath, Irshad; Acharya, Aravind; Vasista, Vinay
- ACM Transactions on Architecture and Code Optimization, Vol. 12, Issue 2
Comparison of accurate methods for the integration of hyperbolic equations
journal, January 1972
- Kreiss, Heinz-Otto; Oliger, Joseph
- Tellus, Vol. 24, Issue 3
Quantifying Performance Bottlenecks of Stencil Computations Using the Execution-Cache-Memory Model
conference, January 2015
- Stengel, Holger; Treibig, Jan; Hager, Georg
- Proceedings of the 29th ACM on International Conference on Supercomputing - ICS '15
Essentially non-oscillatory and weighted essentially non-oscillatory schemes for hyperbolic conservation laws
book, January 1998
- Shu, Chi-Wang
- Lecture Notes in Mathematics
Compiler-Directed Transformation for Higher-Order Stencils
conference, May 2015
- Basu, Protonu; Hall, Mary; Williams, Samuel
- 2015 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
Communication lower bounds and optimal algorithms for numerical linear algebra
journal, May 2014
- Ballard, G.; Carson, E.; Demmel, J.
- Acta Numerica, Vol. 23
High-order finite-volume methods for hyperbolic conservation laws on mapped multiblock grids
journal, May 2015
- McCorquodale, P.; Dorr, M. R.; Hittinger, J. A. F.
- Journal of Computational Physics, Vol. 288
On the GPU performance of cell-centered finite volume method over unstructured tetrahedral meshes
conference, January 2013
- Langguth, Johannes; Wu, Nan; Chai, Jun
- Proceedings of the 3rd Workshop on Irregular Applications Architectures and Algorithms - IA^3 '13
Weighted Essentially Non-oscillatory Schemes
journal, November 1994
- Liu, Xu-Dong; Osher, Stanley; Chan, Tony
- Journal of Computational Physics, Vol. 115, Issue 1
Fully multidimensional flux-corrected transport algorithms for fluids
journal, June 1979
- Zalesak, Steven T.
- Journal of Computational Physics, Vol. 31, Issue 3
Optimization of a lattice Boltzmann computation on state-of-the-art multicore platforms
journal, September 2009
- Williams, Samuel; Carter, Jonathan; Oliker, Leonid
- Journal of Parallel and Distributed Computing, Vol. 69, Issue 9
Solving the compressible navier-stokes equations on up to 1.97 million cores and 4.1 trillion grid points
conference, November 2013
- Bermejo-Moreno, Iván; Bodart, Julien; Larsson, Johan
- SC13: International Conference for High Performance Computing, Networking, Storage and Analysis, Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis
A 30 Year Retrospective on Dennard's MOSFET Scaling Paper
journal, January 2007
- Bohr, Mark
- IEEE Solid-State Circuits Newsletter, Vol. 12, Issue 1
A high-order finite-volume method for conservation laws on locally refined grids
journal, January 2011
- McCorquodale, Peter; Colella, Phillip
- Communications in Applied Mathematics and Computational Science, Vol. 6, Issue 1
The Piecewise Parabolic Method (PPM) for gas-dynamical simulations
journal, April 1984
- Colella, Phillip; Woodward, Paul R.
- Journal of Computational Physics, Vol. 54, Issue 1
Strong Stability-Preserving High-Order Time Discretization Methods
journal, January 2001
- Gottlieb, Sigal; Shu, Chi-Wang; Tadmor, Eitan
- SIAM Review, Vol. 43, Issue 1
Hierarchical N-body Simulations with Autotuning for Heterogeneous Systems
journal, May 2012
- Yokota, Rio; Barba, Lorena
- Computing in Science & Engineering, Vol. 14, Issue 3
A performance analysis framework for identifying potential benefits in GPGPU applications
journal, September 2012
- Sim, Jaewoong; Dasgupta, Aniruddha; Kim, Hyesoon
- ACM SIGPLAN Notices, Vol. 47, Issue 8
Performance modeling of serial and parallel implementations of the fractional Adams-Bashforth-Moulton method
journal, June 2014
- Zhang, Wei; Wei, Wenjie; Cai, Xing
- Fractional Calculus and Applied Analysis, Vol. 17, Issue 3
Managing application complexity in the SAMRAI object-oriented framework
journal, January 2002
- Hornung, Richard D.; Kohn, Scott R.
- Concurrency and Computation: Practice and Experience, Vol. 14, Issue 5
Efficient Implementation of Weighted ENO Schemes
journal, June 1996
- Jiang, Guang-Shan; Shu, Chi-Wang
- Journal of Computational Physics, Vol. 126, Issue 1
High-order, finite-volume methods in mapped coordinates
journal, April 2011
- Colella, P.; Dorr, M. R.; Hittinger, J. A. F.
- Journal of Computational Physics, Vol. 230, Issue 8
High throughput software for direct numerical simulations of compressible two-phase flows
conference, November 2012
- Hejazialhosseini, Babak; Rossinelli, Diego; Conti, Christian
- 2012 SC - International Conference for High Performance Computing, Networking, Storage and Analysis, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis
Roofline: an insightful visual performance model for multicore architectures
journal, April 2009
- Williams, Samuel; Waterman, Andrew; Patterson, David
- Communications of the ACM, Vol. 52, Issue 4
An efficient mixed-precision, hybrid CPU–GPU implementation of a nonlinearly implicit one-dimensional particle-in-cell algorithm
journal, June 2012
- Chen, G.; Chacón, L.; Barnes, D. C.
- Journal of Computational Physics, Vol. 231, Issue 16
Flux-corrected transport. I. SHASTA, a fluid transport algorithm that works
journal, January 1973
- Boris, Jay P.; Book, David L.
- Journal of Computational Physics, Vol. 11, Issue 1