@PREAMBLE{ " \newcommand{\noop}[1]{} " } @Article{2016geofe, author = {{Ye}, R. and {de Hoop}, M.~V. and {Petrovitch}, C.~L. and {Pyrak-Nolte}, L.~J. and {Wilcox}, L.~C.}, title = {A discontinuous Galerkin method with a modified penalty flux for the propagation and scattering of acousto-elastic waves}, journal = {Geophysical Journal International}, archiveprefix = "arXiv", eprint = {1511.00675}, primaryclass = "physics.comp-ph", keywords = {Numerical approximations and analysis, Interface waves, Seismic anisotropy, Wave propagation}, year = 2016, month = may, volume = 205, pages = {1267-1289}, doi = {10.1093/gji/ggw070}, adsurl = {http://adsabs.harvard.edu/abs/2016GeoJI.205.1267Y}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} } @Article{absrobingpml, author = {{Grote}, M.~J. and {Sim}, I.}, title = {Efficient PML for the wave equation}, journal = {ArXiv e-prints}, archiveprefix = "arXiv", eprint = {1001.0319}, primaryclass = "math.NA", keywords = {Mathematics - Numerical Analysis, 35L05, 35L20, 65M06, 65M12}, year = 2010, month = jan, adsurl = {http://adsabs.harvard.edu/abs/2010arXiv1001.0319G}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} } @InProceedings{activelearning, author = {Duplyakin, Dmitry and Brown, Jed and Ricci, Robert}, booktitle = {Cluster Computing (CLUSTER), 2016 IEEE International Conference on}, date-added = {2017-03-04 18:44:01 +0000}, date-modified = {2017-03-06 13:58:01 +0000}, organization = {IEEE}, pages = {182--191}, title = {{Active Learning in Performance Analysis}}, year = {2016} } @Article{alkhalifah2000, author = {{Alkhalifah}, Tariq}, title = {An acoustic wave equation for anisotropic media}, journal = {Geophysics}, year = 2000, volume = 65, number = 4, pages = {1239-1250} } @Article{alnaes2014, title = {{U}nified {F}orm {L}anguage: a domain-specific language for weak formulations of partial differential equations}, author = {Aln{\ae}s, Martin S. and Logg, Anders and {\O}lgaard, Kristian B. and Rognes, Marie E. and Wells, Garth N.}, journal = {ACM Transactions on Mathematical Software (TOMS)}, volume = {40}, number = {2}, pages = {9}, year = {2014}, publisher = {ACM} } @TechReport{alpal, author = {Cook Jr, Grant O}, institution = {Lawrence Livermore National Lab., CA (USA)}, title = {ALPAL: A tool for the development of large-scale simulation codes}, year = {1988} } @Unpublished{am2016, author = {Amestoy, Patrick and Buttari, Alfredo and L'Excellent, Jean-Yves and Mary, Théo}, title = {Sparse direct solvers towards seismic imaging of large 3D domains (78th EAGE Conference, workshop methods and challenges of seismic wave modelling for seismic imaging, Vienna, Austria, 30/05/2016-02/06/2016)}, year = {2016}, booktitle = {EAGE Annual Conference Proceedings}, language = {anglais}, url = {http://mary.perso.enseeiht.fr/doc/EAGE16.pdf} } @InProceedings{andreolli2014genetic, author = {Andreolli, C and Thierry, P and Borges, L and Yount, C and Skinner, G}, booktitle = {EAGE Workshop on High Performance Computing for Upstream}, date-added = {2017-02-09 11:27:13 +0000}, date-modified = {2017-02-09 11:27:13 +0000}, title = {Genetic algorithm based auto-tuning of seismic applications on multi and manycore computers}, year = {2014} } @InCollection{andreolli2015, address = {Boston}, author = {Cedric Andreolli and Philippe Thierry and Leonardo Borges and Gregg Skinner and Chuck Yount}, booktitle = {High Performance Parallelism Pearls}, doi = {http://dx.doi.org/10.1016/B978-0-12-802118-7.00023-6}, editor = {Reinders, James and Jeffers, Jim}, isbn = {978-0-12-802118-7}, pages = {377 - 396}, publisher = {Morgan Kaufmann}, title = {Chapter 23 - Characterization and Optimization Methodology Applied to Stencil Computations}, url = {http://www.sciencedirect.com/science/article/pii/B9780128021187000236}, year = {2015}, bdsk-url-1 = {http://www.sciencedirect.com/science/article/pii/B9780128021187000236}, bdsk-url-2 = {http://dx.doi.org/10.1016/B978-0-12-802118-7.00023-6} } @article{aochi2013, doi = {10.1088/1742-6596/454/1/012010}, url = {https://doi.org/10.1088%2F1742-6596%2F454%2F1%2F012010}, year = 2013, month = {aug}, publisher = {{IOP} Publishing}, volume = {454}, pages = {012010}, author = {Hideo Aochi and Thomas Ulrich and Ariane Ducellier and Fabrice Dupros and David Michea}, title = {Finite difference simulations of seismic wave propagation for understanding earthquake physics and predicting ground motions: Advances and challenges}, journal = {Journal of Physics: Conference Series}, abstract = {Seismic waves radiated from an earthquake propagate in the Earth and the ground shaking is felt and recorded at (or near) the ground surface. Understanding the wave propagation with respect to the Earth's structure and the earthquake mechanisms is one of the main objectives of seismology, and predicting the strong ground shaking for moderate and large earthquakes is essential for quantitative seismic hazard assessment. The finite difference scheme for solving the wave propagation problem in elastic (sometimes anelastic) media has been more widely used since the 1970s than any other numerical methods, because of its simple formulation and implementation, and its easy scalability to large computations. This paper briefly overviews the advances in finite difference simulations, focusing particularly on earthquake mechanics and the resultant wave radiation in the near field. As the finite difference formulation is simple (interpolation is smooth), an easy coupling with other approaches is one of its advantages. A coupling with a boundary integral equation method (BIEM) allows us to simulate complex earthquake source processes.} } @TechReport{asanovic2006landscape, author = {Asanovic, Krste and Bodik, Ras and Catanzaro, Bryan Christopher and Gebis, Joseph James and Husbands, Parry and Keutzer, Kurt and Patterson, David A and Plishker, William Lester and Shalf, John and Williams, Samuel Webb and others}, institution = {Technical Report UCB/EECS-2006-183, EECS Department, University of California, Berkeley}, title = {The landscape of parallel computing research: A view from berkeley}, year = {2006} } @InProceedings{autotuning-1, acmid = {2259037}, address = {New York, NY, USA}, author = {Zhang, Yongpeng and Mueller, Frank}, booktitle = {Proceedings of the Tenth International Symposium on Code Generation and Optimization}, date-added = {2016-04-16 10:18:22 +0000}, date-modified = {2016-04-16 10:18:32 +0000}, doi = {10.1145/2259016.2259037}, isbn = {978-1-4503-1206-6}, location = {San Jose, California}, numpages = {10}, pages = {155--164}, publisher = {ACM}, series = {CGO '12}, title = {Auto-generation and Auto-tuning of 3D Stencil Codes on GPU Clusters}, url = {http://doi.acm.org/10.1145/2259016.2259037}, year = {2012}, bdsk-url-1 = {http://doi.acm.org/10.1145/2259016.2259037}, bdsk-url-2 = {http://dx.doi.org/10.1145/2259016.2259037} } @InBook{awi, author = {Mike Warner and Lluís Guasch}, title = {Adaptive waveform inversion: Theory}, booktitle = {SEG Technical Program Expanded Abstracts 2014}, chapter = {}, pages = {1089-1093}, year = {2014}, doi = {10.1190/segam2014-0371.1}, url = {https://library.seg.org/doi/abs/10.1190/segam2014-0371.1}, eprint = {https://library.seg.org/doi/pdf/10.1190/segam2014-0371.1} } @InProceedings{backus1957fortran, author = {Backus, John W and Beeber, Robert J and Best, Sheldon and Goldberg, Richard and Haibt, L Mitchell and Herrick, Harlan L and Nelson, Robert A and Sayre, David and Sheridan, Peter B and Stern, H and others}, booktitle = {Papers presented at the February 26-28, 1957, western joint computer conference: Techniques for reliability}, date-added = {2017-04-27 15:32:19 +0000}, date-modified = {2017-04-27 15:32:19 +0000}, organization = {ACM}, pages = {188--198}, title = {The FORTRAN automatic coding system}, year = {1957} } @article{baines2004, title={Geological storage of carbon dioxide}, author={Baines, Shelagh J and Worden, Richard H}, journal={Geological Society, London, Special Publications}, volume={233}, number={1}, pages={1--6}, year={2004}, publisher={Geological Society of London} } @Article{barba2013will, author = {Barba, Lorena A and Yokota, Rio}, date-added = {2017-03-01 20:01:53 +0000}, date-modified = {2017-03-01 20:01:53 +0000}, journal = {SIAM News}, number = {6}, pages = {1--3}, title = {How will the fast multipole method fare in the exascale era}, volume = {46}, year = {2013} } @InProceedings{bassetti98, acmid = {709707}, address = {London, UK, UK}, author = {Bassetti, Federico and Davis, Kei and Quinlan, Daniel J.}, booktitle = {Proceedings of the Second International Symposium on Computing in Object-Oriented Parallel Environments}, date-added = {2016-03-30 16:05:48 +0000}, date-modified = {2016-03-30 16:05:55 +0000}, isbn = {3-540-65387-2}, numpages = {12}, pages = {107--118}, publisher = {Springer-Verlag}, series = {ISCOPE '98}, title = {Optimizing Transformations of Stencil Operations for Parallel Object-Oriented Scientific Frameworks on Cache-Based Architectures}, url = {http://dl.acm.org/citation.cfm?id=646894.709707}, year = {1998}, bdsk-url-1 = {http://dl.acm.org/citation.cfm?id=646894.709707} } @incollection{bathe08, author = {Klaus Jurgen Bathe}, title = {Finite Element Method}, booktitle = {Wiley Encyclopedia of Computer Science and Engineering}, year = {2008}, crossref = {DBLP:reference/wiley/2008}, url = {https://doi.org/10.1002/9780470050118.ecse159}, doi = {10.1002/9780470050118.ecse159}, timestamp = {Tue, 16 May 2017 14:02:23 +0200}, biburl = {https://dblp.org/rec/bib/reference/wiley/Bathe08}, bibsource = {dblp computer science bibliography, https://dblp.org} } @Article{baysal1983, author = {{Baysal}, Edip and {Kosloff}, Dan D. and and {Sherwood}, John W. C.}, title = {Reverse time migration}, journal = {Geophysics}, year = 1983, month = {november}, volume = 48, number = 11, pages = {1514-1524} } @Article{bigdata, author = {Sheng Xu and Yu Zhang and Bing Tang}, title = {3D angle gathers from reverse time migration}, journal = {GEOPHYSICS}, volume = {76}, number = {2}, pages = {S77-S92}, year = {2011}, doi = {10.1190/1.3536527}, url = {https://doi.org/10.1190/1.3536527}, eprint = {https://doi.org/10.1190/1.3536527} } @misc{bisbas, title={Accelerating real-world stencil computations using temporal blocking: handling sparse sources and receivers}, url={https://figshare.com/articles/Accelerating_real-world_stencil_computations_using_temporal_blocking_handling_sparse_sources_and_receivers/9439829/1}, DOI={10.6084/m9.figshare.9439829.v1}, abstract={Accelerating stencil computations has long been studied on a great variety of platforms and a lot of work has been done on enhancing data locality in order to optimize our computations through use of cache memory. Temporal blocking, also known as time-tiling is a well-known technique for accelerating stencil based computations by enhancing data locality. Solving partial differential equations with the finite difference method yields stencil codes that can solve complex problems described by PDEs. Related work is partly solving the issues encountered in this research area however there is no framework that currently supports the modelling of real-world problems, a high-level syntax of PDE’s, automatic parallel code generation (SIMD/OpenMP/MPI) and loop nest optimizations all together. We introduce an algorithm to apply temporal blocking to phenomena with source injection and is designed to be delivered, and automated as part of the Devito framework.}, publisher={figshare}, author={Bisbas, George and Luporini, Fabio and Louboutin, Mathias and Gorman, Gerard and H.J. Kelly, Paul}, year={2019}, month={Aug} } @article{bostock1999, author = {Bostock, M. G. and Rondenay, S.}, title = {Migration of scattered teleseismic body waves}, journal = {Geophysical Journal International}, volume = {137}, number = {3}, pages = {732-746}, year = {1999}, month = {06}, abstract = {The retrieval of near-receiver mantle structure from scattered waves associated with teleseismic P and S and recorded on three-component, linear seismic arrays is considered in the context of inverse scattering theory. A Ray + Born formulation is proposed which admits linearization of the forward problem and economy in the computation of the elastic wave Green's function. The high-frequency approximation further simplifies the problem by enabling (1) the use of an earth-flattened, 1-D reference model, (2) a reduction in computations to 2-D through the assumption of 2.5-D experimental geometry, and (3) band-diagonalization of the Hessian matrix in the inverse formulation. The final expressions are in a form reminiscent of the classical diffraction stack of seismic migration. Implementation of this procedure demands an accurate estimate of the scattered wave contribution to the impulse response, and thus requires the removal of both the reference wavefield and the source time signature from the raw record sections. An approximate separation of direct and scattered waves is achieved through application of the inverse free-surface transfer operator to individual station records and a Karhunen–Loeve transform to the resulting record sections. This procedure takes the full displacement field to a wave vector space wherein the first principal component of the incident wave-type section is identified with the direct wave and is used as an estimate of the source time function. The scattered displacement field is reconstituted from the remaining principal components using the forward free-surface transfer operator, and may be reduced to a scattering impulse response upon deconvolution of the source estimate. An example employing pseudo-spectral synthetic seismograms demonstrates an application of the methodology.}, issn = {0956-540X}, doi = {10.1046/j.1365-246x.1999.00813.x}, url = {https://doi.org/10.1046/j.1365-246x.1999.00813.x}, eprint = {http://oup.prod.sis.lan/gji/article-pdf/137/3/732/1695373/137-3-732.pdf}, } @InBook{bubesatti2016, abstract = { Second-order pseudo-acoustic systems are commonly used in reverse-time migration in tranversely isotropic media to model quasi-P-wave propagation, and have also been used in full-waveform inversion. Systems with the shear wavespeeds set to zero can su er from instabilities. Systems with positive shear wavespeeds behave better, but they can occasionally have di culties when material parameters are strongly variable, particularly when applying the adjoint system in full-waveform inversion. We present here a new family of 2 × 2 second-order pseudo-acoustic systems with positive shear wavespeeds for tranversely isotropic media that are time-reversible, self-adjoint, and conserve a complete energy. They are thus stable for both forward and adjoint computations. In addition, they require only two-thirds of the number of spatial derivatives compared with previous systems. Presentation Date: Tuesday, October 18, 2016 Start Time: 9:15:00 AM Location: 146 Presentation Type: ORAL }, author = {Kenneth Bube and John Washbourne and Raymond Ergas and Tamas Nemeth}, publisher = {SEG}, booktitle = {SEG Technical Program Expanded Abstracts 2016}, doi = {10.1190/segam2016-13878451.1}, eprint = {https://library.seg.org/doi/pdf/10.1190/segam2016-13878451.1}, pages = {1110-1114}, title = {Self-adjoint, energy-conserving second-order pseudoacoustic systems for VTI and TTI media for reverse time migration and full-waveform inversion}, url = {e}, year = {2016}, bdsk-url-1 = {https://library.seg.org/doi/abs/10.1190/segam2016-13878451.1}, bdsk-url-2 = {http://dx.doi.org/10.1190/segam2016-13878451.1} } @Article{bubetti2012, abstract = {We studied second-order wave propagation systems for vertical transversely isotropic (VTI) and tilted transversely isotropic (TTI) acoustic media with variable axes of symmetry that have their shear-wave speeds set to zero. Acoustic TTI systems are commonly used in reverse-time migration, but these second-order systems are susceptible to instablities appearing as nonphysical stationary noise growing linearly in time, particularly in variable-tilt TTI media. We found an explanation of the cause of this phenomenon. The instabilities are not caused only by the numerical schemes; they are inherent to the differential equations. These instabilities are present even in homogeneous VTI media. These instabilities are caused by zero wave speeds at a wide variety of wavenumbers --- a direct consequence of setting the shear-wave speeds to zero --- coupled with the second time derivative in these systems. Although the second-order isotropic wave equation allows smooth time-growing solutions, a larger class of time-growing solutions exists for the second-order acoustic TI systems, including nonsmooth solutions. Boundary conditions appear to be less effective in controlling these time-growing solutions than they are for the isotropic wave equation. These systems conserve an incomplete energy that does not prevent the instabilities. The corresponding steady-state systems are no longer elliptic differential equations and can have nonsmooth solutions that are related to the instabilities. We started initially with homogeneous VTI media, and then extended these results to heterogeneous variable-tilt TTI media. We also developed a second-order acoustic system for heterogeneous variable-tilt TTI media derived directly from the full-elastic system for heterogeneous variable-tilt TTI media. All second-order systems with a dispersion relation obtained by setting the shear-wave speeds to zero in the elastic dispersion relation allowed these nonphysical time-growing solutions; however, knowing the cause of these instabilities, it may be possible to prevent or control the activation of these solutions. }, author = {Kenneth P. Bube and Tamas Nemeth and Joseph P. Stefani and Ray Ergas and Wei Liu and Kurt T. Nihei and Linbin Zhang}, doi = {10.1190/geo2011-0250.1}, eprint = {https://doi.org/10.1190/geo2011-0250.1}, journal = {GEOPHYSICS}, publisher = {SEG}, number = {5}, pages = {T171-T186}, title = {On the instability in second-order systems for acoustic VTI and TTI media}, url = {https://doi.org/10.1190/geo2011-0250.1}, volume = {77}, year = {2012}, bdsk-url-1 = {https://doi.org/10.1190/geo2011-0250.1}, bdsk-url-2 = {http://dx.doi.org/10.1190/geo2011-0250.1} } @InBook{bubetti2014, abstract = { Summary Second-order pseudo-acoustic systems are commonly used in reverse-time migration in tranversely isotropic media to model quasi-P-wave propagation. Systems with the shear wavespeeds set to zero can su er from instabilities. Systems with positive shear wavespeeds behave better, but they can occasionally have di culties when material parameters are strongly variable, and also need to be adjusted in isotropic regions to prevent spurious artificial shear-wave modes. We describe a simple projection method to allow using zero shear wavespeeds in isotropic and elliptic subregions. We also obtain a more thorough resolution to these issues by modifying a popular second-order system with positive shear wavespeeds so that it both has a complete conserved energy, and also decouples into a single second-order equation and a second identically-zero equation in isotropic and elliptically anisotropic regions. This new system stays stable even in strongly variable media when the simple but usually effective projection method fails. }, author = {Kenneth P. Bube* and Ray Ergas and Tamas Nemeth}, booktitle = {SEG Technical Program Expanded Abstracts 2014}, doi = {10.1190/segam2014-0986.1}, eprint = {https://library.seg.org/doi/pdf/10.1190/segam2014-0986.1}, pages = {3439-3443}, title = {Stability and energy conservation for second-order acoustic systems for VTI andTTI media with positive shear wavespeeds}, publisher = {SEG}, url = {https://library.seg.org/doi/abs/10.1190/segam2014-0986.1}, year = {2014}, bdsk-url-1 = {https://library.seg.org/doi/abs/10.1190/segam2014-0986.1}, bdsk-url-2 = {http://dx.doi.org/10.1190/segam2014-0986.1} } @Article{cacheawareroofline, author = {Ilic, Aleksandar and Pratas, Frederico and Sousa, Leonel}, date-added = {2017-03-04 18:31:29 +0000}, date-modified = {2017-03-06 13:59:16 +0000}, journal = {IEEE Computer Architecture Letters}, number = {1}, pages = {21--24}, publisher = {IEEE}, title = {Cache-aware roofline model: {Upgrading} the loft}, volume = {13}, year = {2014} } @Article{cauchy1847methode, author = {Cauchy, Augustin-Louis}, citeulike-article-id={7133110}, day = {18}, journal = {Compte Rendu des S\'{e}ances de L'Acad\'{e}mie des Sciences XXV}, keywords = {first, steepest-descent}, month = oct, number = {25}, pages = {536--538}, posted-at = {2010-05-07 01:16:11}, priority = {2}, title = {{M\'{e}thode g\'{e}n\'{e}rale pour la r\'{e}solution des syst\`{e}mes d'\'{e}quations simultan\'{e}es}}, volume = {S\'{e}rie A}, year = {1847} } @article{cerjan, author = {Charles Cerjan and Dan Kosloff and Ronnie Kosloff and Moshe Reshef}, title = {A nonreflecting boundary condition for discrete acoustic and elastic wave equations}, journal = {GEOPHYSICS}, volume = {50}, number = {4}, pages = {705-708}, year = {1985}, doi = {10.1190/1.1441945}, URL = { https://doi.org/10.1190/1.1441945 }, eprint = { https://doi.org/10.1190/1.1441945 } , abstract = { One of the nagging problems which arises in application of discrete solution methods for wave‐propagation calculations is the presence of reflections or wraparound from the boundaries of the numerical mesh. These undesired events eventually override the actual seismic signals which propagate in the modeled region. The solution to avoiding boundary effects used to be to enlarge the numerical mesh, thus delaying the side reflections and wraparound longer than the range of times involved in the modeling. Obviously this solution considerably increases the expense of computation. More recently, nonreflecting boundary conditions were introduced for the finite‐difference method (Clayton and Enquist, 1977; Reynolds, 1978). These boundary conditions are based on replacing the wave equation in the boundary region by one‐way wave equations which do not permit energy to propagate from the boundaries into the numerical mesh. This approach has been relatively successful, except that its effectiveness degrades for events which impinge on the boundaries at shallow angles. It is also not clear how to apply this type of boundary condition to global discrete methods such as the Fourier method for which all grid points are coupled. } } @Misc{ cfd12, author = {Lorena A. Barba and Gilbert F. Forsyth}, title = { CFD Python: the 12 steps to Navier-Stokes equations.}, year = {2018}, journal = {Journal of Open Source Education}, volume = {1}, number = {9}, doi = {https://doi.org/10.21105/jose.00021} } @Article{cfl, author = {Wen‐Jing Wu and Larry R. Lines and Han‐Xing Lu}, date-modified = {2017-03-06 14:03:33 +0000}, doi = {10.1190/1.1444009}, eprint = {http://dx.doi.org/10.1190/1.1444009}, journal = {GEOPHYSICS}, number = {3}, pages = {845-856}, title = {Analysis of higher‐order, finite‐difference schemes in 3-{D} reverse‐time migration}, url = {http://dx.doi.org/10.1190/1.1444009}, volume = {61}, year = {1996}, bdsk-url-1 = {http://dx.doi.org/10.1190/1.1444009} } @Misc{ cgen, author = {Andreas Klöckner}, howpublished = {\url{https://github.com/inducer/cgen}}, title = {CGen - C/C++ source generation from an AST}, year = {2016} } @InProceedings{chan2013software, author = {Chan, Cy and Unat, Didem and Lijewski, Michael and Zhang, Weiqun and Bell, John and Shalf, John}, booktitle = {International Supercomputing Conference}, date-added = {2017-02-10 11:59:08 +0000}, date-modified = {2017-02-10 11:59:08 +0000}, organization = {Springer}, pages = {196--212}, title = {Software design space exploration for exascale combustion co-design}, year = {2013} } @Article{chu2011, author = {Chunlei Chu and Brian K. Macy and Phil D. Anno}, title = {Approximation of pure acoustic seismic wave propagation in TTI media}, journal = {GEOPHYSICS}, volume = {76}, number = {5}, pages = {WB97-WB107}, year = {2011}, doi = {10.1190/geo2011-0092.1}, url = {https://doi.org/10.1190/geo2011-0092.1}, eprint = {https://doi.org/10.1190/geo2011-0092.1}, abstract = { Pseudoacoustic anisotropic wave equations are simplified elastic wave equations obtained by setting the S-wave velocity to zero along the anisotropy axis of symmetry. These pseudoacoustic wave equations greatly reduce the computational cost of modeling and imaging compared to the full elastic wave equation while preserving P-wave kinematics very well. For this reason, they are widely used in reverse time migration (RTM) to account for anisotropic effects. One fundamental shortcoming of this pseudoacoustic approximation is that it only prevents S-wave propagation along the symmetry axis and not in other directions. This problem leads to the presence of unwanted S-waves in P-wave simulation results and brings artifacts into P-wave RTM images. More significantly, the pseudoacoustic wave equations become unstable for anisotropy parameters ϵ<δ and for heterogeneous models with highly varying dip and azimuth angles in tilted transversely isotropic (TTI) media. Pure acoustic anisotropic wave equations completely decouple the P-wave response from the elastic wavefield and naturally solve all the above-mentioned problems of the pseudoacoustic wave equations without significantly increasing the computational cost. In this work, we propose new pure acoustic TTI wave equations and compare them with the conventional coupled pseudoacoustic wave equations. Our equations can be directly solved using either the finite-difference method or the pseudospectral method. We give two approaches to derive these equations. One employs Taylor series expansion to approximate the pseudodifferential operator in the decoupled P-wave equation, and the other uses isotropic and elliptically anisotropic dispersion relations to reduce the temporal frequency order of the P-SV dispersion equation. We use several numerical examples to demonstrate that the newly derived pure acoustic wave equations produce highly accurate P-wave results, very close to results produced by coupled pseudoacoustic wave equations, but completely free from S-wave artifacts and instabilities. } } @InProceedings{cire-1, author = {Kaushik Datta and Samuel Williams and Vasily Volkov and Jonathan Carter and Leonid Oliker and John Shalf and Katherine Yelick}, title = {Auto-tuning the 27-point stencil for multicore}, booktitle = {In In Proc. iWAPT2009: The Fourth International Workshop on Automatic Performance Tuning}, year = {2009} } @InProceedings{cire-2, author = {Deitz, Steven J. and Chamberlain, Bradford L. and Snyder, Lawrence}, title = {Eliminating Redundancies in Sum-of-product Array Computations}, booktitle = {Proceedings of the 15th International Conference on Supercomputing}, series = {ICS '01}, year = {2001}, isbn = {1-58113-410-X}, location = {Sorrento, Italy}, pages = {65--77}, numpages = {13}, url = {http://doi.acm.org/10.1145/377792.377807}, doi = {10.1145/377792.377807}, acmid = {377807}, publisher = {ACM}, address = {New York, NY, USA} } @InProceedings{cire-3, title = {Redundancy Elimination in the ExaStencils Code Generator}, author = {Stefan Kronawitter and Sebastian Kuckuk and Christian Lengauer}, booktitle = {ICA3PP Workshops}, year = {2016} } @Book{ claerboutfd, author = {Claerbout, Jon F.}, title = {Imaging the Earth's Interior}, year = {1985}, isbn = {0-86542-304-0}, publisher = {Blackwell Scientific Publications, Inc.}, address = {Cambridge, MA, USA} } @Article{clayton1529, author = {Clayton, Robert and Engquist, Bj{\"o}rn}, title = {Absorbing boundary conditions for acoustic and elastic wave equations}, volume = {67}, number = {6}, pages = {1529--1540}, year = {1977}, publisher = {Bulletin of the Seismological Society of America}, abstract = {Boundary conditions are derived for numerical wave simulation that minimize artificial reflections from the edges of the domain of computation. In this way acoustic and elastic wave propagation in a limited area can be efficiently used to describe physical behavior in an unbounded domain. The boundary conditions are based on paraxial approximations of the scalar and elastic wave equations. They are computationally inexpensive and simple to apply, and they reduce reflections over a wide range of incident angles.}, issn = {0037-1106}, url = {http://bssa.geoscienceworld.org/content/67/6/1529}, eprint = {http://bssa.geoscienceworld.org/content/67/6/1529.full.pdf}, journal = {Bulletin of the Seismological Society of America} } @InProceedings{cohen-timetiling, acmid = {2628106}, address = {New York, NY, USA}, author = {Bondhugula, Uday and Bandishti, Vinayaka and Cohen, Albert and Potron, Guillain and Vasilache, Nicolas}, booktitle = {Proceedings of the 23rd International Conference on Parallel Architectures and Compilation}, doi = {10.1145/2628071.2628106}, isbn = {978-1-4503-2809-8}, keywords = {automatic parallelization, periodic, polyhedral model, stencils, tiling}, location = {Edmonton, AB, Canada}, numpages = {12}, pages = {39--50}, publisher = {ACM}, series = {PACT '14}, title = {Tiling and Optimizing Time-iterated Computations on Periodic Domains}, url = {http://doi.acm.org/10.1145/2628071.2628106}, year = {2014}, bdsk-url-1 = {http://doi.acm.org/10.1145/2628071.2628106}, bdsk-url-2 = {http://dx.doi.org/10.1145/2628071.2628106} } @article{courant1943, author = "Courant, R.", fjournal = "Bulletin of the American Mathematical Society", journal = "Bull. Amer. Math. Soc.", month = "01", number = "1", pages = "1--23", publisher = "American Mathematical Society", title = "Variational methods for the solution of problems of equilibrium and vibrations", url = "https://projecteuclid.org:443/euclid.bams/1183504922", volume = "49", year = "1943" } @Article{courant1967, author = {R. Courant and K. Friedrichs and H. Lewy}, journal = {International Business Machines (IBM) Journal of Research and Development}, title = {On the Partial Difference Equations of Mathematical Physics}, year = {1967}, volume = {11}, number = {2}, pages = {215-234}, keywords = {}, doi = {10.1147/rd.112.0215}, issn = {0018-8646}, month = {March} } @InProceedings{ctadel, author = {Van Engelen, Robert and Wolters, Lex and Cats, Gerard}, booktitle = {Proceedings of the 10th international conference on Supercomputing}, organization = {ACM}, pages = {86--93}, title = {Ctadel: A generator of multi-platform high performance codes for pde-based scientific applications}, year = {1996} } @Misc{ cx2-helen, author = {Imperial College High Performance Computing Service}, title = {The cx2/Helen cluster}, doi = {10.14469/hpc/2232 } } @InBook{dasilva2016, author = {Curt {Da Silva} and Felix Herrmann}, title = {A unified 2D/3D software environment for large-scale time-harmonic full-waveform inversion}, booktitle = {SEG Technical Program Expanded Abstracts 2016}, chapter = {}, pages = {1169-1173}, year = {2016}, doi = {10.1190/segam2016-13869051.1}, url = {http://library.seg.org/doi/abs/10.1190/segam2016-13869051.1}, eprint = {http://library.seg.org/doi/pdf/10.1190/segam2016-13869051.1} } @PhDThesis{datta2009auto, author = {Datta, Kaushik and Yelick, Katherine A}, date-added = {2017-02-10 15:21:28 +0000}, date-modified = {2017-02-10 15:21:28 +0000}, school = {University of California, Berkeley}, title = {Auto-tuning stencil codes for cache-based multicore platforms}, year = {2009} } @Article{datta2009optimization, author = {Datta, Kaushik and Kamil, Shoaib and Williams, Samuel and Oliker, Leonid and Shalf, John and Yelick, Katherine}, date-added = {2017-03-04 18:44:59 +0000}, date-modified = {2017-03-04 18:44:59 +0000}, journal = {SIAM review}, number = {1}, pages = {129--159}, publisher = {SIAM}, title = {Optimization and performance modeling of stencil computations on modern microprocessors}, volume = {51}, year = {2009} } @Article{dblp, author = {Lars Ruthotto and Eran Treister and Eldad Haber}, title = {jInv - a flexible Julia package for {PDE} parameter estimation}, journal = {CoRR}, volume = {abs/1606.07399}, year = {2016}, url = {http://arxiv.org/abs/1606.07399}, archiveprefix = {arXiv}, eprint = {1606.07399}, timestamp = {Wed, 07 Jun 2017 14:42:13 +0200}, biburl = {https://dblp.org/rec/bib/journals/corr/RuthottoTH16}, bibsource = {dblp computer science bibliography, https://dblp.org} } @ARTICLE{dehoop2016, author = {{Ye}, Ruichao and {de Hoop}, Maarten V. and {Petrovitch}, Christopher L. and {Pyrak-Nolte}, Laura J. and {Wilcox}, Lucas C.}, title = {A discontinuous Galerkin method with a modified penalty flux for the propagation and scattering of acousto-elastic waves}, journal = {Geophysical Journal International}, keywords = {Numerical approximations and analysis, Interface waves, Seismic anisotropy, Wave propagation, Physics - Computational Physics, Physics - Geophysics}, year = "2016", month = "May", volume = {205}, number = {2}, pages = {1267-1289}, doi = {10.1093/gji/ggw070}, archivePrefix = {arXiv}, eprint = {1511.00675}, primaryClass = {physics.comp-ph}, adsurl = {https://ui.adsabs.harvard.edu/abs/2016GeoJI.205.1267Y}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} } @Article{deqsol, author = {Umetani, Yukio}, journal = {Proc. IFIP TC2/WG22, 1985}, pages = {147--164}, title = {DEQSOL A numerical Simulation Language for Vector/Parallel Processors}, volume = {5}, year = {1985} } @article{devito-api, author = {Louboutin, M. and Lange, M. and Luporini, F. and Kukreja, N. and Witte, P. A. and Herrmann, F. J. and Velesko, P. and Gorman, G. J.}, title = {Devito (v3.1.0): an embedded domain-specific language for finite differences and geophysical exploration}, journal = {Geoscientific Model Development}, volume = {12}, year = {2019}, number = {3}, pages = {1165--1187}, url = {https://www.geosci-model-dev.net/12/1165/2019/}, doi = {10.5194/gmd-12-1165-2019} } @ARTICLE{devito-arm, author = {{Senger}, Hermes and {de Souza}, Jaime F. and {Gomi}, Edson S. and {Luporini}, Fabio and {Gorman}, Gerard J.}, title = {Performance of Devito on HPC-Optimised ARM Processors}, journal = {arXiv e-prints}, keywords = {Computer Science - Performance}, year = "2019", month = "Aug", eid = {arXiv:1908.03653}, pages = {arXiv:1908.03653}, archivePrefix = {arXiv}, eprint = {1908.03653}, primaryClass = {cs.PF}, adsurl = {https://ui.adsabs.harvard.edu/abs/2019arXiv190803653S}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} } @Article{devito-compiler, author = { {Luporini}, F. and {Lange}, M. and {Louboutin}, M. and {Kukreja}, N. and {H{\"u}ckelheim}, J. and {Yount}, C. and {Witte}, P. and {Kelly}, P.~H.~J. and {Gorman}, G.~J. and {Herrmann}, F.~J. }, title = { Architecture and performance of Devito, a system for automated stencil computation }, journal = { CoRR }, volume = { abs/1807.03032 }, month = { jul }, year = { 2018 }, url = { http://arxiv.org/abs/1807.03032 }, archiveprefix = { arXiv }, eprint = { 1807.03032 } } @article {devito-gpu, title = {Automatic code generation for GPUs using Devito}, year = {2019}, note = {Submitted to Rice Oil and Gas High Performance Computing Conference 2020 on November 27, 2019}, abstract = {This paper focuses on adding GPU support to the finite difference domain specific language Devito. Devito is already capable of generating highly optimized finite difference code for CPUs, including the Intel KNL, ARM and Power architectures, parallelized using OpenMP and MPI. It is typically used as a wave propagator and to calculate gradients using the adjoint-state method in RTM and FWI. We consider a range of GPU programming models for automatic code generation with respect to simplicity, performance and portability to GPUs from different chip manufacturers.}, keywords = {GPU, HPC, Devito, finite-differences}, author = {Fabio Luporini and Gerard Gorman} } @Misc{ devito-performance, author = {{Zenodo/devito-performance}}, title = {{Devito Experimentation Framework}}, month = {July}, year = {2018}, doi = {TODO} } @Misc{ devito31, author = {{Zenodo/devito}}, title = {{Devito v3.1}}, month = {October}, year = {2017}, doi = {10.5281/zenodo.836688} } @Misc{ devito_zenodo, author = {Michael Lange and Fabio Luporini and Mathias Louboutin and Navjot Kukreja and Vincenzo Pandolfo and Paulius Kazakas and Paulius Velesko and Shuhao Zhang and Peng Peng and Gerard Gorman}, title = {opesci/devito: Devito-3.0}, month = may, year = 2017, doi = {10.5281/zenodo.581384}, url = {https://doi.org/10.5281/zenodo.581384} } @Article{dijkstra1963design, author = {Dijkstra, Edsger W}, date-added = {2017-04-28 13:48:57 +0000}, date-modified = {2017-04-28 13:48:57 +0000}, journal = {Annual Review in Automatic Programming}, pages = {27--42}, publisher = {Elsevier}, title = {On the design of machine independent programming languages}, volume = {3}, year = {1963} } @Article{dispersion, author = {Lines, L. and Slawinski, R. and Bording, R.}, doi = {10.1190/1.1444605}, journal = {GEOPHYSICS}, note = {\url{http://library.seg.org/doi/pdf/10.1190/1.1444605}}, number = {3}, pages = {967-969}, title = {A recipe for stability of finite-difference wave-equation computations}, volume = {64}, year = {1999}, bdsk-url-1 = {http://dx.doi.org/10.1190/1.1444605} } @Article{doi:10.1137/120900745, author = {P. E. Farrell and C. J. Cotter and S. W. Funke}, title = {A Framework for the Automation of Generalized Stability Theory}, journal = {SIAM Journal on Scientific Computing}, volume = {36}, number = {1}, pages = {C25-C48}, year = {2014}, doi = {10.1137/120900745}, url = {http://dx.doi.org/10.1137/120900745}, eprint = {http://dx.doi.org/10.1137/120900745} } @Article{doi:10.1190/1.3008547, author = {Mostafa Naghizadeh and Mauricio D. Sacchi}, title = { f-x adaptive seismic-trace interpolation}, journal = {GEOPHYSICS}, volume = {74}, number = {1}, pages = {V9-V16}, year = {2009}, doi = {10.1190/1.3008547}, url = { https://doi.org/10.1190/1.3008547 }, eprint = { https://doi.org/10.1190/1.3008547 } } @Article{doi:10.1190/geo2010-0231.1, author = {Yang Liu and Sergey Fomel}, title = {Seismic data interpolation beyond aliasing using regularized nonstationary autoregression}, journal = {GEOPHYSICS}, volume = {76}, number = {5}, pages = {V69-V77}, year = {2011}, doi = {10.1190/geo2010-0231.1}, url = { https://doi.org/10.1190/geo2010-0231.1 }, eprint = { https://doi.org/10.1190/geo2010-0231.1 } } @Article{dolfinadjoint, author = {P. E. Farrell and D. A. Ham and S. W. Funke and M. E. Rognes}, title = {Automated Derivation of the Adjoint of High-Level Transient Finite Element Programs}, journal = {SIAM Journal on Scientific Computing}, volume = {35}, number = {4}, pages = {C369-C393}, year = {2013}, doi = {10.1137/120873558}, url = {http://dx.doi.org/10.1137/120873558}, eprint = {http://dx.doi.org/10.1137/120873558} } @InProceedings{dongarra, acmid = {742568}, address = {London, UK, UK}, author = {Dongarra, Jack}, booktitle = {Proceedings of the 1st International Conference on Supercomputing}, date-modified = {2017-03-06 13:57:52 +0000}, isbn = {3-540-18991-2}, numpages = {19}, pages = {456--474}, publisher = {Springer-Verlag}, title = {The {LINPACK} Benchmark: {An} Explanation}, url = {http://dl.acm.org/citation.cfm?id=647970.742568}, year = {1988}, bdsk-url-1 = {http://dl.acm.org/citation.cfm?id=647970.742568} } @InProceedings{dongarra:1987:lbe:647970.742568, acmid = {742568}, address = {London, UK, UK}, author = {Dongarra, Jack}, booktitle = {Proceedings of the 1st International Conference on Supercomputing}, isbn = {3-540-18991-2}, numpages = {19}, pages = {456--474}, publisher = {Springer-Verlag}, title = {The LINPACK Benchmark: An Explanation}, url = {http://dl.acm.org/citation.cfm?id=647970.742568}, year = {1988}, bdsk-url-1 = {http://dl.acm.org/citation.cfm?id=647970.742568} } @Book{ dragonbook, acknowledgement={#ack-nhfb#}, address = {Boston, MA, USA}, bibdate = {Tue Jan 30 16:21:16 MST 2007}, bibsource = {http://www.math.utah.edu/pub/tex/bib/master.bib; z3950.loc.gov:7090/Voyager}, edition = {Second}, editor = {Alfred V. Aho and Monica S. Lam and Ravi Sethi and Jeffrey D. Ullman}, isbn = {0-321-48681-1}, isbn-13 = {978-0-321-48681-3}, lccn = {QA76.76.C65 A37 2007}, libnote = {Not yet in my library.}, pages = {xxiv + 1009}, publisher = {Pearson/Addison Wesley}, remark = {Revised edition of \cite{Aho:CPT86}}, subject = {Compilers (Computer programs)}, title = {Compilers: principles, techniques, and tools}, url = {http://www.loc.gov/catdir/toc/ecip0618/2006024333.html}, year = {2007}, bdsk-url-1 = {http://www.loc.gov/catdir/toc/ecip0618/2006024333.html} } @Article{dse-ref-glore, author = {Ding, Yufei and Shen, Xipeng}, title = {GLORE: Generalized Loop Redundancy Elimination Upon LER-notation}, journal = {Proc. ACM Program. Lang.}, issue_date = {October 2017}, volume = {1}, number = {OOPSLA}, month = oct, year = {2017}, issn = {2475-1421}, pages = {74:1--74:28}, articleno = {74}, numpages = {28}, url = {http://doi.acm.org/10.1145/3133898}, doi = {10.1145/3133898}, acmid = {3133898}, publisher = {ACM}, address = {New York, NY, USA}, keywords = {loop redundancy elimination, operation minimization, program optimization} } @Article{duveneck, author = {Eric Duveneck and Peter M. Bakker}, title = {Stable P-wave modeling for reverse-time migration in tilted TI media}, journal = {GEOPHYSICS}, volume = {76}, number = {2}, pages = {S65-S75}, year = {2011}, doi = {10.1190/1.3533964}, url = {https://doi.org/10.1190/1.3533964}, eprint = {https://doi.org/10.1190/1.3533964} } @InProceedings{early-tiling, author = {Irigoin, Fran{\c{c}}ois and Triolet, R{\'e}mi}, booktitle = {Proceedings of the 15th ACM SIGPLAN-SIGACT symposium on Principles of programming languages}, date-added = {2016-05-27 09:37:48 +0000}, date-modified = {2016-05-27 09:37:57 +0000}, organization = {ACM}, pages = {319--329}, title = {Supernode partitioning}, year = {1988} } @Article{ecmmodel, author = {Hofmann, Johannes and Eitzinger, Jan and Fey, Dietmar}, date-added = {2017-03-04 18:01:42 +0000}, date-modified = {2017-03-06 13:58:58 +0000}, journal = {arXiv preprint arXiv:1509.03118}, title = {Execution-cache-memory performance model: {Introduction} and validation}, year = {2015} } @InProceedings{ecmmodelstencil, author = {Stengel, Holger and Treibig, Jan and Hager, Georg and Wellein, Gerhard}, booktitle = {Proceedings of the 29th ACM on International Conference on Supercomputing}, date-added = {2017-03-04 17:56:44 +0000}, date-modified = {2017-03-04 17:57:04 +0000}, organization = {ACM}, pages = {207--216}, title = {Quantifying performance bottlenecks of stencil computations using the execution-cache-memory model}, year = {2015} } @InProceedings{epicoco2014roofline, author = {Epicoco, Italo and Mocavero, Silvia and Macchia, Francesca and Aloisio, Giovanni}, booktitle = {High Performance Computing \& Simulation (HPCS), 2014 International Conference on}, date-added = {2017-02-09 10:40:30 +0000}, date-modified = {2017-02-09 10:40:30 +0000}, organization = {IEEE}, pages = {732--737}, title = {The roofline model for oceanic climate applications}, year = {2014} } @InBook{erlanggaherrkrylov, author = {Yogi A. Erlangga and Felix J. Herrmann}, title = {An iterative multilevel method for computing wavefields in frequency‐domain seismic inversion}, booktitle = {SEG Technical Program Expanded Abstracts 2008}, chapter = {}, pages = {1956-1960}, year = {2008}, doi = {10.1190/1.3059279}, url = {https://library.seg.org/doi/abs/10.1190/1.3059279}, eprint = {https://library.seg.org/doi/pdf/10.1190/1.3059279} } @Article{erlanggakrylov, author = {Y. A. Erlangga and C. W. Oosterlee and C. Vuik}, title = {A Novel Multigrid Based Preconditioner For Heterogeneous Helmholtz Problems}, journal = {SIAM Journal on Scientific Computing}, volume = {27}, number = {4}, pages = {1471-1492}, year = {2006}, doi = {10.1137/040615195}, url = { https://doi.org/10.1137/040615195 }, eprint = { https://doi.org/10.1137/040615195 } } @ARTICLE{ esser2016, author = {E.[] Esser and L.[] Guasch and T.[] van Leeuwen and A.[] Y.[] Aravkin and F.[] J.[] Herrmann}, title = {Total-variation regularization strategies in full-waveform inversion}, journal = {ArXiv e-prints}, archivePrefix = "arXiv", eprint = {1608.06159}, primaryClass = "math.OC", keywords = {Mathematics - Optimization and Control, 65K05, 65K10, 86-08}, year = 2016, month = aug, adsurl = {http://adsabs.harvard.edu/abs/2016arXiv160806159E}, adsnote = {Provided by the SAO/NASA Astrophysics Data System} } @Article{exasat, author = {Unat, Didem and Chan, Cy and Zhang, Weiqun and Williams, Samuel and Bachan, John and Bell, John and Shalf, John}, date-added = {2017-03-04 18:46:21 +0000}, date-modified = {2017-03-06 14:02:51 +0000}, journal = {The International Journal of High Performance Computing Applications}, number = {2}, pages = {209--232}, publisher = {SAGE Publications Sage UK: London, England}, title = {{ExaSAT}: An exascale co-design tool for performance modeling}, volume = {29}, year = {2015} } @InProceedings{exastencils, author = {Christian Lengauer and Sven Apel and Matthias Bolten and Armin Grosslinger and Frank Hannig and Harald Kostler and Ulrich Rude and Jurgen Teich and Alexander Grebhahn and Stefan Kronawitter and Sebastian Kuckuk and Hannah Rittich and Christian Schmitt}, title = {ExaStencils: Advanced Stencil-Code Engineering}, booktitle = {Euro-Par 2014: Parallel Processing Workshops - Euro-Par 2014 International Workshops, Porto, Portugal, August 25-26, 2014, Revised Selected Papers, Part {II}}, pages = {553--564}, year = {2014}, broken_crossref={DBLP:conf/europar/2014w2}, url = {https://doi.org/10.1007/978-3-319-14313-2_47}, doi = {10.1007/978-3-319-14313-2_47}, timestamp = {Mon, 06 Nov 2017 11:58:38 +0100}, biburl = {http://dblp.org/rec/bib/conf/europar/LengauerABGHKRTGKKRS14}, bibsource = {dblp computer science bibliography, http://dblp.org} } @PhDThesis{fabio-thesis, author = {Luporini, Fabio}, school = {Imperial College London}, title = {Automated optimization of numerical methods for partial differential equations}, year = {2017} } @Book{ fem-scott, address = {New York, NY}, author = {Brenner, Susanne C. and Scott, L. Ridgway}, citeulike-article-id={10835016}, citeulike-linkout-0={http://dx.doi.org/10.1007/978-0-387-75934-0}, citeulike-linkout-1={http://www.springerlink.com/content/978-0-387-75933-3}, doi = {10.1007/978-0-387-75934-0}, isbn = {978-0-387-75933-3}, posted-at = {2012-06-27 16:58:34}, priority = {2}, publisher = {Springer New York}, title = {{The Mathematical Theory of Finite Element Methods}}, url = {http://dx.doi.org/10.1007/978-0-387-75934-0}, volume = {15}, year = {2008} } @Book{ fenics, author = {Anders Logg and Kent-Andre Mardal and Garth N. Wells and others}, doi = {10.1007/978-3-642-23099-8}, isbn = {978-3-642-23098-1}, publisher = {Springer}, title = {Automated Solution of Differential Equations by the Finite Element Method}, year = {2012}, bdsk-url-1 = {http://dx.doi.org/10.1007/978-3-642-23099-8} } @Article{firedrake, author = {Florian Rathgeber and David A. Ham and Lawrence Mitchell and Michael Lange and Fabio Luporini and Andrew T. T. McRae and Gheorghe{-}Teodor Bercea and Graham R. Markall and Paul H. J. Kelly}, bibsource = {dblp computer science bibliography, http://dblp.org}, biburl = {http://dblp.uni-trier.de/rec/bib/journals/corr/RathgeberHMLLMBMK15}, date-added = {2016-04-16 10:07:43 +0000}, date-modified = {2016-04-16 10:07:43 +0000}, journal = {CoRR}, title = {Firedrake: automating the finite element method by composing abstractions}, url = {http://arxiv.org/abs/1501.01809}, volume = {abs/1501.01809}, year = {2015}, bdsk-url-1 = {http://arxiv.org/abs/1501.01809} } @Article{firedrake0, author = {Florian Rathgeber and David A. Ham and Lawrence Mitchell and Michael Lange and Fabio Luporini and Andrew T. T. McRae and Gheorghe{-}Teodor Bercea and Graham R. Markall and Paul H. J. Kelly}, bibsource = {dblp computer science bibliography, http://dblp.org}, biburl = {http://dblp.uni-trier.de/rec/bib/journals/corr/RathgeberHMLLMBMK15}, date-added = {2016-04-16 10:07:43 +0000}, date-modified = {2016-04-16 10:07:43 +0000}, journal = {CoRR}, title = {Firedrake: automating the finite element method by composing abstractions}, url = {http://arxiv.org/abs/1501.01809}, volume = {abs/1501.01809}, year = {2015}, bdsk-url-1 = {http://arxiv.org/abs/1501.01809} } @Article{fletcher, author = {Robin P. Fletcher and Xiang Du and Paul J. Fowler}, title = {Reverse time migration in tilted transversely isotropic (TTI) media}, journal = {GEOPHYSICS}, volume = {74}, number = {6}, pages = {WCA179-WCA187}, year = {2009}, doi = {10.1190/1.3269902}, url = {https://doi.org/10.1190/1.3269902}, eprint = {https://doi.org/10.1190/1.3269902} } @article{fletcher1964, author = {Fletcher, R. and Reeves, C. M.}, title = {Function minimization by conjugate gradients}, journal = {The Computer Journal}, volume = {7}, number = {2}, pages = {149-154}, year = {1964}, month = {01}, abstract = {A quadratically convergent gradient method for locating an unconstrained local minimum of a function of several variables is described. Particular advantages are its simplicity and its modest demands on storage, space for only three vectors being required. An ALGOL procedure is presented, and the paper includes a discussion of results obtained by its used on various test functions.}, issn = {0010-4620}, doi = {10.1093/comjnl/7.2.149}, url = {https://doi.org/10.1093/comjnl/7.2.149}, eprint = {http://oup.prod.sis.lan/comjnl/article-pdf/7/2/149/959725/070149.pdf}, } @InProceedings{florins, acmid = {964469}, address = {New York, NY, USA}, author = {Florins, Murielle and Vanderdonckt, Jean}, booktitle = {Proceedings of the 9th International Conference on Intelligent User Interfaces}, doi = {10.1145/964442.964469}, isbn = {1-58113-815-6}, keywords = {continuity, design, graceful degradation, multiplatform systems, multiple computing platforms}, location = {Funchal, Madeira, Portugal}, numpages = {8}, pages = {140--147}, publisher = {ACM}, series = {IUI '04}, title = {Graceful Degradation of User Interfaces As a Design Method for Multiplatform Systems}, url = {http://doi.acm.org/10.1145/964442.964469}, year = {2004}, bdsk-url-1 = {http://doi.acm.org/10.1145/964442.964469}, bdsk-url-2 = {http://dx.doi.org/10.1145/964442.964469} } @Article{fowlertti2010, abstract = { Reverse time migration (RTM) images reflectors by using time-extrapolation modeling codes to synthesize source and receiver wavefields in the subsurface. Asymptotic analysis of wave propagation in transversely isotropic (TI) media yields a dispersion relation describing coupled P- and SV-wave modes. This dispersion relation can be converted into a fourth-order scalar partial differential equation (PDE). Increased computational efficiency can be achieved using equivalent coupled second-order PDEs. Analysis of the corresponding dispersion relations as matrix eigenvalue systems allows one to characterize all possible coupled linear second-order systems equivalent to a given linear fourth-order PDE and to determine which ones yield optimally efficient finite-difference implementations. Setting the shear velocity along the axis of symmetry to zero yields a simpler approximate TI wave equation that is more efficient to implement. This simpler approximation, however, can become unstable for some plausible combinations of anisotropic parameters. The same eigensystem analysis can be applied using finite vertical shear velocity to obtain solutions that avoid these instability problems. }, author = {Paul J. Fowler and Xiang Du and Robin P. Fletcher}, doi = {10.1190/1.3294572}, eprint = {https://doi.org/10.1190/1.3294572}, journal = {GEOPHYSICS}, number = {1}, pages = {S11-S22}, title = {Coupled equations for reverse time migration in transversely isotropic media}, url = {https://doi.org/10.1190/1.3294572}, volume = {75}, year = {2010}, bdsk-url-1 = {https://doi.org/10.1190/1.3294572}, bdsk-url-2 = {http://dx.doi.org/10.1190/1.3294572} } @article{frederiksen2000, author = {Frederiksen, A. W. and Bostock, M. G.}, title = {Modelling teleseismic waves in dipping anisotropic structures}, journal = {Geophysical Journal International}, volume = {141}, number = {2}, pages = {401-412}, year = {2000}, month = {05}, abstract = {The existence of seismic discontinuities within the continental upper mantle has long been recognized, with more recent studies often indicating an association with elastic anisotropy. Their near-vertical sampling renders teleseismic P and S waves suitable for characterization of mantle discontinuities, but computationally efficient methods of calculating synthetic seismograms are required for structures that exhibit lateral variability. We consider lithospheric models consisting of planar, homogeneous anisotropic layers with arbitrary dip. We adopt the traveltime equation of Diebold for dipping, plane-layered media as the basis for a high-frequency asymptotic method that does not require ray tracing. Traveltimes of plane waves in anisotropic media are calculated from simple analytic formulae involving the depths of layers beneath a station and the vertical components of phase slowness within the layers. We compute amplitudes using the reflection and transmission matrices for planar interfaces separating homogeneous anisotropic media. Modelling indicates that upper-mantle seismic responses depend in a complex fashion on both layer dip and anisotropy, particularly in the case of converted phases. Azimuthal anisotropy generally displays a distinctive 180° backazimuthal periodicity in Ps conversion amplitude, as opposed to the 360° symmetry produced by dip. In contrast, anisotropy with a steeply plunging axis may under certain conditions be difficult to distinguish from interface dip, as both exhibit a 360° symmetry. We demonstrate the application of the method on Ps and Sp conversion data from the Yellowknife Array, which show evidence for both dipping and anisotropic layering, attributed to layers of anisotropi c fabric in the upper mantle associated with ancient subducted slabs.}, issn = {0956-540X}, doi = {10.1046/j.1365-246x.2000.00090.x}, url = {https://doi.org/10.1046/j.1365-246x.2000.00090.x}, eprint = {http://oup.prod.sis.lan/gji/article-pdf/141/2/401/1825416/141-2-401.pdf}, } @Article{frigo, author = {M. Frigo and S. Johnson}, doi = {10.1109/JPROC.2004.840301}, issn = {0018-9219}, journal = {Proceedings of the IEEE}, month = {Feb}, number = {2}, pages = {216-231}, title = {The Design and Implementation of FFTW3}, volume = {93}, year = {2005}, bdsk-url-1 = {http://dx.doi.org/10.1109/JPROC.2004.840301} } @Article{gpr:gpr977, author = {Symes, William W. and Sun, Dong and Enriquez, Marco}, doi = {10.1111/j.1365-2478.2011.00977.x}, issn = {1365-2478}, journal = {Geophysical Prospecting}, keywords = {Adjoint, Born modelling, Initialization}, number = {5}, pages = {814--833}, publisher = {Blackwell Publishing Ltd}, title = {From modelling to inversion: designing a well-adapted simulator}, url = {http://dx.doi.org/10.1111/j.1365-2478.2011.00977.x}, volume = {59}, year = {2011}, bdsk-url-1 = {http://dx.doi.org/10.1111/j.1365-2478.2011.00977.x} } @Article{griewank2000ara, ad_theotech = {Checkpointing}, author = {Andreas Griewank and Andrea Walther}, coden = {ACMSCU}, issn = {0098-3500}, journal = {{ACM} Transactions on Mathematical Software}, note = {Also appeared as Technical University of Dresden, Technical Report IOKOMO-04-1997.}, number = {1}, pages = {19--45}, title = {Algorithm 799: {R}evolve: {A}n Implementation of Checkpoint for the Reverse or Adjoint Mode of Computational Differentiation}, volume = {26}, year = {2000} } @Article{haber10tremp, abstract = {Often, parameter estimation problems of parameter-dependent PDEs involve multiple right-hand sides. The computational cost and memory requirements of such problems increase linearly with the number of right-hand sides. For many applications this is the main bottleneck of the computation. In this paper we show that problems with multiple right-hand sides can be reformulated as stochastic programming problems by combining the right-hand sides into a few {\quotedblbase}simultaneous{\textquotedblright} sources. This effectively reduces the cost of the forward problem and results in problems that are much cheaper to solve. We discuss two solution methodologies: namely sample average approximation and stochastic approximation. To illustrate the effectiveness of our approach we present two model problems, direct current resistivity and seismic tomography.}, author = {Eldad Haber and Matthias Chung and Felix J. Herrmann}, journal = {SIAM Journal on Optimization}, keywords = {FWI, Optimization, SLIM}, month = {7}, number = {3}, publisher = {UBC-Earth and Ocean Sciences Department}, title = {An effective method for parameter estimation with {PDE} constraints with multiple right hand sides}, url = {http://dx.doi.org/10.1137/11081126X}, volume = {22}, year = {2012}, bdsk-url-1 = {http://dx.doi.org/10.1137/11081126X} } @InProceedings{halide, author = {Ragan-Kelley, Jonathan and Barnes, Connelly and Adams, Andrew and Paris, Sylvain and Durand, Fr{\'e}do and Amarasinghe, Saman}, title = {Halide: A Language and Compiler for Optimizing Parallelism, Locality, and Recomputation in Image Processing Pipelines}, booktitle = {Proceedings of the 34th ACM SIGPLAN Conference on Programming Language Design and Implementation}, series = {PLDI '13}, year = {2013}, isbn = {978-1-4503-2014-6}, location = {Seattle, Washington, USA}, pages = {519--530}, numpages = {12}, url = {http://doi.acm.org/10.1145/2491956.2462176}, doi = {10.1145/2491956.2462176}, acmid = {2462176}, publisher = {ACM}, address = {New York, NY, USA}, keywords = {autotuning, compiler, domain specific language, gpu, image processing, locality, optimization,parallelism, redundant computation, vectorization} } @Article{hawick2009turning, author = {Hawick, KA and Playne, DP}, date-added = {2017-03-29 15:54:44 +0000}, date-modified = {2017-03-29 15:54:44 +0000}, journal = {Computer Science, Massey University, Tech. Rep}, title = {Turning partial differential equations into scalable software}, year = {2009} } @InProceedings{hbm-tiling-2016, author = {Charles Yount and Alejandro Duran}, title = {Effective Use of Large High-Bandwidth Memory Caches in {HPC} Stencil Computation via Temporal Wave-Front Tiling}, booktitle = {Proceedings of the 7th International Workshop in Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems held as part of {ACM/IEEE} Supercomputing 2016 ({SC16})}, series = {PMBS'16}, month = {Nov}, year = {2016}, location = {Salt Lake City, Utah} } @Misc{ helen-top, author = {TOP500}, howpublished = {\url{https://www.top500.org/system/178845}}, title = {cx2/Helen cluster specification in the TOP500 ranking.}, year = {2016} } @article{hestenes1952, added-at = {2008-10-07T16:03:39.000+0200}, author = {Hestenes, M. R. and Stiefel, E.}, biburl = {https://www.bibsonomy.org/bibtex/2362a03244482b96f9f9ee4af6eb7f80c/brefeld}, interhash = {bcb34a6f8b9fb2f2371e92430116a8ad}, intrahash = {362a03244482b96f9f9ee4af6eb7f80c}, journal = {Journal of research of the National Bureau of Standards}, keywords = {imported}, pages = {409--436}, timestamp = {2008-10-07T16:03:39.000+0200}, title = {Methods of conjugate gradients for solving linear systems}, volume = 49, year = 1952 } @conference {herrmann2019EAGEHPCaii, title = {Accelerating ideation \& innovation cheaply in the Cloud the power of abstraction, collaboration \& reproducibility}, booktitle = {4th EAGE Workshop on High-performance Computing}, year = {2019}, note = {(EAGE HPC Workshop, Dubai)}, month = {10}, keywords = {cloud, devito, EAGE, HPC, JUDI}, presentation = {https://slim.gatech.edu/Publications/Public/Conferences/EAGEHPC/2019/herrmann2019EAGEHPCaii/herrmann2019EAGEHPCaii_pres.pdf}, author = {Felix J. Herrmann and Charles Jones and Gerard Gorman and Jan H{\"u}ckelheim and Keegan Lensink and Paul Kelly and Navjot Kukreja and Henryk Modzelewski and Michael Lange and Mathias Louboutin and Fabio Luporini and James Selvages and Philipp A. Witte} } @conference {herrmann2019NIPSliwcuc, title = {Learned imaging with constraints and uncertainty quantification}, booktitle = {Neural Information Processing Systems (NeurIPS)}, year = {2019}, note = {Accepted on October 1, 2019}, month = {12}, abstract = {We outline new approaches to incorporate ideas from convolutional networks into wave-based least-squares imaging. The aim is to combine hand-crafted constraints with deep convolutional networks allowing us to directly train a network capable of generating samples from the posterior. The main contributions include combination of weak deep priors with hard handcrafted constraints and a possible new way to sample the posterior.}, keywords = {constraint, deep learning, Imaging, Uncertainty quantification}, url = {https://slim.gatech.edu/Publications/Public/Conferences/NIPS/2019/herrmann2019NIPSliwcuc/herrmann2019NIPSliwcuc.html}, presentation = {https://slim.gatech.edu/Publications/Public/Conferences/NIPS/2019/herrmann2019NIPSliwcuc/herrmann2019NIPSliwcuc_pres.pdf}, url2 = {https://openreview.net/pdf?id=Hyet2Q29IS}, author = {Felix J. Herrmann and Ali Siahkoohi and Gabrio Rizzuti} } @Article{higdon-bcs, issn = {00255718, 10886842}, url = {http://www.jstor.org/stable/2008250}, abstract = {We develop a theory of difference approximations to absorbing boundary conditions for the scalar wave equation in several space dimensions. This generalizes the work of the author described in [8]. The theory is based on a representation of analytical absorbing boundary conditions proven in [8]. These conditions are defined by compositions of first-order, one-dimensional differential operators. Here the operators are discretized individually, and their composition is used as a discretization of the boundary condition. The analysis of stability and reflection properties reduces to separate studies of the individual factors. A representation of the discrete boundary conditions makes it possible to perform the analysis geometrically, with little explicit calculation.}, author = {Robert L. Higdon}, journal = {Mathematics of Computation}, number = {179}, pages = {65--90}, publisher = {American Mathematical Society}, title = {Numerical Absorbing Boundary Conditions for the Wave Equation}, volume = {49}, year = {1987} } @InProceedings{hipacc, author = {Membarth, Richard and Hannig, Frank and Teich, J{\"u}rgen and K{\"o}stler, Harald}, booktitle = {High Performance Computing, Networking, Storage and Analysis (SCC), 2012 SC Companion:}, organization = {IEEE}, owner = {aluno}, pages = {1133--1138}, timestamp = {2016.08.22}, title = {Towards domain-specific computing for stencil codes in HPC}, year = {2012} } @InProceedings{historyoffortran, author = {Backus, John}, booktitle = {History of programming languages I}, date-added = {2017-04-27 15:19:21 +0000}, date-modified = {2017-04-27 15:19:30 +0000}, organization = {ACM}, pages = {25--74}, title = {The history of Fortran I, II, and III}, year = {1978} } @InProceedings{holt1960man, author = {Holt, AW and Turanski, WJ}, booktitle = {Papers presented at the May 3-5, 1960, western joint IRE-AIEE-ACM computer conference}, date-added = {2017-04-28 12:54:09 +0000}, date-modified = {2017-04-28 12:54:09 +0000}, organization = {ACM}, pages = {329--339}, title = {Man-to-machine communication and automatic code translation}, year = {1960} } @Book{hooke, author={Hooke, Robert and Papin, Denis and Sturmy, Samuel and Young, James}, title={Lectures de potentia restitutiva}, abstract={Lectures de potentia restitutiva, or, Of spring : explaining the power of springing bodies : to which are added some collections viz. a description of Dr. Pappins wind-fountain and force-pump, Mr. Young's observation concerning natural fountains, some other considerations concerning that subject, Captain Sturmy's remarks of a subterraneous cave and cistern, Mr. G.T. observations made on the Pike of Teneriff, 1674, some reflections and conjectures occasioned thereupon, a relation of a late eruption in the Isle of Palma}, year={1678}, publisher={London : Printed for John Martyn ...}, url={http://lib.ugent.be/catalog/rug01:001559640} } @InProceedings{hoppera0, author = {Hopper, Grace Murray}, booktitle = {Proceedings of the 1952 ACM national meeting (Pittsburgh)}, date-added = {2017-04-27 15:30:17 +0000}, date-modified = {2017-04-27 15:30:26 +0000}, organization = {ACM}, pages = {243--249}, title = {The education of a computer}, year = {1952} } @Book{ hpc-pearls, author = {Jeffers, Jim and Reinders, James}, title = {High Performance Parallelism Pearls Volume Two: Multicore and Many-core Programming Approaches}, year = {2015}, isbn = {0128038195, 9780128038192}, edition = {1st}, publisher = {Morgan Kaufmann Publishers Inc.}, address = {San Francisco, CA, USA} } @article{ huang2017, author = {Huang, Guanghui and Nammour, Rami and Symes, William}, title = {Full-waveform inversion via source-receiver extension}, volume = {82}, number = {3}, pages = {R153--R171}, year = {2017}, doi = {10.1190/geo2016-0301.1}, publisher = {Society of Exploration Geophysicists}, issn = {0016-8033}, URL = {http://geophysics.geoscienceworld.org/content/82/3/R153}, eprint = {http://geophysics.geoscienceworld.org/content/82/3/R153.full.pdf}, journal = {GEOPHYSICS} } @Book{ id2138, author = {Heiner Igel}, edition = {1.}, editor = {Oxford University Press}, month = {nov}, publisher = {Oxford University Press}, title = {{Computational Seismology: A Practical Introduction}}, year = {2016}, isbn = {9780198717409}, language = {en}, url = {https://global.oup.com/academic/product/computational-seismology-9780198717409?cc=de{\&}lang=en{\&}} } @Book{ iverson1962, author = {Iverson, Kenneth E.}, title = {A Programming Language}, year = {1962}, isbn = {0-471430-14-5}, source = {Library of Congress Catalog Card Number: 62-15180}, publisher = {John Wiley \& Sons, Inc.}, address = {New York, NY, USA} } @Article{iwave, author = {Symes, William W. and Sun, Dong and Enriquez, Marco}, doi = {10.1111/j.1365-2478.2011.00977.x}, issn = {1365-2478}, journal = {Geophysical Prospecting}, keywords = {Adjoint, Born modelling, Initialization}, number = {5}, pages = {814--833}, publisher = {Blackwell Publishing Ltd}, title = {From modelling to inversion: designing a well-adapted simulator}, url = {http://dx.doi.org/10.1111/j.1365-2478.2011.00977.x}, volume = {59}, year = {2011}, bdsk-url-1 = {http://dx.doi.org/10.1111/j.1365-2478.2011.00977.x} } @Article{jacobssbli, author = {Christian T. Jacobs and Satya P. Jammy and Neil D. Sandham}, bibsource = {dblp computer science bibliography, http://dblp.org}, biburl = {http://dblp.uni-trier.de/rec/bib/journals/corr/JacobsJS16}, journal = {CoRR}, timestamp = {Mon, 03 Oct 2016 17:51:10 +0200}, title = {OpenSBLI: {A} framework for the automated derivation and parallel execution of finite difference solvers on a range of computer architectures}, url = {http://arxiv.org/abs/1609.01277}, volume = {abs/1609.01277}, year = {2016}, bdsk-url-1 = {http://arxiv.org/abs/1609.01277} } @PhDThesis{jones1954survey, author = {Jones, John L}, date-added = {2017-04-28 10:49:52 +0000}, date-modified = {2017-04-28 10:49:52 +0000}, school = {Massachusetts Institute of Technology}, title = {A survey of automatic coding techniques for digital computers}, year = {1954} } @InProceedings{kahanscalar, author = {Hofmann, Johannes and Fey, Dietmar and Riedmann, Michael and Eitzinger, Jan and Hager, Georg and Wellein, Gerhard}, booktitle = {International Conference on Parallel Processing and Applied Mathematics}, date-added = {2017-03-04 18:50:57 +0000}, date-modified = {2017-03-06 13:59:08 +0000}, organization = {Springer}, pages = {63--73}, title = {Performance analysis of the {Kahan}-enhanced scalar product on current multicore processors}, year = {2015} } @InProceedings{kamil2006, acmid = {1178605}, address = {New York, NY, USA}, author = {Kamil, Shoaib and Datta, Kaushik and Williams, Samuel and Oliker, Leonid and Shalf, John and Yelick, Katherine}, booktitle = {Proceedings of the 2006 Workshop on Memory System Performance and Correctness}, date-modified = {2017-03-06 13:59:28 +0000}, doi = {10.1145/1178597.1178605}, isbn = {1-59593-578-9}, location = {San Jose, California}, numpages = {10}, pages = {51--60}, publisher = {ACM}, series = {MSPC '06}, title = {Implicit and Explicit Optimizations for {Stencil} Computations}, url = {http://doi.acm.org/10.1145/1178597.1178605}, year = {2006}, bdsk-url-1 = {http://doi.acm.org/10.1145/1178597.1178605}, bdsk-url-2 = {http://dx.doi.org/10.1145/1178597.1178605} } @Article{kennedy2005telescoping, author = {Kennedy, Ken and Broom, Bradley and Chauhan, Arun and Fowler, Robert J and Garvin, John and Koelbel, Charles and McCosh, Cheryl and Mellor-Crummey, John}, date-added = {2017-04-28 09:55:46 +0000}, date-modified = {2017-04-28 09:55:46 +0000}, journal = {Proceedings of the IEEE}, number = {2}, pages = {387--408}, publisher = {IEEE}, title = {Telescoping languages: A system for automatic generation of domain languages}, volume = {93}, year = {2005} } @InProceedings{kerncraft, author = {Hammer, Julian and Hager, Georg and Eitzinger, Jan and Wellein, Gerhard}, booktitle = {Proceedings of the 6th International Workshop on Performance Modeling, Benchmarking, and Simulation of High Performance Computing Systems}, date-added = {2017-03-04 18:12:30 +0000}, date-modified = {2017-03-06 13:58:47 +0000}, organization = {ACM}, pages = {4}, title = {Automatic loop kernel analysis and performance modeling with {KernCraft}}, year = {2015} } @Article{kim2011performance, author = {Kim, Ki-Hwan and Kim, KyoungHo and Park, Q-Han}, date-added = {2017-02-10 15:10:25 +0000}, date-modified = {2017-03-06 14:00:04 +0000}, journal = {Computer Physics Communications}, number = {6}, pages = {1201--1207}, publisher = {Elsevier}, title = {Performance analysis and optimization of three-dimensional {FDTD} on {GPU} using roofline model}, volume = {182}, year = {2011} } @Article{kimapnum, author = {Seongjai Kim, Hyeona Lim}, journal = {Applied Numerical Mathematics}, pages = {402--414}, publisher = {Elsevier B.V.}, title = {High-order schemes for acoustic waveform simulation}, url = {http://www.hl107.math.msstate.edu/pdfs/rein/HighANM_final.pdf}, volume = {57}, year = {2007}, bdsk-url-1 = {http://www.hl107.math.msstate.edu/pdfs/rein/HighANM_final.pdf} } @Book{ knupp2002verification, title = {Verification of Computer Codes in Computational Science and Engineering}, author = {Knupp, P. and Salari, K.}, isbn = {9781420035421}, lccn = {2002073821}, series = {Discrete Mathematics and Its Applications}, url = {https://books.google.com/books?id=PiEAeZ19JZgC}, year = {2002}, publisher = {CRC Press} } @Article{knuth1962history, author = {Knuth, Donald E}, date-added = {2017-04-27 15:54:49 +0000}, date-modified = {2017-04-27 15:54:49 +0000}, journal = {Computers and Automation}, number = {12}, pages = {8--18}, title = {A history of writing compilers}, volume = {11}, year = {1962} } @InProceedings{konstantinidis2015, author = {E. Konstantinidis and Y. Cotronis}, booktitle = {2015 23rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing}, date-modified = {2017-03-06 14:00:14 +0000}, doi = {10.1109/PDP.2015.51}, issn = {1066-6192}, month = {March}, pages = {651-658}, title = {A Practical Performance Model for Compute and Memory Bound {GPU} Kernels}, year = {2015}, bdsk-url-1 = {http://dx.doi.org/10.1109/PDP.2015.51} } @InProceedings{koster2014platform, author = {K{\"o}ster, Marcel and Lei{\ss}a, Roland and Hack, Sebastian and Membarth, Richard and Slusallek, Philipp}, booktitle = {Proceedings of the 1st International Workshop on High-Performance Stencil Computations}, date-added = {2017-03-29 14:51:10 +0000}, date-modified = {2017-03-29 14:51:10 +0000}, pages = {1--6}, title = {Platform-Specific Optimization and Mapping of Stencil Codes through Refinement}, year = {2014} } @article{krebsdg, author = "Krebs, J.R. and Collis, S.S. and Downey, N.J. and Ober, C.C. and Overfelt, J.R. and Smith, T.M. and van Bloemen-Waanders, B.G. and Young, J.G.", title = "Full Wave Inversion Using a Spectral-Element Discontinuous Galerkin Method", year = "2014", volume = "2014", number = "1", pages = "1-5", doi = "https://doi.org/10.3997/2214-4609.20140707", url = "https://www.earthdoc.org/content/papers/10.3997/2214-4609.20140707", publisher = "European Association of Geoscientists & Engineers", issn = "2214-4609", } @Article{krishnamoorthy07, acmid = {1250761}, address = {New York, NY, USA}, author = {Krishnamoorthy, Sriram and Baskaran, Muthu and Bondhugula, Uday and Ramanujam, J. and Rountev, Atanas and Sadayappan, P}, date-added = {2016-03-30 16:10:37 +0000}, date-modified = {2016-03-30 16:10:45 +0000}, doi = {10.1145/1273442.1250761}, issn = {0362-1340}, issue_date = {June 2007}, journal = {SIGPLAN Not.}, keywords = {automatic parallelization, load, stencil computations, tiling}, month = jun, number = {6}, numpages = {10}, pages = {235--244}, publisher = {ACM}, title = {Effective Automatic Parallelization of Stencil Computations}, url = {http://doi.acm.org/10.1145/1273442.1250761}, volume = {42}, year = {2007}, bdsk-url-1 = {http://doi.acm.org/10.1145/1273442.1250761}, bdsk-url-2 = {http://dx.doi.org/10.1145/1273442.1250761} } @Article{kukreja2016devito, author = {Kukreja, Navjot and Louboutin, Mathias and Vieira, Felippe and Luporini, Fabio and Lange, Michael and Gorman, Gerard}, date-added = {2017-05-22 14:34:19 +0000}, date-modified = {2017-05-22 14:34:30 +0000}, journal = {WolfHPC}, title = {Devito: automated fast finite difference computation}, year = {2016} } @Article{kukreja2018high, title = {High-level python abstractions for optimal checkpointing in inversion problems}, author = {Kukreja, Navjot and H{\"u}ckelheim, Jan and Lange, Michael and Louboutin, Mathias and Walther, Andrea and Funke, Simon W and Gorman, Gerard}, journal = {arXiv preprint arXiv:1802.02474}, year = {2018} } @unpublished{ kukreja2019PASCccd, title = {Combining checkpointing and data compression to accelerate adjoint-based optimization problems}, year = {2019}, note = {Submitted to PASC19 on January 16, 2019}, abstract = {Seismic inversion and imaging are adjoint-based optimization problems that processes up to terabytes of data, regularly exceeding the memory capacity of available computers. Data compression is an effective strategy to reduce this memory requirement by a certain factor, particularly if some loss in accuracy is acceptable. A popular alternative is checkpointing, where data is stored at selected points in time, and values at other times are recomputed as needed from the last stored state. This allows arbitrarily large adjoint computations with limited memory, at the cost of additional recomputations. In this paper we combine compression and checkpointing for the first time to compute a realistic seismic inversion. The combination of checkpointing and compression allows larger adjoint computations compared to using only compression, and reduces the recomputation overhead significantly compared to using only checkpointing.}, keywords = {Adjoint-state, checkpointing, compression, FD, HPC, Inverse problems, private}, url = {https://slim.gatech.edu/Publications/Private/Submitted/2019/kukreja2019PASCccd/kukreja2019PASCccd.pdf}, author = {Navjot Kukreja and Jan H{\"u}ckelheim and Mathias Louboutin and Kaiyuan Hou and Paul Hovland and Gerard Gorman} } @Article{kumar2015sss, title = {Source separation for simultaneous towed-streamer marine acquisition {\textendash}- a compressed sensing approach}, journal = {Geophysics}, volume = {80}, number = {06}, year = {2015}, note = {(Geophysics)}, month = {11}, pages = {WD73-WD88}, abstract = {Simultaneous marine acquisition is an economic way to sample seismic data and speed up acquisition, wherein single or multiple source vessels fire sources at near-simultaneous or slightly random times, resulting in overlapping shot records. The current paradigm for simultaneous towed-streamer marine acquisition incorporates {\textquotedblleft}low variability{\textquotedblright} in source firing times, i.e., 0 <= 1 or 2 s because the sources and receivers are moving. This results in a low degree of randomness in simultaneous data, which is challenging to separate (into its constituent sources) using compressed-sensing-based separation techniques because randomization is key to successful recovery via compressed sensing. We have addressed the challenge of source separation for simultaneous towed-streamer acquisitions via two compressed-sensing-based approaches, i.e., sparsity promotion and rank minimization. We have evaluated the performance of the sparsity-promotion- and rank-minimization-based techniques by simulating two simultaneous towed-streamer acquisition scenarios, i.e., over/under and simultaneous long offset. A field data example from the Gulf of Suez for the over/under acquisition scenario was also developed. We observed that the proposed approaches gave good and comparable recovery qualities of the separated sources, but the rank-minimization technique outperformed the sparsity-promoting technique in terms of the computational time and memory. We also compared these two techniques with the normal-moveout-based median-filtering-type approach, which had comparable results.}, keywords = {2D, Acquisition, inversion, marine, Optimization, Rank, source separation, sparsity}, doi = {10.1190/geo2015-0108.1}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Journals/Geophysics/2015/kumar2015sss/kumar2015sss_revised.pdf}, url2 = {http://library.seg.org/doi/abs/10.1190/geo2015-0108.1}, author = {Rajiv Kumar and Haneet Wason and Felix J. Herrmann} } @InProceedings{lai2013performance, author = {Lai, Junjie and Seznec, Andr{\'e}}, booktitle = {Code Generation and Optimization (CGO), 2013 IEEE/ACM International Symposium on}, date-added = {2017-02-09 15:23:46 +0000}, date-modified = {2017-03-06 14:00:36 +0000}, organization = {IEEE}, pages = {1--10}, title = {Performance upper bound analysis and optimization of {SGEMM} on {Fermi} and {Kepler} {GPUs}}, year = {2013} } @Article{lange2016devito, author = {Lange, Michael and Kukreja, Navjot and Louboutin, Mathias and Luporini, Fabio and Vieira, Felippe and Pandolfo, Vincenzo and Velesko, Paulius and Kazakas, Paulius and Gorman, Gerard}, date-added = {2017-05-22 14:35:16 +0000}, date-modified = {2017-05-22 14:35:22 +0000}, journal = {pyHPC}, title = {Devito: towards a generic finite difference DSL using symbolic Python}, year = {2016} } @InProceedings{lange2017scipy, author = { {M}. {L}ange and {N}. {K}ukreja and {F}. {L}uporini and {M}. {L}ouboutin and {C}. {Y}ount and {J}. {H}\"uckelheim and {G}. {J}. {G}orman }, title = { {O}ptimised finite difference computation from symbolic equations }, booktitle = { {P}roceedings of the 15th {P}ython in {S}cience {C}onference }, pages = { 89 - 96 }, year = { 2017 }, editor = { {K}aty {H}uff and {D}avid {L}ippa and {D}illon {N}iederhut and {M} {P}acer } } @Article{larner, author = {Ken Larner and Craig Beasley}, title = {Cascaded migrations: Improving the accuracy of finite‐difference migration}, journal = {GEOPHYSICS}, volume = {52}, number = {5}, pages = {618-643}, year = {1987}, doi = {10.1190/1.1442331}, url = { https://doi.org/10.1190/1.1442331 }, eprint = { https://doi.org/10.1190/1.1442331 } } @misc{lensink2017, author = {{Lensink}, Keegan}, title = {{SegyIO.jl}}, year = {2017}, publisher = {GitHub}, journal = {GitHub repository}, howpublished = {\url{https://github.com/slimgroup/SegyIO.jl}}, urldate = {2019-11-07}, commit = {d2054c52e7904816f198268145a37b0bf03b32a9} } @Article{libadjoint, author = {P. E. Farrell and S. W. Funke}, title = {A new approach for developing adjoint models}, journal = {AGU Fall Meeting Abstracts}, keywords = {0560 COMPUTATIONAL GEOPHYSICS / Numerical solutions, 0594 COMPUTATIONAL GEOPHYSICS / Instruments and techniques, Data Assimilation}, year = {2011}, month = {dec}, url = {http://adsabs.harvard.edu/abs/2011AGUFM.A33A0182F}, note = {Provided by the SAO/NASA Astrophysics Data System} } @Book{ lionsjl1971, author = {Lions, J. L.}, edition = {1st}, isbn = {978-3-642-65026-0}, publisher = {Springer-Verlag Berlin Heidelberg}, title = {Optimal control of systems governed by partial differential equations}, year = {1971} } @InProceedings{liu2009stable, author = {Liu, W and Bube, KP and Zhang, LB and Nihei, K}, booktitle = {71st EAGE Conference and Exhibition incorporating SPE EUROPEC 2009}, date-modified = {2017-03-06 14:01:05 +0000}, title = {Stable reverse-time migration in variable-tilt {TI} media}, year = {2009} } @InProceedings{loopy, author = {{Kl{\"o}ckner}, Andreas}, title = {Loo.py: transformation-based code~generation for GPUs and CPUs}, booktitle = {Proceedings of ARRAY `14: ACM SIGPLAN Workshop on Libraries, Languages, and Compilers for Array Programming}, year = 2014, publisher = {Association for Computing Machinery}, address = {Edinburgh, Scotland.}, doi = {10.1145/2627373.2627387} } @Conference{louboutin2015segtcs, abstract = {Time-domain Full-Waveform Inversion (FWI) aims to image the subsurface of the earth accurately from field recorded data and can be solved via the reduced adjoint-state method. However, this method requires access to the forward and adjoint wavefields that are meet when computing gradient updates. The challenge here is that the adjoint wavefield is computed in reverse order during time stepping and therefore requires storage or other type of mitigation because storing the full time history of the forward wavefield is too expensive in realistic 3D settings. To overcome this challenge, we propose an approximate adjoint-state method where the wavefields are subsampled randomly, which drastically the amount of storage needed. By using techniques from stochastic optimization, we control the errors induced by the subsampling. Examples of the proposed technique on a synthetic but realistic 2D model show that the subsampling-related artifacts can be reduced significantly by changing the sampling for each source after each model update. Combination of this gradient approximation with a quasi-Newton method shows virtually artifact free inversion results requiring only 5\% of storage compared to saving the history at Nyquist. In addition, we avoid having to recompute the wavefields as is required by checkpointing.}, author = {Mathias Louboutin and Felix J. Herrmann}, booktitle = {SEG Technical Program Expanded Abstracts}, doi = {10.1190/segam2015-5924937.1}, keywords = {acoustic, Full-waveform inversion, inversion, SEG, Stochastic optimization, Subsampling, Time-domain}, month = {10}, note = {(SEG, New Orleans)}, pages = {5153-5157}, presentation = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/SEG/2015/louboutin2015SEGtcs/louboutin2015SEGtcs_poster.pdf}, title = {Time compressively sampled full-waveform inversion with stochastic optimization}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/SEG/2015/louboutin2015SEGtcs/louboutin2015SEGtcs.html}, year = {2015}, bdsk-url-1 = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/SEG/2015/louboutin2015SEGtcs/louboutin2015SEGtcs.html}, bdsk-url-2 = {http://dx.doi.org/10.1190/segam2015-5924937.1} } @Article{louboutin2016ppf, author = {Mathias Louboutin and Michael Lange and Felix J. Herrmann and Navjot Kukreja and Gerard Gorman}, title = {Performance prediction of finite-difference solvers for different computer architectures}, journal = {Computers \& Geosciences}, volume = {105}, year = {2017}, month = {08}, pages = {148-157}, abstract = {The life-cycle of a partial differential equation (PDE) solver is often characterized by three development phases: the development of a stable numerical discretization; development of a correct (verified) implementation; and the optimization of the implementation for different computer architectures. Often it is only after significant time and effort has been invested that the performance bottlenecks of a PDE solver are fully understood, and the precise details varies between different computer architectures. One way to mitigate this issue is to establish a reliable performance model that allows a numerical analyst to make reliable predictions of how well a numerical method would perform on a given computer architecture, before embarking upon potentially long and expensive implementation and optimization phases. The availability of a reliable performance model also saves developer effort as it both informs the developer on what kind of optimisations are beneficial, and when the maximum expected performance has been reached and optimisation work should stop. We show how discretization of a wave-equation can be theoretically studied to understand the performance limitations of the method on modern computer architectures. We focus on the roofline model, now broadly used in the high-performance computing community, which considers the achievable performance in terms of the peak memory bandwidth and peak floating point performance of a computer with respect to algorithmic choices. A first principles analysis of operational intensity for key time-stepping finite-difference algorithms is presented. With this information available at the time of algorithm design, the expected performance on target computer systems can be used as a driver for algorithm design.}, keywords = {finite differences, HPC, Modeling, performance}, doi = {https://doi.org/10.1016/j.cageo.2017.04.014} } @conference{ louboutin2017eageess, title = {Extending the search space of time-domain adjoint-state FWI with randomized implicit time shifts}, booktitle = {EAGE Annual Conference Proceedings}, year = {2017}, note = {(EAGE, Paris)}, month = {06}, abstract = {Adjoint-state full-waveform inversion aims to obtain subsurface properties such as velocity, density or anisotropy parameters, from surface recorded data. As with any (non-stochastic) gradient based optimization procedure, the solution of this inversion procedure is to a large extend determined by the quality of the starting model. If this starting model is too far from the true model, these derivative-based optimizations will likely end up in local minima and erroneous inversion results. In certain cases, extension of the search space, e.g. by making the wavefields or focused matched sources additional unknowns, has removed some of these non-uniqueness issues but these rely on time-harmonic formulations. Here, we follow a different approach by combining an implicit extension of the velocity model, time compression techniques and recent results on stochastic sampling in non-smooth/non-convex optimization}, keywords = {cycle skipping, EAGE, FWI, inversion, nonconvex, time domain}, doi = {10.3997/2214-4609.201700831}, url = {https://slim.gatech.edu/Publications/Public/Conferences/EAGE/2017/louboutin2017EAGEess/louboutin2017EAGEess.html}, presentation = {https://slim.gatech.edu/Publications/Public/Conferences/EAGE/2017/louboutin2017EAGEess/louboutin2017EAGEess_pres.pdf}, author = {Mathias Louboutin and Felix J. Herrmann} } @article {louboutin2017fwi, title = {Full-Waveform Inversion - Part 1: forward modeling}, journal = {The Leading Edge}, volume = {36}, number = {12}, year = {2017}, note = {(The Leading Edge)}, month = {12}, pages = {1033-1036}, abstract = {Since its re-introduction by Pratt (1999), full-waveform inversion (FWI) has gained a lot of attention in geophysical exploration because of its ability to build high resolution velocity models more or less automatically in areas of complex geology. While there is an extensive and growing literature on the topic, publications focus mostly on technical aspects, making this topic inaccessible for a broader audience due to the lack of simple introductory resources for newcomers to geophysics. We will accomplish this by providing a hands-on walkthrough of FWI using Devito (Lange et al. 2016), a system based on domain-specific languages that automatically generates code for time-domainfinite-differences.}, keywords = {devito, finite-differences, FWI, Modeling, tutorial}, doi = {10.1190/tle36121033.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/TheLeadingEdge/2017/louboutin2017fwi/louboutin2017fwi.html}, author = {Mathias Louboutin and Philipp A. Witte and Michael Lange and Navjot Kukreja and Fabio Luporini and Gerard Gorman and Felix J. Herrmann} } @article {louboutin2017fwip2, title = {Full-Waveform Inversion - Part 2: adjoint modeling}, journal = {The Leading Edge}, volume = {37}, number = {1}, year = {2018}, note = {(The Leading Edge)}, month = {1}, pages = {69-72}, abstract = {This tutorial is the second part of a three part tutorial series on full-waveform inversion (FWI), in which we provide a step by step walk through of setting up forward and adjoint wave equation solvers and an optimization framework for inversion. In part 1 (Louboutin et al., 2017), we demonstrated how to discretize the acoustic wave equation and how to set up a basic forward modeling scheme using Devito, a domain-specific language (DSL) in Python for automated finite-difference (FD) computations (Lange et al., 2016). Devito allows us to define wave equations as symbolic Python expressions (Meurer et al., 2017), from which optimized FD stencil code is automatically generated at run time. In part 1, we show how we can use Devito to set up and solve acoustic wave equations with (impulsive) seismic sources and sample wavefields at the receiver locations to model shot records.}, keywords = {acoustic, devito, finite-difference, FWI, tutorial}, doi = {10.1190/tle37010069.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/TheLeadingEdge/2018/louboutin2017fwip2/louboutin2017fwip2.html}, author = {Mathias Louboutin and Philipp A. Witte and Michael Lange and Navjot Kukreja and Fabio Luporini and Gerard Gorman and Felix J. Herrmann} } @Conference{louboutin2018segeow, abstract = {In order to obtain accurate images of the subsurface, anisotropic modeling and imaging is necessary. However, the twenty-one parameter complete wave-equation is too computationally expensive to be of use in this case. The transverse tilted isotropic wave-equation is then the next best feasible representation of the physics to use for imaging. The main complexity arising from transverse tilted isotropic imaging is to model the receiver wavefield (back propagation of the data or data residual) for the imaging condition. Unlike the isotropic or the full physics wave-equations, the transverse tilted isotropic wave-equation is not not self-adjoint. This difference means that time-reversal will not model the correct receiver wavefield and this can lead to incorrect subsurface images. In this work, we derive and implement the adjoint wave-equation to demonstrate the necessity of exact adjoint modeling for anisotropic modeling and compare our result with adjoint-free time-reversed imaging.}, author = {Mathias Louboutin and Philipp A. Witte and Felix J. Herrmann}, booktitle = {SEG Technical Program Expanded Abstracts}, doi = {10.1190/segam2018-2996274.1}, keywords = {Adjoint, anisotropy, finite-differences, Imaging, RTM, SEG, TTI}, month = {10}, note = {(SEG, Anaheim)}, pages = {331-335}, title = {Effects of wrong adjoints for RTM in TTI media}, url = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2018/louboutin2018SEGeow/louboutin2018SEGeow.html}, year = {2018}, bdsk-url-1 = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2018/louboutin2018SEGeow/louboutin2018SEGeow.html}, bdsk-url-2 = {http://dx.doi.org/10.1190/segam2018-2996274.1} } @phdthesis{louboutin2020THmfi, title = {Modeling for inversion in exploration geophysics}, year = {2020}, note = {(PhD)}, month = {03}, school = {Georgia Institute of Technology}, type = {phd}, address = {Atlanta}, abstract = {Seismic inversion, and more generally geophysical exploration, aims at better understanding the earth{\textquoteright}s subsurface, which is one of today{\textquoteright}s most important challenges. Firstly, it contains natural resources that are critical to our technologies such as water, minerals and oil and gas. Secondly, monitoring the subsurface in the context of CO2 sequestration, earthquake detection and global seismology are of major interests with regard to safety and the environment hazards. However, the technologies to monitor the subsurface or find resources are scientifically extremely challenging. Seismic inversion can be formulated as a mathematical optimization problem that minimizes the difference between field recorded data and numerically modeled synthetic data. The process of solving this optimization problem then requires to numerically model, thousands of times, wave-propagation in large three-dimensional representations of part of the earth subsurface. The mathematical and computational complexity of this problem, therefore, calls for software design that abstracts these requirements and facilitates algorithm and software development. My thesis addresses some of the challenges that arise from these problems; mainly the computational cost and access to the right software for research and development. In the first part, I will discuss a performance metric that improves the current runtime-only benchmarks in exploration geophysics. This metric, the roofline model, first provides insight at the hardware level of the performance of a given implementation relative to the maximum achievable performance. Second, this study demonstrates that the choice of numerical discretization has a major impact on the achievable performance depending on the hardware at hand and shows that a flexible framework with respect to the discretization parameters is necessary. In the second part, I will introduce and describe Devito, a symbolic finite-difference DSL that provides a high-level interface to the definition of partial differential equations (PDE) such as the wave equation. Devito, from the symbolic definition of PDEs, then generates and compiles highly optimized C code on-the-fly to compute the solution of the PDE. The combination of the high-level abstractions and the just-in-time compiler enable research for geophysical exploration and PDE-constrainted optimization based on the paradigm of separation of concerns. This allows researchers to concentrate on their respective field of study while having access to computationally performant solvers with a flexible and easy to use interface to successfully implement complex representations of the physics. The second part of my thesis will be split into two sub-parts; first describing the symbolic application programming interface (API), before describing and benchmarking the just-in-time compiler. I will end my thesis with concluding remarks, the latest developments and a brief description of projects that were enabled by Devito.}, keywords = {finite-differences, FWI, HPC, Imaging, inversion, Modeling, performance, PhD, RTM}, url = {https://slim.gatech.edu/Publications/Public/Thesis/2020/louboutin2020THmfi/louboutin2020THmfi.pdf}, presentation = {https://slim.gatech.edu/Publications/Public/Thesis/2020/louboutin2020THmfi/louboutin2020THmfi_pres.pdf}, author = {Mathias Louboutin} } @article{ lumley20104, title={4D seismic monitoring of CO 2 sequestration}, author={Lumley, David}, journal={The Leading Edge}, volume={29}, number={2}, pages={150--155}, year={2010}, publisher={Society of Exploration Geophysicists} } @Article{luporini-coffee, acmid = {2687415}, address = {New York, NY, USA}, articleno = {57}, author = {Luporini, Fabio and Varbanescu, Ana Lucia and Rathgeber, Florian and Bercea, Gheorghe-Teodor and Ramanujam, J. and Ham, David A. and Kelly, Paul H. J.}, date-added = {2016-05-25 08:03:55 +0000}, date-modified = {2016-05-25 08:05:18 +0000}, doi = {10.1145/2687415}, issn = {1544-3566}, issue_date = {January 2015}, journal = {ACM Trans. Archit. Code Optim.}, keywords = {Finite element integration, SIMD vectorization, compilers, local assembly, optimizations}, month = jan, number = {4}, numpages = {25}, pages = {57:1--57:25}, publisher = {ACM}, title = {Cross-Loop Optimization of Arithmetic Intensity for Finite Element Local Assembly}, url = {http://doi.acm.org/10.1145/2687415}, volume = {11}, year = {2015}, bdsk-url-1 = {http://doi.acm.org/10.1145/2687415}, bdsk-url-2 = {http://dx.doi.org/10.1145/2687415} } @Article{luporini2015, author = {Luporini, Fabio and Varbanescu, Ana Lucia and Rathgeber, Florian and Bercea, Gheorghe-Teodor and Ramanujam, J. and Ham, David A. and Kelly, Paul H. J.}, title = {Cross-Loop Optimization of Arithmetic Intensity for Finite Element Local Assembly}, journal = {ACM Trans. Archit. Code Optim.}, issue_date = {January 2015}, volume = {11}, number = {4}, month = jan, year = {2015}, issn = {1544-3566}, pages = {57:1--57:25}, articleno = {57}, numpages = {25}, url = {http://doi.acm.org/10.1145/2687415}, doi = {10.1145/2687415}, acmid = {2687415}, publisher = {ACM}, address = {New York, NY, USA} } @Misc{ madagascar, author = {{Fomel et. al}, Sergey}, title = {Madagascar: open-source software project for multidimensional data analysis and reproducible computational experiments}, journal = {Journal of Open Research Software}, volume = 1, number = 1, year = {2013}, doi = {http://doi.org/10.5334/jors.ag} } @article{marmouelas, author = {Gary S. Martin and Robert Wiley and Kurt J. Marfurt}, title = {Marmousi2: An elastic upgrade for Marmousi}, journal = {The Leading Edge}, volume = {25}, number = {2}, pages = {156-166}, year = {2006}, doi = {10.1190/1.2172306}, URL = {https://doi.org/10.1190/1.2172306}, eprint = {https://doi.org/10.1190/1.2172306}, abstract = { The original Marmousi model was created by a consortium led by the Institut Français du Pétrole (IFP) in 1988. Since its creation, the model and its acoustic finite-difference synthetic data have been used by hundreds of researchers throughout the world for a multitude of geophysical purposes, and to this day remains one of the most published geophysical data sets. The advancement in computer hardware capabilities since the late 1980s has made it possible to perform a major upgrade to the model and data set, thereby extending the usefulness of the model for, hopefully, some time to come. This paper outlines the creation of an updated and upgraded Marmousi model and data set which we have named Marmousi2. } } @Article{mccalpin1995, abstract = {The ratio of cpu speed to memory speed in current high-performance computers is growing rapidly, with significant implications for the design and implementation of algorithms in scientific computing. I present the results of a broad survey of memory bandwidth and machine balance for a large variety of current computers, including uniprocessors, vector processors, shared-memory systems, and districuted-memory systems. The results are analyzed in terms of the sustainable data transfer rates for uncached unit-stride vector operation for each machine, and for each class.}, author = {John D. McCalpin}, journal = {IEEE Computer Society Technical Committee on Computer Architecture (TCCA) Newsletter}, month = dec, pages = {19--25}, pdf = {http://tab.computer.org/tcca/NEWS/DEC95/dec95_mccalpin.ps}, title = {Memory Bandwidth and Machine Balance in Current High Performance Computers}, year = {1995} } @TechReport{mccalpin2007, address = {Charlottesville, Virginia}, author = {John D. McCalpin}, institution = {University of Virginia}, note = {A continually updated technical report. http://www.cs.virginia.edu/stream/}, title = {STREAM: Sustainable Memory Bandwidth in High Performance Computers}, url = {http://www.cs.virginia.edu/stream/}, year = {1991-2007}, bdsk-url-1 = {http://www.cs.virginia.edu/stream/} } @Article{mcmechan, author = {McMechan, G. A.}, title = {MIGRATION BY EXTRAPOLATION OF TIME-DEPENDENT BOUNDARY VALUES}, journal = {Geophysical Prospecting}, volume = {31}, number = {3}, publisher = {Blackwell Publishing Ltd}, issn = {1365-2478}, url = {http://dx.doi.org/10.1111/j.1365-2478.1983.tb01060.x}, doi = {10.1111/j.1365-2478.1983.tb01060.x}, pages = {413--420}, year = {1983} } @Article{mint, acmid = {2225167}, address = {Piscataway, NJ, USA}, author = {Unat, Didem and Zhou, Jun and Cui, Yifeng and Baden, Scott B. and Cai, Xing}, date-added = {2016-04-16 10:17:14 +0000}, date-modified = {2016-04-16 10:17:21 +0000}, doi = {10.1109/MCSE.2012.44}, issn = {1521-9615}, issue_date = {May 2012}, journal = {Computing in Science and Engg.}, keywords = {Code generation, optimization, emerging technologies, Earth and atmospheric sciences, scientific computing}, month = may, number = {3}, numpages = {12}, pages = {48--59}, publisher = {IEEE Educational Activities Department}, title = {Accelerating a 3D Finite-Difference Earthquake Simulation with a C-to-CUDA Translator}, url = {http://dx.doi.org/10.1109/MCSE.2012.44}, volume = {14}, year = {2012}, bdsk-url-1 = {http://dx.doi.org/10.1109/MCSE.2012.44} } @Article{mittet, author = {Rune Mittet}, title = {Implementation of the Kirchhoff integral for elastic waves in staggered-grid modeling schemes}, journal = {GEOPHYSICS}, volume = {59}, number = {12}, pages = {1894-1901}, year = {1994}, doi = {10.1190/1.1443576}, url = { http://dx.doi.org/10.1190/1.1443576 }, eprint = { http://dx.doi.org/10.1190/1.1443576 } } @InProceedings{moser1954compiler, author = {Moser, Nora B}, booktitle = {Proc. Symp. on Automatic Programming for Digital Computers}, date-added = {2017-04-27 15:51:37 +0000}, date-modified = {2017-04-27 15:51:37 +0000}, title = {Compiler method of automatic programming}, year = {1954} } @article{ mulder2004, author = {Mulder, W. A. and Plessix, R.-E.}, title = {How to choose a subset of frequencies in frequency-domain finite-difference migration}, journal = {Geophysical Journal International}, volume = {158}, number = {3}, pages = {801-812}, year = {2004}, doi = {10.1111/j.1365-246X.2004.02336.x}, URL = {http://dx.doi.org/10.1111/j.1365-246X.2004.02336.x}, eprint = {/oup/backfile/content_public/journal/gji/158/3/10.1111/j.1365-246x.2004.02336.x/2/158-3-801.pdf} } @article{ mulder2008, author = {Mulder, W.A. and Plessix, R.-E.}, title = {Exploring some issues in acoustic full waveform inversion}, journal = {Geophysical Prospecting}, volume = {56}, number = {6}, publisher = {Blackwell Publishing Ltd}, issn = {1365-2478}, url = {http://dx.doi.org/10.1111/j.1365-2478.2008.00708.x}, urldate = {2018-07-27}, doi = {10.1111/j.1365-2478.2008.00708.x}, pages = {827--841}, year = {2008}, } @Misc{ notebook-bcs, author = {{Mathias Louboutin, Fabio Luporini}}, title = {{Boundary conditions in Devito}}, year = {\noop{2100}in preparation (2018)} } @article{officer, author = {Officer,C. B. and Morse,Philip M. }, title = {Introduction to the Theory of Sound Transmission}, journal = {Physics Today}, volume = {12}, number = {12}, pages = {66-67}, year = {1959}, doi = {10.1063/1.3060620}, URL = {https://doi.org/10.1063/1.3060620}, eprint = {https://doi.org/10.1063/1.3060620} } @Manual{openfoam, author = {The OpenFOAM Foundation}, title = {OpenFOAM v5 User Guide}, url = {https://cfd.direct/openfoam/user-guide/}, date = {March 2, 2017} } @Article{opensbli, author = {Christian T. Jacobs and Satya P. Jammy and Neil D. Sandham}, bibsource = {dblp computer science bibliography, http://dblp.org}, biburl = {http://dblp.uni-trier.de/rec/bib/journals/corr/JacobsJS16}, journal = {CoRR}, timestamp = {Mon, 03 Oct 2016 17:51:10 +0200}, title = {OpenSBLI: {A} framework for the automated derivation and parallel execution of finite difference solvers on a range of computer architectures}, url = {http://arxiv.org/abs/1609.01277}, volume = {abs/1609.01277}, year = {2016}, bdsk-url-1 = {http://arxiv.org/abs/1609.01277} } @Article{opertomod, author = {Hustedt, Bernhard and Operto, Stéphane and Virieux, Jean}, title = {A multi-level direct-iterative solver for seismic wave propagation modelling: space and wavelet approaches}, journal = {Geophysical Journal International}, volume = {155}, number = {3}, pages = {953-980}, year = {2003}, doi = {10.1111/j.1365-246X.2003.02098.x}, url = { + http://dx.doi.org/10.1111/j.1365-246X.2003.02098.x}, eprint = {/oup/backfile/content_public/journal/gji/155/3/10.1111/j.1365-246x.2003.02098.x/3/155-3-953.pdf} } @Misc{ opesci-project, author = {Gerard Gorman and Marcos de Aguiar and David Ham and Felix Herrmann and Christian Jacobs and Paul Kelly and Michael Lange and Chris Pain and Matthew Piggott and Spencer Sherwin and Felippe Vieira Zacarias and Mike Warner and Fabio Luporini and Navjot Kukreja}, date-added = {2016-06-23 12:42:32 +0000}, date-modified = {2016-07-07 09:31:00 +0000}, howpublished = {\url{http://www.opesci.org}}, title = {{OPESCI} project web page}, year = {2015} } @InProceedings{ops, acmid = {2691173}, address = {Piscataway, NJ, USA}, author = {Reguly, Istv\'{a}n Z. and Mudalige, Gihan R. and Giles, Michael B. and Curran, Dan and McIntosh-Smith, Simon}, booktitle = {Proceedings of the Fourth International Workshop on Domain-Specific Languages and High-Level Frameworks for High Performance Computing}, doi = {10.1109/WOLFHPC.2014.7}, isbn = {978-1-4799-7020-9}, keywords = {domain specific languages, high performance computing, software design, structured grid}, location = {New Orleans, Louisiana}, numpages = {10}, pages = {58--67}, publisher = {IEEE Press}, series = {WOLFHPC '14}, title = {The OPS Domain Specific Abstraction for Multi-block Structured Grid Computations}, url = {http://dx.doi.org/10.1109/WOLFHPC.2014.7}, year = {2014}, bdsk-url-1 = {http://dx.doi.org/10.1109/WOLFHPC.2014.7} } @InProceedings{osuna2014stella, author = {Osuna, Carlos and Fuhrer, Oliver and Gysi, Tobias and Bianco, Mauro}, booktitle = {Poster Presentation at the Platform for Advanced Scientific Computing (PASC) Conference, Zurich, Switzerland}, date-added = {2017-03-29 16:55:37 +0000}, date-modified = {2017-03-29 16:55:37 +0000}, title = {STELLA: A domain-specific language for stencil methods on structured grids}, year = {2014} } @article{osti_1468379, title = {Computation of Kernels for Full Waveform Seismic Inversion Using Parelasti.}, author = {Preston, Leiph}, abstractNote = {Full waveform inversion allows the seismologist to utilize an entire waveform and all the information it contains to help image the 3-D structure of the interior of the earth. This report summarizes the basic theory that has been developed in full waveform seismic inversion, primarily related to computation of sensitivity kernels. It then describes the implementation of this theory using Sandia Geophysics Department's Parelasti code, a 3-D full waveform elastic simulation algorithm. Finally, the code is validated using synthetics from simple homogeneous elastic earth models.}, doi = {10.2172/1468379}, place = {United States}, year = {2018}, month = {8} } @article{osti_1561580, title = {Paraniso 1.0: 3-D Full Waveform Seismic Simulation in General Anisotropic Media.}, author = {Preston, Leiph}, abstractNote = {Many geologic materials and minerals are seismically anisotropic, with the most general anisotropic material having up to 21 independent elastic coefficients. This report outlines the development of a 3-D, generally anisotropic, linear elastic full waveform finite- difference solver. First, a mathematical description of the solution equations will be described. The finite-difference implementation of these equations will then be shown. Finally, a comparison of results from this new solver to other solutions will be provided as verification that the new algorithm can accurately replicate these solutions. ACKNOWLEDGEMENTS The author also acknowledges the National Nuclear Security Administration, Defense Nuclear Nonproliferation Research and Development (DNN R&D), and the Source Physics Experiment (SPE) working group, a multi-institutional and interdisciplinary group of scientists and engineers.}, doi = {10.2172/1561580}, place = {United States}, year = {2019}, month = {9} } @article{osti_1561581, title = {ParelastiFWI 1.0 User Guide.}, author = {Preston, Leiph}, abstractNote = {ParelastiFWl is a python-based frontend to the seismic full waveform inversion process using Sandia Geophysics Department's 3-D isotropic elastic full waveform simulation code, Parelasti. The arguments one provides to ParelastiFWl guide the full waveform inversion process, including resolution of the inversion grid and basic regularization. This report outlines the user flags and ParelastiFWI usage to control the full waveform inversion procedure. ACKNOWLEDGEMENTS The author also acknowledges the National Nuclear Security Administration, Defense Nuclear Nonproliferation Research and Development (DNN R&D), and the Source Physics Experiment (SPE) working group, a multi-institutional and interdisciplinary group of scientists and engineers.}, doi = {10.2172/1561581}, place = {United States}, year = {2019}, month = {9} } @article{pandolfo, author = {Vincenzo Pandolfo}, title = {Investigating the {OPS} intermediate representation to target GPUs in the Devito {DSL}}, journal = {CoRR}, volume = {abs/1906.10811}, year = {2019}, url = {http://arxiv.org/abs/1906.10811}, archivePrefix = {arXiv}, eprint = {1906.10811}, timestamp = {Thu, 27 Jun 2019 18:54:51 +0200}, biburl = {https://dblp.org/rec/bib/journals/corr/abs-1906-10811}, bibsource = {dblp computer science bibliography, https://dblp.org} } @Book{ patterson, address = {San Francisco, CA, USA}, author = {Patterson, David A. and Hennessy, John L.}, edition = {3rd}, isbn = {0123706068, 9780123706065}, publisher = {Morgan Kaufmann Publishers Inc.}, title = {Computer Organization and Design: The Hardware/Software Interface}, year = {2007} } @InProceedings{patus, acmid = {2059499}, address = {Washington, DC, USA}, author = {Christen, Matthias and Schenk, Olaf and Burkhart, Helmar}, booktitle = {Proceedings of the 2011 IEEE International Parallel \& Distributed Processing Symposium}, date-added = {2016-04-16 10:21:24 +0000}, date-modified = {2016-04-16 10:21:29 +0000}, doi = {10.1109/IPDPS.2011.70}, isbn = {978-0-7695-4385-7}, numpages = {12}, pages = {676--687}, publisher = {IEEE Computer Society}, series = {IPDPS '11}, title = {PATUS: A Code Generation and Autotuning Framework for Parallel Iterative Stencil Computations on Modern Microarchitectures}, url = {http://dx.doi.org/10.1109/IPDPS.2011.70}, year = {2011}, bdsk-url-1 = {http://dx.doi.org/10.1109/IPDPS.2011.70} } @Article{pdel, author = {C{\'a}rdenas, Alfonso F and Karplus, Walter J}, journal = {Communications of the ACM}, number = {3}, pages = {184--191}, publisher = {ACM}, title = {PDEL---a language for partial differential equations}, volume = {13}, year = {1970} } @Article{pennycook2016metric, author = {Pennycook, Simon J and Sewall, JD and Lee, VW}, title = {A metric for performance portability}, journal = {arXiv preprint arXiv:1611.07409}, year = {2016} } @InProceedings{perfboundscodenarayanan, author = {Narayanan, Sri Hari Krishna and Norris, Boyana and Hovland, Paul D}, booktitle = {Parallel Processing Workshops (ICPPW), 2010 39th International Conference on}, date-added = {2017-03-04 18:41:18 +0000}, date-modified = {2017-03-04 18:41:47 +0000}, organization = {IEEE}, pages = {197--206}, title = {Generating performance bounds from source code}, year = {2010} } @InCollection{perls:3dfd, title = {Characterization and Optimization Methodology Applied to Stencil Computations}, author = {Cedric Andreolli and Philippe Thierry and Leo Borges and Gregg Skinner and Charles Yount}, chapter = {23}, year = {2016}, booktitle = {High Performance Parallelism Pearls, 1st Edition}, publisher = {Elsevier}, editor = {Reinders, James and Jeffers, James}, hidden_url = {http://store.elsevier.com/High-Performance-Parallelism-Pearls/James-Reinders/isbn-9780128021187/} } @Article{peters2016cvp, title = {Constraints versus penalties for edge-preserving full-waveform inversion}, journal = {The Leading Edge}, volume = {36}, number = {1}, year = {2017}, note = {(The Leading Edge)}, month = {01}, pages = {94-100}, abstract = {Full-waveform inversion is challenging in complex geologic areas. Even when provided with an accurate starting model, the inversion algorithms often struggle to update the velocity model. Compared with other areas in applied geophysics, including prior information in full-waveform inversion is still in its relative infancy. In part, this is due to the fact that it is difficult to incorporate prior information that relates to geologic settings where strong discontinuities in the velocity model dominate, because these settings call for nonsmooth regularizations. We tackle this problem by including constraints on the spatial variations and value ranges of the inverted velocities, as opposed to adding penalties to the objective, which is more customary in mainstream geophysical inversion. By demonstrating the lack of predictability of edge-preserving inversion when the regularization is in the form of an added penalty term, we advocate the inclusion of constraints instead. Our examples show that the latter leads to more predictable results and to significant improvements in the delineation of salt bodies when these constraints are relaxed gradually in combination with extending the search space to approximately fit the observed data but not the noise.}, keywords = {algorithm, full waveform inversion, noise, Optimization, penalty methods, salt dome, total variation}, doi = {10.1190/tle36010094.1}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Journals/TheLeadingEdge/2016/peters2016cvp/peters2016cvp.html}, author = {Bas Peters and Felix J. Herrmann} } @article{ peters2017, author = {Bas Peters and Felix J. Herrmann }, title = {Constraints versus penalties for edge-preserving full-waveform inversion}, journal = {The Leading Edge}, volume = {36}, number = {1}, pages = {94-100}, year = {2017}, doi = {10.1190/tle36010094.1}, URL = {http://dx.doi.org/10.1190/tle36010094.1}, eprint = {http://dx.doi.org/10.1190/tle36010094.1} } @article{ peters2019, author = {Bas Peters and Brendan R. Smithyman and Felix J. Herrmann}, title = {Projection methods and applications for seismic nonlinear inverse problems with multiple constraints}, journal = {GEOPHYSICS}, volume = {84}, number = {2}, pages = {R251-R269}, year = {2019}, doi = {10.1190/geo2018-0192.1}, URL = {https://doi.org/10.1190/geo2018-0192.1}, eprint = {https://doi.org/10.1190/geo2018-0192.1} } @Article{plessixasfwi, author = {Plessix, R.-E.}, doi = {10.1111/j.1365-246X.2006.02978.x}, issn = {1365-246X}, journal = {Geophysical Journal International}, keywords = {adjoint state, gradient, migration, tomography}, number = {2}, pages = {495--503}, publisher = {Blackwell Publishing Ltd}, title = {A review of the adjoint-state method for computing the gradient of a functional with geophysical applications}, url = {http://dx.doi.org/10.1111/j.1365-246X.2006.02978.x}, volume = {167}, year = {2006}, bdsk-url-1 = {http://dx.doi.org/10.1111/j.1365-246X.2006.02978.x} } @InProceedings{pluto, acmid = {1375595}, address = {New York, NY, USA}, author = {Bondhugula, Uday and Hartono, Albert and Ramanujam, J. and Sadayappan, P.}, booktitle = {Proceedings of the 2008 ACM SIGPLAN Conference on Programming Language Design and Implementation}, doi = {10.1145/1375581.1375595}, isbn = {978-1-59593-860-2}, keywords = {affine transformations, automatic parallelization, locality optimization, loop transformations, polyhedral model, tiling}, location = {Tucson, AZ, USA}, numpages = {13}, pages = {101--113}, publisher = {ACM}, series = {PLDI '08}, title = {A Practical Automatic Polyhedral Parallelizer and Locality Optimizer}, url = {http://doi.acm.org/10.1145/1375581.1375595}, year = {2008}, bdsk-url-1 = {http://doi.acm.org/10.1145/1375581.1375595}, bdsk-url-2 = {http://dx.doi.org/10.1145/1375581.1375595} } @Article{prattmod, author = {R. Gerhard Pratt}, title = {Seismic waveform inversion in the frequency domain, Part 1: Theory and verification in a physical scale model}, journal = {GEOPHYSICS}, volume = {64}, number = {3}, pages = {888-901}, year = {1999}, doi = {10.1190/1.1444597}, url = { https://doi.org/10.1190/1.1444597 }, eprint = { https://doi.org/10.1190/1.1444597 } } @Article{pyfr, title = "PyFR: An open source framework for solving advection–diffusion type problems on streaming architectures using the flux reconstruction approach", journal = "Computer Physics Communications", volume = "185", number = "11", pages = "3028 - 3040", year = "2014", issn = "0010-4655", doi = "https://doi.org/10.1016/j.cpc.2014.07.011", url = "http://www.sciencedirect.com/science/article/pii/S0010465514002549", author = "F.D. Witherden and A.M. Farrington and P.E. Vincent", keywords = "High-order, Flux reconstruction, Parallel algorithms, Heterogeneous computing" } @InProceedings{pyop2, author = {G.~R. Markall and F. Rathgeber and L. Mitchell and N. Loriant and C. Bertolli and D.~A. Ham and P.~H.~J.~ Kelly}, booktitle = {Proceedings of the International Supercomputing Conference (ISC) '13}, month = {June}, note = {In press}, series = {Lecture Notes in Computer Science}, title = {Performance portable finite element assembly using {PyOP2} and {FEniCS}}, volume = {7905}, year = {2013} } @Misc{ pysit2013, author = {Hewett, R.J. and Demanet, L. and the PySIT Team}, note = {Release 0.5}, title = {{PySIT}: {Python} Seismic Imaging Toolbox v0.5}, url = {http://www.pysit.org}, year = {2013}, bdsk-url-1 = {http://www.pysit.org} } @InProceedings{rahman2011understanding, author = {Rahman, Shah M Faizur and Yi, Qing and Qasem, Apan}, booktitle = {Proceedings of the 8th ACM International Conference on Computing Frontiers}, date-added = {2017-03-04 18:47:17 +0000}, date-modified = {2017-03-04 18:47:17 +0000}, organization = {ACM}, pages = {30}, title = {Understanding stencil code performance on multicore architectures}, year = {2011} } @Article{raknesr45, author = {Raknes, Espen Birger and Weibull, Wiktor}, title = {Efficient 3D elastic full-waveform inversion using wavefield reconstruction methods}, volume = {81}, number = {2}, pages = {R45--R55}, year = {2016}, doi = {10.1190/geo2015-0185.1}, publisher = {Society of Exploration Geophysicists}, abstract = {In reverse time migration (RTM) or full-waveform inversion (FWI), forward and reverse time propagating wavefields are crosscorrelated in time to form either the image condition in RTM or the misfit gradient in FWI. The crosscorrelation condition requires both fields to be available at the same time instants. For large-scale 3D problems, it is not possible, in practice, to store snapshots of the wavefields during forward modeling due to extreme storage requirements. We have developed an approximate wavefield reconstruction method that uses particle velocity field recordings on the boundaries to reconstruct the forward wavefields during the computation of the reverse time wavefields. The method is computationally effective and requires less storage than similar methods. We have compared the reconstruction method to a boundary reconstruction method that uses particle velocity and stress fields at the boundaries and to the optimal checkpointing method. We have tested the methods on a 2D vertical transversely isotropic model and a large-scale 3D elastic FWI problem. Our results revealed that there are small differences in the results for the three methods.}, issn = {0016-8033}, url = {http://geophysics.geoscienceworld.org/content/81/2/R45}, eprint = {http://geophysics.geoscienceworld.org/content/81/2/R45.full.pdf}, journal = {Geophysics} } @Book{ ref1, author = "Kernighan, Brian W. and Ritchie, Dennis M.", title = {The C Programming Language Second Edition}, publisher = "Prentice-Hall, Inc.", year = 1988 } @Book{ ref2, author = "George P. Burdell", title = {Myths and Their Origins Plus Some Extra Words to Make the Title Really Long and Extend to the Next Line}, publisher = "Real Books, Inc.", year = 2017 } @Article{ref3, title = {Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?}, author = {Einstein, A. and Podolsky, B. and Rosen, N.}, journal = {Phys. Rev.}, volume = {47}, issue = {10}, pages = {777--780}, numpages = {0}, year = {1935}, month = {5}, publisher = {American Physical Society}, doi = {10.1103/PhysRev.47.777}, url = {http://link.aps.org/doi/10.1103/PhysRev.47.777} } @InProceedings{rooflinemodeltoolkit, author = {Lo, Yu Jung and Williams, Samuel and Van Straalen, Brian and Ligocki, Terry J and Cordery, Matthew J and Wright, Nicholas J and Hall, Mary W and Oliker, Leonid}, booktitle = {International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems}, date-added = {2017-03-04 18:48:45 +0000}, date-modified = {2017-03-04 18:49:23 +0000}, organization = {Springer}, pages = {129--148}, title = {Roofline model toolkit: A practical tool for architectural and program analysis}, year = {2014} } @InProceedings{rosen1964programming, author = {Rosen, Saul}, booktitle = {Proceedings of the April 21-23, 1964, spring joint computer conference}, date-added = {2017-04-28 12:50:25 +0000}, date-modified = {2017-04-28 12:50:25 +0000}, organization = {ACM}, pages = {1--15}, title = {Programming systems and languages: a historical survey}, year = {1964} } @InProceedings{sato2009performance, author = {Sato, Yoshiei and Nagaoka, Ryuichi and Musa, Akihiro and Egawa, Ryusuke and Takizawa, Hiroyuki and Okabe, Koki and Kobayashi, Hiroaki}, booktitle = {Proceedings of the 10th workshop on MEmory performance: DEaling with Applications, systems and architecture}, date-added = {2017-02-09 13:43:05 +0000}, date-modified = {2017-02-09 13:43:05 +0000}, organization = {ACM}, pages = {7--14}, title = {Performance tuning and analysis of future vector processors based on the roofline model}, year = {2009} } @article{sava1, author = {Köhn, D. and Hellwig, O. and De Nil, D. and Rabbel, W.}, title = {Waveform inversion in triclinic anisotropic media—a resolution study}, journal = {Geophysical Journal International}, volume = {201}, number = {3}, pages = {1642-1656}, year = {2015}, month = {04}, abstract = {Seismic full waveform inversion (FWI) has been applied to simple elastic problems with certain symmetries, such as isotropic, transverse isotropic or vertical transversely isotropic media. In this study, the FWI concept is extended to the most general anisotropic case with 21 independent elastic material parameters and no symmetry plane (triclinic). Beside a short description of the 3-D finite-difference scheme to solve the forward problem and the FWI optimization algorithm, we present a sensitivity study for a simple anisotropic medium. This test problem consists of a homogenous triclinic anisotropic full space, which contains 21 spatially separated spheres. In each sphere one component of the elastic tensor deviates by 5 per cent from the background medium. The resolution of the different spheres, ambiguities between the different elastic parameters, as well as the effect of the acquisition geometry can be systematically investigated. Due to the high computational costs of the triclinic forward problem a few compromises have to be made regarding the acquisition geometries. Point sources are replaced by plane wave sources which lead to a limitation of incidence angles and therefore a strong decrease in resolution of the nondiagonal elastic tensor components. It is shown that, despite these limitations, a tomographic acquisition geometry would be able to resolve to some extent a monoclinic symmetry via FWI. Restricting the acquisition geometries (e.g. VSP combined with reflection seismic or reflection seismic only) significantly reduces the number of resolvable tensor elements in strict dependence of the covered incidence angles.}, issn = {0956-540X}, doi = {10.1093/gji/ggv097}, url = {https://doi.org/10.1093/gji/ggv097}, eprint = {https://academic.oup.com/gji/article-pdf/201/3/1642/17367789/ggv097.pdf}, } @article{sava2, title = "Rasterizing geological models for parallel finite difference simulation using seismic simulation as an example", journal = "Computers \& Geosciences", volume = "86", pages = "83 - 91", year = "2016", issn = "0098-3004", doi = "https://doi.org/10.1016/j.cageo.2015.10.008", url = "http://www.sciencedirect.com/science/article/pii/S009830041530073X", author = "Björn Zehner and Olaf Hellwig and Maik Linke and Ines Görz and Stefan Buske", keywords = "3D, Rasterization, Voxelization, Scan conversion, Seismic, Finite difference simulation, Parallel computation", abstract = "3D geological underground models are often presented by vector data, such as triangulated networks representing boundaries of geological bodies and geological structures. Since models are to be used for numerical simulations based on the finite difference method, they have to be converted into a representation discretizing the full volume of the model into hexahedral cells. Often the simulations require a high grid resolution and are done using parallel computing. The storage of such a high-resolution raster model would require a large amount of storage space and it is difficult to create such a model using the standard geomodelling packages. Since the raster representation is only required for the calculation, but not for the geometry description, we present an algorithm and concept for rasterizing geological models on the fly for the use in finite difference codes that are parallelized by domain decomposition. As a proof of concept we implemented a rasterizer library and integrated it into seismic simulation software that is run as parallel code on a UNIX cluster using the Message Passing Interface. We can thus run the simulation with realistic and complicated surface-based geological models that are created using 3D geomodelling software, instead of using a simplified representation of the geological subsurface using mathematical functions or geometric primitives. We tested this set-up using an example model that we provide along with the implemented library." } @InProceedings{schmidtbergfriedlandermurphy:2009, address = {Clearwater Beach, Florida}, author = {M. Schmidt and E. van den Berg and M. P. Friedlander and K. Murphy}, booktitle = {Proceedings of The Twelfth International Conference on Artificial Intelligence and Statistics (AISTATS) 2009}, editor = {D. van Dyk and M. Welling}, month = {April}, pages = {456-463}, title = {Optimizing Costly Functions with Simple Constraints: A Limited-Memory Projected Quasi-Newton Algorithm}, volume = 5, year = 2009 } @InProceedings{schmitt2014exaslang, author = {Schmitt, Christian and Kuckuk, Sebastian and Hannig, Frank and K{\"o}stler, Harald and Teich, J{\"u}rgen}, booktitle = {Proceedings of the Fourth International Workshop on Domain-Specific Languages and High-Level Frameworks for High Performance Computing}, date-added = {2017-04-13 11:56:24 +0000}, date-modified = {2017-04-13 11:56:24 +0000}, organization = {IEEE Press}, pages = {42--51}, title = {Exaslang: a domain-specific language for highly scalable multigrid solvers}, year = {2014} } @article{segy, author = {Barry, K. M. and Cavers, D. A. and Kneale, C. W.}, title = {Recommended standards for digital tape formats}, journal = {Geophysics}, volume = {40}, number = {2}, pages = {344-352}, year = {1975}, month = {04}, abstract = {Recently a new demand for demultiplexed formats has arisen in the seismic industry due to the utilization of minicomputers in digital field recording systems and because of a growing need to standardize an acceptable data exchange format.}, issn = {0016-8033}, doi = {10.1190/1.1440530}, url = {https://doi.org/10.1190/1.1440530}, eprint = {https://pubs.geoscienceworld.org/geophysics/article-pdf/40/2/344/3156872/344.pdf}, } @InProceedings{seismic-sims-on-knl:tobin-isc17, author = {Josh Tobin and Alexander Breuer and Alexander Heinecke and Charles Yount and Yifeng Cui}, title = {Accelerating Seismic Simulations using the Intel Xeon Phi Knights Landing Processor}, booktitle = {Proceedings of ISC High Performance 2017 (ISC17)}, month = {to appear}, year = {2017}, location = {Frankfurt, Germany} } @presentation{sevendwarfs, Author = {Phillip Colella}, Date-Added = {2017-05-09 18:31:43 +0000}, Date-Modified = {2017-05-09 18:31:43 +0000}, Journal = {presentation}, Title = {Defining Software Requirements for Scientific Computing}, Year = {2004} } @misc{shahbp2007tti, Author = {H. Shah}, Journal = {presentation}, Title = {The 2007 BP anisotropic velocity-analysis benchmark (presented at the 70th EAGE annual meeting, workshop)}, Year = {2007} } @Article{shell1959share, author = {Shell, Donald L}, date-added = {2017-04-28 12:52:28 +0000}, date-modified = {2017-04-28 12:52:28 +0000}, journal = {Journal of the ACM (JACM)}, number = {2}, pages = {123--127}, publisher = {ACM}, title = {The SHARE 709 system: a cooperative effort}, volume = {6}, year = {1959} } @book{ sheriff1995, title={Exploration seismology}, author={Sheriff, Robert E and Geldart, Lloyd P}, year={1995}, publisher={Cambridge University Press} } @inproceedings{sheriff1991encyclopedic, title={Encyclopedic dictionary of exploration geophysics}, author={Robert E. Sheriff}, year={1991} } @article {siahkoohi2019itl, title = {The importance of transfer learning in seismic modeling and imaging}, journal = {Geophysics}, year = {2019}, note = {(Accepted in GEOPHYSICS)}, abstract = {Accurate forward modeling is essential for solving inverse problems in exploration seismology. Unfortunately, it is often not possible to afford being physically or numerically accurate. To overcome this conundrum, we make use of raw and processed data from nearby surveys. We propose to use this data, consisting of shot records or velocity models, to pre-train a neural network to correct for the effects of, for instance, the free surface or numerical dispersion, both of which can be considered as proxies for incomplete or inaccurate physics. Given this pre-trained neural network, we apply transfer learning to finetune this pre-trained neural network so it performs well on its task of mapping low-cost, but low-fidelity, solutions to high-fidelity solutions for the current survey. As long as we can limit ourselves during finetuning to using only a small fraction of high-fidelity data, we gain processing the current survey while using information from nearby surveys. We demonstrate this principle by removing surface-related multiples and ghosts from shot records and the effects of numerical dispersion from migrated images and wave simulations}, keywords = {deep learning, Imaging, Modeling, SRME, transfer learning}, doi = {10.1190/geo2019-0056.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/Geophysics/2019/siahkoohi2019itl/siahkoohi2019itl.html}, author = {Ali Siahkoohi and Mathias Louboutin and Felix J. Herrmann} } @Article{simflowny, author = {Arbona, A and Mi{\~n}ano, B and Rigo, A and Bona, C and Palenzuela, C and Artigues, A and Bona-Casas, C and Mass{\'o}, J}, date-added = {2017-03-29 14:40:26 +0000}, date-modified = {2017-03-29 14:40:33 +0000}, journal = {arXiv preprint arXiv:1702.04715}, title = {Simflowny 2: An upgraded platform for scientific modeling and simulation}, year = {2017} } @article{sirgue, author = "Sirgue, L. and I. Barkved, O. and P. Van Gestel, J. and J. Askim, O. and H. Kommedal, J.", title = "3D Waveform Inversion on Valhall Wide-azimuth OBC", year = "2009", doi = "https://doi.org/10.3997/2214-4609.201400395", url = "https://www.earthdoc.org/content/papers/10.3997/2214-4609.201400395", publisher = "European Association of Geoscientists & Engineers", issn = "2214-4609", eid = "cp-127-00388", } @InProceedings{stargates, address = {Las Vegas, USA}, author = {K. A. Hawick and D. P. Playne}, booktitle = {11th International Conference on Software Engineering Research and Practice (SERP'13)}, institution = {Computer Science, Massey University, Auckland, New Zealand}, keywords = {generative programming; DSL; partial differential equation}, month = {22-25 July}, number = {CSTN-187}, owner = {kahawick}, pages = {SER3829}, publisher = {WorldComp}, timestamp = {2013.04.15}, title = {Simulation Software Generation using a Domain-Specific Language for Partial Differential Field Equations}, year = {2013} } @Article{stencil-tiling-fgcs-2017, title = "Multi-level spatial and temporal tiling for efficient HPC stencil computation on many-core processors with large shared caches", journal = "Future Generation Computer Systems", publisher = elsevier, year = "2017", issn = "0167-739X", doi = "https://doi.org/10.1016/j.future.2017.10.041", url = "http://www.sciencedirect.com/science/article/pii/S0167739X17304648", author = "Charles Yount and Alejandro Duran and Josh Tobin", keywords = "Finite-difference method, Seismic modeling, Intel Xeon Phi, Temporal wave-front tiling, Vector-folding" } @Proceedings{stencilcomp, author = {Kaushik Datta and Mark Murphy and Vasily Volkov and Samuel Williams and Jonathan Carter and Leonid Oliker and David A. Patterson and John Shalf and Katherine A. Yelick}, bibsource = {dblp computer science bibliography, http://dblp.org}, biburl = {http://dblp2.uni-trier.de/rec/bib/conf/sc/DattaMVWCOPSY08}, booktitle = {Proceedings of the {ACM/IEEE} Conference on High Performance Computing, {SC} 2008, November 15-21, 2008, Austin, Texas, {USA}}, date-modified = {2017-03-27 13:59:27 +0000}, doi = {10.1145/1413370.1413375}, pages = {4}, timestamp = {Tue, 30 Jun 2015 16:34:04 +0200}, title = {Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures}, url = {http://doi.acm.org/10.1145/1413370.1413375}, year = {2008}, bdsk-url-1 = {http://doi.acm.org/10.1145/1413370.1413375}, bdsk-url-2 = {http://dx.doi.org/10.1145/1413370.1413375} } @InProceedings{stencildsl-1, acmid = {2467268}, address = {New York, NY, USA}, author = {Henretty, Tom and Veras, Richard and Franchetti, Franz and Pouchet, Louis-Noel and Ramanujam, J. and Sadayappan, P.}, booktitle = {Proceedings of the 27th International ACM Conference on International Conference on Supercomputing}, doi = {10.1145/2464996.2467268}, isbn = {978-1-4503-2130-3}, keywords = {dsl, multicore, simd, split tiling, stencils}, location = {Eugene, Oregon, USA}, numpages = {12}, pages = {13--24}, publisher = {ACM}, series = {ICS '13}, title = {A Stencil Compiler for Short-vector SIMD Architectures}, url = {http://doi.acm.org/10.1145/2464996.2467268}, year = {2013}, bdsk-url-1 = {http://doi.acm.org/10.1145/2464996.2467268}, bdsk-url-2 = {http://dx.doi.org/10.1145/2464996.2467268} } @article{stolk2013, title = "A rapidly converging domain decomposition method for the Helmholtz equation", journal = "Journal of Computational Physics", volume = "241", pages = "240 - 252", year = "2013", issn = "0021-9991", doi = "https://doi.org/10.1016/j.jcp.2013.01.039", url = "http://www.sciencedirect.com/science/article/pii/S0021999113000922", author = "Christiaan C. Stolk", keywords = "Helmholtz equation, Domain decomposition, Perfectly matched layers, High-frequency waves", abstract = "A new domain decomposition method is introduced for the heterogeneous 2-D and 3-D Helmholtz equations. Transmission conditions based on the perfectly matched layer (PML) are derived that avoid artificial reflections and match incoming and outgoing waves at the subdomain interfaces. We focus on a subdivision of the rectangular domain into many thin subdomains along one of the axes, in combination with a certain ordering for solving the subdomain problems and a GMRES outer iteration. When combined with multifrontal methods, the solver has near-linear cost in examples, due to very small iteration numbers that are essentially independent of problem size and number of subdomains. It is to our knowledge only the second method with this property next to the moving PML sweeping method." } @article{stolk2014, author = {Stolk, Christiaan C. and Ahmed, Mostak. and Bhowmik, Samir Kumar.}, title = {A Multigrid Method for the Helmholtz Equation with Optimized Coarse Grid Corrections}, journal = {SIAM Journal on Scientific Computing}, volume = {36}, number = {6}, pages = {A2819-A2841}, year = {2014}, doi = {10.1137/13092349X}, URL = { https://doi.org/10.1137/13092349X }, eprint = { https://doi.org/10.1137/13092349X } } @InProceedings{strout-common-problems, author = {Michelle Mills Strout}, booktitle = {Proceedings of Workshop on Optimizing Stencil Computations (WOSC)}, date-added = {2016-03-30 16:00:04 +0000}, date-modified = {2016-03-30 16:00:18 +0000}, month = {October 27,}, title = {Compilers for Regular and Irregular Stencils: Some Shared Problems and Solutions}, year = {2013} } @TechReport{sun2010iwave, author = {Sun, Dong and Symes, William W}, institution = {Tech. Rep. 10-06, Department of Computational and Applied Mathematics, Rice University, Houston, Texas, USA}, title = {IWAVE implementation of adjoint state method}, url = {https://pdfs.semanticscholar.org/6c17/cfe41b76f6b745c435891ea6ba6f4e2c2dbf.pdf}, year = {2010}, bdsk-url-1 = {https://pdfs.semanticscholar.org/6c17/cfe41b76f6b745c435891ea6ba6f4e2c2dbf.pdf} } @Article{symes2007, author = {Symes}, doi = {10.1190/1.2742686}, eprint = {http://library.seg.org/doi/pdf/10.1190/1.2742686}, journal = {GEOPHYSICS}, number = {5}, pages = {SM213-SM221}, title = {Reverse Time Migration with Optimal Checkpointing}, url = {http://library.seg.org/doi/abs/10.1190/1.2742686}, volume = {72}, year = {2007}, bdsk-url-1 = {http://library.seg.org/doi/abs/10.1190/1.2742686}, bdsk-url-2 = {http://dx.doi.org/10.1190/1.2742686} } @Article{symes2015acoustic, author = {Symes, William W}, journal = {THE RICE INVERSION PROJECT}, pages = {141}, title = {Acoustic Staggered Grid Modeling in IWAVE}, url = {http://www.trip.caam.rice.edu/reports/2014/book.pdf#page=145}, year = {2015}, bdsk-url-1 = {http://www.trip.caam.rice.edu/reports/2014/book.pdf#page=145} } @Article{symes2015iwave, author = {Symes, William W}, journal = {THE RICE INVERSION PROJECT}, pages = {85}, title = {IWAVE structure and basic use cases}, url = {http://www.trip.caam.rice.edu/reports/2014/book.pdf#page=89}, year = {2015}, bdsk-url-1 = {http://www.trip.caam.rice.edu/reports/2014/book.pdf#page=89} } @InProceedings{ symes2017, author = {{Symes}, William W.}, title = {The Search for a Cycle-Skipping Cure: {An} overview}, booktitle = {Institute for Pure and Applied Mathematics (IPAM): Computational Issues in Oil Field Applications}, year = {2017}, urldate = {2018-07-27} } @Article{sympy, title = {SymPy: symbolic computing in Python}, author = {Meurer, Aaron and Smith, Christopher P. and Paprocki, Mateusz and \v{C}ert\'{i}k, Ond\v{r}ej and Kirpichev, Sergey B. and Rocklin, Matthew and Kumar, AMiT and Ivanov, Sergiu and Moore, Jason K. and Singh, Sartaj and Rathnayake, Thilina and Vig, Sean and Granger, Brian E. and Muller, Richard P. and Bonazzi, Francesco and Gupta, Harsh and Vats, Shivam and Johansson, Fredrik and Pedregosa, Fabian and Curry, Matthew J. and Terrel, Andy R. and Rou\v{c}ka, \v{S}t\v{e}p\'{a}n and Saboo, Ashutosh and Fernando, Isuru and Kulal, Sumith and Cimrman, Robert and Scopatz, Anthony}, year = 2017, month = jan, keywords = {Python, Computer algebra system, Symbolics}, abstract = { SymPy is an open source computer algebra system written in pure Python. It is built with a focus on extensibility and ease of use, through both interactive and programmatic applications. These characteristics have led SymPy to become a popular symbolic library for the scientific Python ecosystem. This paper presents the architecture of SymPy, a description of its features, and a discussion of select submodules. The supplementary material provide additional examples and further outline details of the architecture and features of SymPy. }, volume = 3, pages = {e103}, journal = {PeerJ Computer Science}, issn = {2376-5992}, url = {https://doi.org/10.7717/peerj-cs.103}, doi = {10.7717/peerj-cs.103} } @Article{sympy-bib, title = {SymPy: symbolic computing in Python}, author = {Meurer, Aaron and Smith, Christopher P. and Paprocki, Mateusz and \v{C}ert\'{i}k, Ond\v{r}ej and Kirpichev, Sergey B. and Rocklin, Matthew and Kumar, AMiT and Ivanov, Sergiu and Moore, Jason K. and Singh, Sartaj and Rathnayake, Thilina and Vig, Sean and Granger, Brian E. and Muller, Richard P. and Bonazzi, Francesco and Gupta, Harsh and Vats, Shivam and Johansson, Fredrik and Pedregosa, Fabian and Curry, Matthew J. and Terrel, Andy R. and Rou\v{c}ka, \v{S}t\v{e}p\'{a}n and Saboo, Ashutosh and Fernando, Isuru and Kulal, Sumith and Cimrman, Robert and Scopatz, Anthony}, year = 2017, month = jan, volume = 3, pages = {e103}, journal = {PeerJ Computer Science}, issn = {2376-5992}, url = {https://doi.org/10.7717/peerj-cs.103}, doi = {10.7717/peerj-cs.103} } @InProceedings{tang2011pochoir, author = {Tang, Yuan and Chowdhury, Rezaul Alam and Kuszmaul, Bradley C and Luk, Chi-Keung and Leiserson, Charles E}, booktitle = {Proceedings of the twenty-third annual ACM symposium on Parallelism in algorithms and architectures}, date-added = {2017-03-29 14:41:10 +0000}, date-modified = {2017-03-29 14:41:10 +0000}, organization = {ACM}, pages = {117--128}, title = {The pochoir stencil compiler}, year = {2011} } @Article{tarantola, author = {Tarantola, Albert}, title = {Inversion of seismic reflection data in the acoustic approximation}, journal = {GEOPHYSICS}, volume = {49}, number = {8}, pages = {1259}, year = {1984}, doi = {10.1190/1.1441754}, url = { + http://dx.doi.org/10.1190/1.1441754}, eprint = {/gsw/content_public/journal/geophysics/49/8/10.1190_1.1441754/5/1259.pdf} } @Article{taylor1960flow, author = {Taylor, Alan}, date-added = {2017-04-28 13:43:20 +0000}, date-modified = {2017-04-28 13:43:20 +0000}, journal = {Annual Review in Automatic Programming}, pages = {196--206}, publisher = {Elsevier}, title = {The FLOW-MATIC and MATH-MATIC automatic programming systems}, volume = {1}, year = {1960} } @Article{thomsen1986, author = {{Thomsen}, Leon}, title = {Weak elastic anisotropy}, journal = {Geophysics}, year = 1986, month = {october}, volume = 51, number = 10, pages = {1964-1966} } @article{thorbecke, author = {Jan W. Thorbecke and Deyan Draganov}, title = {Finite-difference modeling experiments for seismic interferometry}, journal = {GEOPHYSICS}, volume = {76}, number = {6}, pages = {H1-H18}, year = {2011}, doi = {10.1190/geo2010-0039.1}, URL = { https://doi.org/10.1190/geo2010-0039.1}, eprint = {https://doi.org/10.1190/geo2010-0039.1}, abstract = { In passive seismic interferometry, new reflection data can be retrieved by crosscorrelating recorded noise data. The quality of the retrieved reflection data is, among others, dependent on the duration and number of passive sources present during the recording time, the source distribution, and the source strength. To investigate these relations we set up several numerical modeling studies. To carry out the modeling in a feasible time, we design a finite-difference algorithm for the simulation of long-duration passive seismic measurements of band-limited noise signatures in the subsurface. Novel features of the algorithm include the modeling of thousands of randomly placed sources during one modeling run. The modeling experiments explore the dependency relation between the retrieved reflections and source-signature length, source positions, number of sources, and source amplitude variations. From these experiments we observed that the positions of the passive sources and the length of the source signals are of direct influence on the quality of the retrieved reflections. Random amplitude variations among source signals do not seem to have a big impact on the retrieved reflections. } } @Article{tle1, author = { Mathias Louboutin and Philipp Witte and Michael Lange and Navjot Kukreja and Fabio Luporini and Gerard Gorman and Felix J. Herrmann }, title = {Full-waveform inversion, Part 1: Forward modeling}, journal = {The Leading Edge}, volume = {36}, number = {12}, pages = {1033-1036}, year = {2017}, doi = {10.1190/tle36121033.1}, url = {https://doi.org/10.1190/tle36121033.1}, eprint = {https://doi.org/10.1190/tle36121033.1} } @Article{tti-main, author = {Yu Zhang and Houzhu Zhang and Guanquan Zhang}, doi = {10.1190/1.3554411}, eprint = {http://dx.doi.org/10.1190/1.3554411}, journal = {GEOPHYSICS}, number = {3}, pages = {WA3-WA11}, title = {A stable TTI reverse time migration and its implementation}, url = {http://dx.doi.org/10.1190/1.3554411}, volume = {76}, year = {2011}, bdsk-url-1 = {http://dx.doi.org/10.1190/1.3554411} } @Conference{tu2013eagelsm, abstract = {The advent of modern computing has made it possible to do seismic imaging using least-squares reverse-time migration. We obtain superior images by solving an optimization problem that recovers the true-amplitude images. However, its success hinges on overcoming several issues, including overwhelming problem size, unknown source wavelet, and interfering coherent events like multiples. In this abstract, we reduce the problem size by using ideas from compressive sensing, and estimate source wavelet by generalized variable projection. We also demonstrate how to invert for subsurface information encoded in surface-related multiples by incorporating the free-surface operator as an areal source in reverse-time migration. Our synthetic examples show that multiples help to improve the resolution of the image, as well as remove the amplitude ambiguity in wavelet estimation.}, author = {Ning Tu and Aleksandr Y. Aravkin and Tristan van Leeuwen and Felix J. Herrmann}, booktitle = {EAGE Annual Conference Proceedings}, doi = {10.3997/2214-4609.20130727}, keywords = {EAGE, Imaging, multiples, source estimation, sparse}, month = {06}, presentation = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/EAGE/2013/tu2013EAGElsm/tu2013EAGElsm_pres.pdf}, title = {Fast least-squares migration with multiples and source estimation}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/EAGE/2013/tu2013EAGElsm/tu2013EAGElsm.pdf}, year = {2013}, bdsk-url-1 = {https://www.slim.eos.ubc.ca/Publications/Public/Conferences/EAGE/2013/tu2013EAGElsm/tu2013EAGElsm.pdf}, bdsk-url-2 = {http://dx.doi.org/10.3997/2214-4609.20130727} } @Article{ufl, title = {{U}nified {F}orm {L}anguage: a domain-specific language for weak formulations of partial differential equations}, author = {Aln{\ae}s, Martin S. and Logg, Anders and {\O}lgaard, Kristian B. and Rognes, Marie E. and Wells, Garth N.}, journal = {ACM Transactions on Mathematical Software (TOMS)}, volume = {40}, number = {2}, pages = {9}, year = {2014}, publisher = {ACM} } @InProceedings{unat2011mint, author = {Unat, Didem and Cai, Xing and Baden, Scott B}, booktitle = {Proceedings of the international conference on Supercomputing}, date-added = {2017-03-29 16:53:32 +0000}, date-modified = {2017-03-29 16:53:32 +0000}, organization = {ACM}, pages = {214--224}, title = {Mint: realizing CUDA performance in 3D stencil methods with annotated C}, year = {2011} } @Article{vanleeuwen2015ippmp, title = {A penalty method for {PDE}-constrained optimization in inverse problems}, journal = {Inverse Problems}, volume = {32}, number = {1}, year = {2015}, note = {(Inverse Problems)}, month = {12}, pages = {015007}, abstract = {Many inverse and parameter estimation problems can be written as PDE-constrained optimization problems. The goal is to infer the parameters, typically coefficients of the PDE, from partial measurements of the solutions of the PDE for several right-hand sides. Such PDE-constrained problems can be solved by finding a stationary point of the Lagrangian, which entails simultaneously updating the parameters and the (adjoint) state variables. For large-scale problems, such an all-at-once approach is not feasible as it requires storing all the state variables. In this case one usually resorts to a reduced approach where the constraints are explicitly eliminated (at each iteration) by solving the PDEs. These two approaches, and variations thereof, are the main workhorses for solving PDE-constrained optimization problems arising from inverse problems. In this paper, we present an alternative method that aims to combine the advantages of both approaches. Our method is based on a quadratic penalty formulation of the constrained optimization problem. By eliminating the state variable, we develop an efficient algorithm that has roughly the same computational complexity as the conventional reduced approach while exploiting a larger search space. Numerical results show that this method indeed reduces some of the nonlinearity of the problem and is less sensitive to the initial iterate.}, keywords = {Inverse problems, Optimization, PDE, penalty method}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Journals/InverseProblems/2015/vanleeuwen2015IPpmp/vanleeuwen2015IPpmp.pdf}, url2 = {http://stacks.iop.org/0266-5611/32/i=1/a=015007}, author = {Tristan van Leeuwen and Felix J. Herrmann} } @Article{versteeg927, author = {Versteeg, Roelof}, eprint = {http://tle.geoscienceworld.org/content/13/9/927.full.pdf}, issn = {1070-485X}, journal = {The Leading Edge}, number = {9}, pages = {927--936}, publisher = {GeoScienceWorld}, title = {The Marmousi experience; velocity model determination on a synthetic complex data set}, url = {http://tle.geoscienceworld.org/content/13/9/927}, volume = {13}, year = {1994}, bdsk-url-1 = {http://tle.geoscienceworld.org/content/13/9/927} } @Article{virieux, author = {J. Virieux and S. Operto}, doi = {10.1190/1.3238367}, eprint = {http://library.seg.org/doi/pdf/10.1190/1.3238367}, journal = {GEOPHYSICS}, number = {5}, pages = {WCC1-WCC26}, title = {An overview of full-waveform inversion in exploration geophysics}, url = {http://library.seg.org/doi/abs/10.1190/1.3238367}, volume = {74}, year = {2009}, bdsk-url-1 = {http://library.seg.org/doi/abs/10.1190/1.3238367}, bdsk-url-2 = {http://dx.doi.org/10.1190/1.3238367} } @Article{virieux86, author = {Jean Virieux}, title = {P-SV wave propagation in heterogeneous media: Velocity-stress finite-difference method}, journal = {GEOPHYSICS}, volume = {51}, number = {4}, pages = {889-901}, year = {1986}, doi = {10.1190/1.1442147}, url = {https://doi.org/10.1190/1.1442147}, eprint = {https://doi.org/10.1190/1.1442147} } @Article{virieuxmodelling, author = {Jean Virieux and Stéphane Operto and Hafedh Ben-Hadj-Ali and Romain Brossier and Vincent Etienne and Florent Sourbier and Luc Giraud and Azzam Haidar}, title = {Seismic wave modeling for seismic imaging}, journal = {The Leading Edge}, volume = {28}, number = {5}, pages = {538-544}, year = {2009}, doi = {10.1190/1.3124928}, url = { https://doi.org/10.1190/1.3124928 }, eprint = { https://doi.org/10.1190/1.3124928 } } @Article{virieuxps, author = {Jean Virieux}, title = {P-SV wave propagation in heterogeneous media: Velocity‐stress finite‐difference method}, journal = {GEOPHYSICS}, volume = {51}, number = {4}, pages = {889-901}, year = {1986}, doi = {10.1190/1.1442147}, url = { https://doi.org/10.1190/1.1442147 }, eprint = { https://doi.org/10.1190/1.1442147 } } @Misc{ vtune, author = {{Intel Corporation}}, date-added = {2016-06-03 13:09:39 +0000}, date-modified = {2016-06-03 13:11:09 +0000}, howpublished = {https://software.intel.com/en-us/intel-vtune-amplifier-xe}, title = {{Intel VTune Performance Analyzer}}, year = {2016} } @InProceedings{wahib2014scalable, author = {Wahib, Mohamed and Maruyama, Naoya}, booktitle = {Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis}, date-added = {2017-02-09 15:28:17 +0000}, date-modified = {2017-03-06 14:03:03 +0000}, organization = {IEEE Press}, pages = {191--202}, title = {Scalable kernel fusion for memory-bound {GPU} applications}, year = {2014} } @Article{wang1090158, author = {Wang, Siyang and Nissen, Anna and Kreiss, Gunilla}, institution = {Uppsala University, Numerical Analysis}, journal = {Computing Research Repository}, number = {1702.01383}, title = {Convergence of finite difference methods for the wave equation in two space dimensions}, url = {https://arxiv.org/abs/1702.01383}, year = {2017} } @Article{wason2016geopctl, title = {Low-cost time-lapse seismic with distributed compressive sensing{\textendash}-{Part} 2: impact on repeatability}, journal = {Geophysics}, volume = {82}, number = {3}, year = {2017}, note = {(Geophysics)}, month = {05}, pages = {P15-P30}, abstract = {Irregular or off-the-grid spatial sampling of sources and receivers is inevitable in field seismic acquisitions. Consequently, time-lapse surveys become particularly expensive since current practices aim to replicate densely sampled surveys for monitoring changes occurring in the reservoir due to hydrocarbon production. We demonstrate that under certain circumstances, high-quality prestack data can be obtained from cheap randomized subsampled measurements that are observed from nonreplicated surveys. We extend our time-jittered marine acquisition to time-lapse surveys by designing acquisition on irregular spatial grids that render simultaneous, subsampled and irregular measurements. Using the fact that different time-lapse data share information and that nonreplicated surveys add information when prestack data are recovered jointly, we recover periodic densely sampled and colocated prestack data by adapting the recovery method to incorporate a regularization operator that maps traces from an irregular spatial grid to a regular periodic grid. The recovery method is, therefore, a combined operation of regularization, interpolation (estimating missing fine-grid traces from subsampled coarse-grid data), and source separation (unraveling overlapping shot records). By relaxing the insistence on replicability between surveys, we find that recovery of the time-lapse difference shows little variability for realistic field scenarios of slightly nonreplicated surveys that suffer from unavoidable natural deviations in spatial sampling of shots (or receivers) and pragmatic compressed-sensing based nonreplicated surveys when compared to the "ideal" scenario of exact replicability between surveys. Moreover, the recovered densely sampled prestack baseline and monitor data improve significantly when the acquisitions are not replicated, and hence can serve as input to extract poststack attributes used to compute time-lapse differences. Our observations are based on experiments conducted for an ocean-bottom cable survey acquired with time-jittered continuous recording assuming source equalization (or same source signature) for the time-lapse surveys and no changes in wave heights, water column velocities or temperature and salinity profiles, etc.}, keywords = {joint recovery method, Marine acquisition, off-the-grid recovery, Optimization, random sampling, time-lapse seismic}, doi = {10.1190/geo2016-0252.1}, url = {https://www.slim.eos.ubc.ca/Publications/Public/Journals/Geophysics/2017/wason2016GEOPctl/wason2016GEOPctl.html}, author = {Haneet Wason and Felix Oghenekohwo and Felix J. Herrmann} } @InBook{watanabe2015, author = "Watanabe, Kazumi", title = "Green's Functions for Laplace and Wave Equations", booktitle = "Integral Transform Techniques for Green's Function", year = "2015", publisher = "Springer International Publishing", address = "Cham", pages = "33--76", abstract = "This chapter shows how to apply the integral transform to the single partial differential equation such as Laplace and Wave equations. The basic technique of the integral transform method is demonstrated. Especially, in the case of the time-harmonic response for the 1 and 2D wave equations, the integration path for the inversion integral is discussed in detail with use of the results in Sect. 1.3. At the end of the chapter, the obtained Green's functions are listed in a table so that the reader can easily find the difference of the functional form among the Green's functions. An evaluation technique for a singular inversion integral which arises in a 2D static problem of Laplace equation is also developed.", isbn = "978-3-319-17455-6", doi = "10.1007/978-3-319-17455-6_2", url = "https://doi.org/10.1007/978-3-319-17455-6_2" } @Article{weiss2013, author = {Robin M. Weiss and Jeffrey Shragge}, date-modified = {2017-03-06 14:02:33 +0000}, doi = {http://dx.doi.org/10.1190/geo2012-0063.1}, journal = {Geophysics}, number = {2}, title = {Solving {3D} anisotropic elastic wave equations on parallel {GPU} devices.}, volume = {78}, year = {2013}, bdsk-url-1 = {http://dx.doi.org/10.1190/geo2012-0063.1} } @Article{whitmore1983, author = {{Whitmore}, N. D.}, title = {Iterative depth migration by backward time propagation}, journal = {1983 SEG Annual Meeting, Expanded Abstracts}, year = 1983 } @Misc{ williams2008, author = {Williams, Samuel and Patterson, David}, journal = {ParLab Summer Retreat}, note = {\url{http://crd.lbl.gov/assets/pubs presos/parlab08-roofline-talk.pdf}}, title = {Roofline Performance Model}, year = {2008} } @Article{williams2009roofline, author = {Williams, Samuel and Waterman, Andrew and Patterson, David}, journal = {communications of the acm}, number = {4}, title = {The Roofline model offers insight on how to improve the performance of software and hardware.}, volume = {52}, year = {2009} } @Article{williams_waterman_patterson_2009, abstractnote = {We propose an easy-to-understand, visual performance model that offers insights to programmers and architects on improving parallel software and hardware for floating point computations.}, author = {Williams, Samuel and Waterman, Andrew and Patterson, David}, journal = {Communications of the Association for Computing Machinery}, month = {Feb}, place = {United States}, title = {Roofline: An Insightful Visual Performance Model for Floating-Point Programs and Multicore Architectures}, url = {http://www.osti.gov/scitech/servlets/purl/963540}, year = {2009}, bdsk-url-1 = {http://www.osti.gov/scitech/servlets/purl/963540} } @Conference{witte2016segpve, title = {Phase velocity error minimizing scheme for the anisotropic pure P-wave equation}, booktitle = {SEG Technical Program Expanded Abstracts}, year = {2016}, note = {(SEG, Dallas)}, month = {10}, pages = {452-457}, abstract = {Pure P-wave equations for acoustic modeling in transverse isotropic media are derived by approximating the exact pure Pwave dispersion relation. In this work, we present an alternative approach to the approximate dispersion relation of Etgen and Brandsberg-Dahl, in which we approximate the exact dispersion relation through a polynomial expansion and determine its coefficients by solving a linear least squares problem that minimizes the phase velocity error over the entire range of phase angles. The coefficients are also optimized over a pre-defined range of Thomsen parameters, so that the phase error is small for models with spatially varying anisotropy. Phase velocity error analysis shows that the optimized pure P-wave equation is up to one order of magnitude more accurate than other popular pure P-wave equations, even for highly non-elliptic anisotropy. The optimized equation can be easily turned into a time-domain forward modeling scheme and comparisons of the modeled waveforms with analytical travel times once more illustrate its high accuracy. We also provide an efficient implementation of our approach for 3D tilted TI media that limits the count of fast Fourier transforms per time step to a number that is comparable to other pure P-wave equations.}, keywords = {anisotropy, least squares, Modelling, SEG, TTI}, doi = {10.1190/segam2016-13844850.1}, url = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2016/witte2016SEGpve/witte2016SEGpve.html}, presentation = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2016/witte2016SEGpve/witte2016SEGpve_poster.pdf}, author = {Philipp A. Witte and Christiaan C. Stolk and Felix J. Herrmann} } @Unpublished{witte2017eagespl, abstract = {Reverse-time migration (RTM) with the conventional cross-correlation imaging condition suffers from low-frequency artifacts that result from backscattered energy in the background velocity models. This problem translates to least-squares reverse-time migration (LS-RTM), where these artifacts slow down the convergence, as many of the initial iterations are spent on removing them. In RTM, this problem has been successfully addressed by the introduction of the so-called inverse scattering imaging condition, which naturally removes these artifacts. In this work, we derive the corresponding linearized forward operator of the inverse scattering imaging operator and incorporate this forward/adjoint operator pair into a sparsity-promoting (SPLS-RTM) workflow. We demonstrate on a challenging salt model, that LS-RTM with the inverse scattering imaging condition is far less prone to low-frequency artifacts than the conventional cross-correlation imaging condition, improves the convergence and does not require any type of additional image filters within the inversion. Through source subsampling and sparsity promotion, we reduce the computational cost in terms of PDE solves to a number comparable to conventional RTM, making our workflow applicable to large-scale problems.}, author = {Philipp A. Witte and Mengmeng Yang and Felix J. Herrmann}, keywords = {EAGE, imaging condition, least-squares migration, linearized Bregman, private}, title = {Sparsity-promoting least-squares migration with the linearized inverse scattering imaging condition}, url = {https://www.slim.eos.ubc.ca/Publications/Private/Conferences/EAGE/2017/witte2017EAGEspl/witte2017EAGEspl.html}, year = {2017}, bdsk-url-1 = {https://www.slim.eos.ubc.ca/Publications/Private/Conferences/EAGE/2017/witte2017EAGEspl/witte2017EAGEspl.html} } @article {witte2018alf, title = {A large-scale framework for symbolic implementations of seismic inversion algorithms in Julia}, journal = {Geophysics}, volume = {84}, number = {3}, year = {2019}, note = {(Geophysics)}, month = {03}, pages = {F57-F71}, abstract = {Writing software packages for seismic inversion is a very challenging task, since problems such as full-waveform inversion or least-squares imaging are both algorithmically and computationally demanding due to the large number of unknown parameters and the fact that we are propagating waves over many wavelengths. Software frameworks therefore need to combine both versatility and performance to provide geophysicists with the means and flexibility to implement complex algorithms that scale to exceedingly large 3D problems. Following these principles, we introduce the Julia Devito Inversion framework, an open-source software package in Julia for large-scale seismic modeling and inversion based on Devito, a domain-specific language compiler for automatic code generation. The framework consists of matrix-free linear operators for implementing seismic inversion algorithms that closely resembles the mathematical notation, a flexible resilient parallelization and an interface to Devito for generating optimized stencil code to solve the underlying wave equations. In comparison to many manually optimized industry codes written in low-level languages, our software is built on the idea of independent layers of abstractions and user interfaces with symbolic operators, making it possible to manage both the complexity of algorithms and performance optimizations, while preserving modularity, which allows for a level of expressiveness needed to formulate a broad range of wave-equation-based inversion problems. Through a series of numerical examples, we demonstrate that this allows users to implement algorithms for waveform inversion and imaging as simple Julia scripts that scale to large-scale 3D problems; thus providing a truly performant research and production framework.}, keywords = {FWI, inversion, LSRTM, Modeling, software}, doi = {10.1190/geo2018-0174.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/Geophysics/2019/witte2018alf/witte2018alf.pdf}, author = {Philipp A. Witte and Mathias Louboutin and Navjot Kukreja and Fabio Luporini and Michael Lange and Gerard J. Gorman and Felix J. Herrmann} } @article {witte2018fwip3, title = {Full-Waveform Inversion - Part 3: optimization}, journal = {The Leading Edge}, volume = {37}, number = {2}, year = {2018}, note = {(The Leading Edge)}, month = {1}, pages = {142-145}, abstract = {This tutorial is the third part of a full-waveform inversion (FWI) tutorial series with a step-by-step walkthrough of setting up forward and adjoint wave equations and building a basic FWI inversion framework. For discretizing and solving wave equations, we use Devito, a Python-based domain-specific language for automated generation of finite-difference code (Lange et al., 2016). The first two parts of this tutorial (Louboutin et al., 2017, 2018) demonstrated how to solve the acoustic wave equation for modeling seismic shot records and how to compute the gradient of the FWI objective function using the adjoint-state method. With these two key ingredients, we will now build an inversion framework that can be used to minimize the FWI least-squares objective function.}, keywords = {devito, finite-differences, FWI, inversion, Modeling, tutorial}, doi = {10.1190/tle37020142.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/TheLeadingEdge/2018/witte2018fwip3/witte2018fwip3.html}, author = {Philipp A. Witte and Mathias Louboutin and Keegan Lensink and Michael Lange and Navjot Kukreja and Fabio Luporini and Gerard Gorman and Felix J. Herrmann} } @article{ witte2018cls, title = {Compressive least-squares migration with on-the-fly Fourier transforms}, journal = {Geophysics}, volume = {84}, number = {5}, year = {2019}, note = {(Geophysics)}, month = {08}, pages = {R655-R672}, abstract = {Least-squares reverse-time migration is a powerful approach for true amplitude seismic imaging of complex geological structures, but the successful application of this method is currently hindered by its enormous computational cost, as well as high memory requirements for computing the gradient of the objective function. We tackle these problems by introducing an algorithm for low-cost sparsity-promoting least-squares migration using on-the-fly Fourier transforms. We formulate the least-squares migration objective function in the frequency domain and compute gradients for randomized subsets of shot records and frequencies, thus significantly reducing data movement and the number of overall wave equations solves. By using on-the-fly Fourier transforms, we can compute an arbitrary number of monochromatic frequency-domain wavefields with a time-domain modeling code, instead of having to solve individual Helmholtz equations for each frequency, which quickly becomes computationally infeasible when moving to high frequencies. Our numerical examples demonstrate that compressive imaging with on-the-fly Fourier transforms provides a fast and memory-efficient alternative to time-domain imaging with optimal checkpointing, whose memory requirements for a fixed background model and source wavelet is independent of the number of time steps. Instead, memory and additional computational cost grow with the number of frequencies and determine the amount of subsampling artifacts and crosstalk. In contrast to optimal checkpointing, this offers the possibility to trade both memory and computational cost for image quality or a larger number of iterations and is advantageous in new computing environments such as the cloud, where compute is often cheaper than memory and data movement.}, keywords = {Fourier, least squares migration, sparsity-promotion}, doi = {10.1190/geo2018-0490.1}, url = {https://slim.gatech.edu/Publications/Public/Journals/Geophysics/2019/witte2018cls/witte2018cls.pdf}, author = {Philipp A. Witte and Mathias Louboutin and Fabio Luporini and Gerard J. Gorman and Felix J. Herrmann} } @conference {witte2019SEGedw, title = {Event-driven workflows for large-scale seismic imaging in the cloud}, booktitle = {SEG Technical Program Expanded Abstracts}, year = {2019}, note = {(SEG, San Antonio)}, month = {09}, pages = {3984-3988}, abstract = {Cloud computing has seen a large rise in popularity in recent years and is becoming a cost effective alternative to on-premise computing, with theoretically unlimited scalability. However, so far little progress has been made in adapting the cloud for high performance computing (HPC) tasks, such as seismic imaging and inversion. As the cloud does not provide the same type of fast and reliable connections as conventional HPC clusters, porting legacy codes developed for HPC environments to the cloud is ineffective and misses out on an opportunity to take advantage of new technologies presented by the cloud. We present a novel approach of bringing seismic imaging and inversion workflows to the cloud, which does not rely on a traditional HPC environment, but is based on serverless and event-driven computations. Computational resources are assigned dynamically in response to events, thus minimizing idle time and providing resilience to hardware failures. We test our workflow on two large-scale imaging examples and demonstrate that cost-effective HPC in the cloud is possible, but requires careful reconsiderations of how to bring software to the cloud.}, keywords = {cloud, Imaging, large-scale, LS-RTM, RTM, SEG, workflow}, doi = {10.1190/segam2019-3215069.1}, url = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2019/witte2019SEGedw/witte2019SEGedw.html}, presentation = {https://slim.gatech.edu/Publications/Public/Conferences/SEG/2019/witte2019SEGedw/witte2019SEGedw_pres.pdf}, author = {Philipp A. Witte and Mathias Louboutin and Henryk Modzelewski and Charles Jones and James Selvage and Felix J. Herrmann} } @unpublished {witte2019TPDedas, author={P. A. {Witte} and M. {Louboutin} and H. {Modzelewski} and C. {Jones} and J. {Selvage} and F. J. {Herrmann}}, journal={IEEE Transactions on Parallel and Distributed Systems}, title={An Event-Driven Approach to Serverless Seismic Imaging in the Cloud}, publisher = {IEEE}, year={2020}, volume={31}, number={9}, pages={2032-2049} } @article {witte2019ecl, title = {Geophysics Bright Spots: Efficient coding of large-scale seismic inversion algorithms}, journal = {The Leading Edge}, volume = {38}, number = {6}, year = {2019}, note = {(The Leading Edge)}, pages = {482-484}, abstract = {In {\textquotedblleft}A large-scale framework for symbolic implementations of seismic inversion algorithms in Julia,{\textquotedblright} Witte et al. describe new developments in how to code complex geophysical algorithms in a concise way. Subsurface seismic imaging and parameter estimation are among the most computationally challenging problems in the scientific community. Codes for solving seismic inverse problems, such as FWI or least-squares reverse time migration (LS-RTM), need to be highly optimized, but at the same time, facilitate the implementation of complex optimization algorithms. Traditionally, production-level codes in the oil and gas industry were exclusively written in low-level languages, such as C or Fortran, with extensive amounts of manual performance optimizations, thus making code maintenance, debugging, and adoption of new algorithms prohibitively challenging. Witte et al. present a paradigm of software engineering for seismic inverse problems based on symbolic user interfaces and code generation with automated performance optimization. Inspired by recent deep learning frameworks, the Julia Devito inversion framework (JUDI; an open-source software package) combines high-level abstractions for expressing seismic inversion algorithms with a domain-specific language compiler called Devito for solving the underlying wave equations. Devito{\textquoteright}s generated code is compiled just in time and outperforms codes with manual performance optimizations. JUDI utilizes Julia{\textquoteright}s high-level parallelization, making the software easily adaptable to a variety of computing environments such as densely connected HPC clusters or the cloud. The numerical examples (Figure 3) demonstrate the ability to implement a variety of complex algorithms for FWI and LS-RTM in a few lines of Julia code and run it on large-scale 3D models. The paper concludes that abstractions and performance are not mutually exclusive, and use of symbolic user interfaces can facilitate the implementation of new and innovative seismic inversion algorithms.}, keywords = {bright spots, inversion, Julia, large-scale}, doi = {10.1190/tle38060482.1}, url = {https://library.seg.org/doi/10.1190/tle38060482.1}, author = {Philipp A. Witte and Mathias Louboutin and Navjot Kukreja and Fabio Luporini and Michael Lange and Gerard J. Gorman and Felix J. Herrmann} } @Article{xu2014, author = {{Xu}, Sheng and {Zhou}, Hongbo}, title = {Accurate simulations of pure quasi-P-waves in complex anisotropic media}, journal = {Geophysics}, year = 2014, month = {november-december}, volume = 79, number = 6, pages = {341-348} } @InProceedings{yask, author = {C. Yount}, booktitle = {2015 IEEE 17th International Conference on High Performance Computing and Communications, 2015 IEEE 7th International Symposium on Cyberspace Safety and Security, and 2015 IEEE 12th International Conference on Embedded Software and Systems}, doi = {10.1109/HPCC-CSS-ICESS.2015.27}, keywords = {data structures;multiprocessing systems;parallel processing;CPU;Intel Xeon Phi Coprocessor;hyper-threaded context;memory access;multidimensional SIMD-vector representation;multidimensional block;scientific-simulation application;sequential data element;stencil computation;stencil performance;vector folding;Jacobian matrices;Layout;Memory management;Registers;Shape;Three-dimensional displays;Intel;SIMD;Xeon Phi;high-performance computing;stencil;vector folding;vectorization}, month = {Aug}, pages = {865-870}, title = {Vector Folding: Improving Stencil Performance via Multi-dimensional SIMD-vector Representation}, year = {2015}, bdsk-url-1 = {http://dx.doi.org/10.1109/HPCC-CSS-ICESS.2015.27} } @InProceedings{yask-main, author = {Charles Yount and Josh Tobin and Alexander Breuer and Alejandro Duran}, title = {YASK--Yet Another Stencil Kernel: a framework for HPC stencil code-generation and tuning}, booktitle = {Proceedings of the 6th International Workshop on Domain-Specific Languages and High-Level Frameworks for High Performance Computing held as part of ACM/IEEE Supercomputing 2016 (SC16)}, series = {WOLFHPC'16}, month = {Nov}, year = {2016}, location = {Salt Lake City, Utah}, doi = {10.1109/WOLFHPC.2016.08} } @InProceedings{yask-vec-folding, author = {Charles Yount}, booktitle = {Proceedings of the IEEE 17th International Conference on High Performance Computing and Communications (HPCC)}, title = {Vector Folding: Improving Stencil Performance via Multi-dimensional SIMD-vector Representation}, month = {Aug}, year = {2015}, pages = {865-870}, doi = {10.1109/HPCC-CSS-ICESS.2015.27} } @book{ yilmaz2001, title={Seismic data analysis: Processing, inversion, and interpretation of seismic data}, author={Yilmaz, Oz}, year={2001}, publisher={Society of Exploration Geophysicists} } @Article{zhan2013, author = {{Zhan}, Ge and {Pestana}, Reynam C. and {Stoffa}, Paul L.}, title = {An efficient hybrid pdeudospectral/finite-difference scheme for solving the TTI pure P-wave equation}, journal = {Journal of Geophyics and Engineering}, year = 2013, volume = 10 } @InProceedings{zhang-stencil, author = {Zhang, Yongpeng and Mueller, Frank}, title = {Auto-generation and Auto-tuning of 3D Stencil Codes on GPU Clusters}, booktitle = {Proceedings of the Tenth International Symposium on Code Generation and Optimization}, series = {CGO '12}, year = {2012}, isbn = {978-1-4503-1206-6}, location = {San Jose, California}, pages = {155--164}, numpages = {10}, url = {http://doi.acm.org/10.1145/2259016.2259037}, doi = {10.1145/2259016.2259037}, acmid = {2259037}, publisher = {ACM}, address = {New York, NY, USA} } @Article{zhang-tti, author = {Yu Zhang and Houzhu Zhang and Guanquan Zhang}, title = {A stable TTI reverse time migration and its implementation}, journal = {GEOPHYSICS}, volume = {76}, number = {3}, pages = {WA3-WA11}, year = {2011}, doi = {10.1190/1.3554411}, url = {https://doi.org/10.1190/1.3554411}, eprint = {https://doi.org/10.1190/1.3554411} } @Article{zhang2005, author = {{Zhang}, Linbin and {Rector III}, James W. and {Hoversten} G. Micheal}, title = {Finite-difference modelling of wave propagation in acoustic tilted TI media}, journal = {Geophysical Prospecting}, year = 2005, volume = 53, pages = {843-852} } @Article{zhang2011, author = {{Zhang}, Yu and {Zhang}, Houzhu and {Zhang}, Guanquan}, title = {A stable TTI reverse time migration and its implementation}, journal = {Geophysics}, year = 2011, month = {may-june}, volume = 76, number = 3, pages = {WA3-WA11} } @book{fehler2011seam, title={SEAM Phase 1: Challenges of subsalt imaging in tertiary basins, with emphasis on deepwater Gulf of Mexico}, author={Fehler, Michael and Keliher, P Joseph}, year={2011}, publisher={Society of Exploration Geophysicists} } @book{fichtner2011, author = {Andreas Fichtner}, publisher = {Springer Verlag}, title = {{Full Seismic Waveform Modelling and Inversion}}, year = {2010}, doi = {10.1007/978-3-642-15807-0}, } @Article{ratcliffe2011, doi = {10.1190/1.3627688}, url = {https://doi.org/10.1190%2F1.3627688}, year = 2011, month = {jan}, publisher = {Society of Exploration Geophysicists}, author = {Andrew Ratcliffe and Caroline Win and Vetle Vinje and Graham Conroy and Mike Warner and Adrian Umpleby and Ivan Stekl and Tenice Nangoo and Alexandre Bertrand}, title = {Full waveform inversion: A North Sea {OBC} case study}, booktitle = {{SEG} Technical Program Expanded Abstracts 2011} } @Article{trinh2019, author = {Phuong-Thu Trinh and Romain Brossier and Ludovic Métivier and Laure Tavard and Jean Virieux}, title = {Efficient time-domain 3D elastic and viscoelastic full-waveform inversion using a spectral-element method on flexible Cartesian-based mesh}, journal = {Geophysics}, year = 2019, month = {January-February}, volume = 84, number = 1, pages = {R75--R97} } @Article{peter2011, author = {Peter, Daniel and Komatitsch, Dimitri and Luo, Yang and Martin, Roland and Le Goff, Nicolas and Casarotti, Emanuele and Le Loher, Pieyre and Magnoni, Federica and Liu, Qinya and Blitz, Céline and Nissen-Meyer, Tarje and Basini, Piero and Tromp, Jeroen}, title = {Forward and adjoint simulations of seismic wave propagation on fully unstructured hexahedral meshes}, journal = {Geophysical Journal International}, volume = {186}, number = {2}, pages = {721-739}, keywords = {Tomography, Interferometry, Computational seismology, Wave propagation}, doi = {10.1111/j.1365-246X.2011.05044.x}, url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1365-246X.2011.05044.x}, eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1365-246X.2011.05044.x}, abstract = {SUMMARY We present forward and adjoint spectral-element simulations of coupled acoustic and (an)elastic seismic wave propagation on fully unstructured hexahedral meshes. Simulations benefit from recent advances in hexahedral meshing, load balancing and software optimization. Meshing may be accomplished using a mesh generation tool kit such as CUBIT, and load balancing is facilitated by graph partitioning based on the SCOTCH library. Coupling between fluid and solid regions is incorporated in a straightforward fashion using domain decomposition. Topography, bathymetry and Moho undulations may be readily included in the mesh, and physical dispersion and attenuation associated with anelasticity are accounted for using a series of standard linear solids. Finite-frequency Fréchet derivatives are calculated using adjoint methods in both fluid and solid domains. The software is benchmarked for a layercake model. We present various examples of fully unstructured meshes, snapshots of wavefields and finite-frequency kernels generated by Version 2.0 ‘Sesame’ of our widely used open source spectral-element package SPECFEM3D.}, year = {2011} } @Manual{petersson2013, author = "Petersson, N.A. and Sj{\"o}green, B.", title="SW4 v1.1 [software]", year="2014", organization="Computational Infrastructure for Geodynamics", optkeywords="SW4", doi="http://doi.org/10.5281/zenodo.571844", opturl="https://geodynamics.org/cig/software/sw4/" } @article{lyu2020, author = {Chao Lyu and Yann Capdeville and Liang Zhao}, title = {Efficiency of the spectral element method with very high polynomial degree to solve the elastic wave equation}, journal = {GEOPHYSICS}, volume = {85}, number = {1}, pages = {T33-T43}, year = {2020}, doi = {10.1190/geo2019-0087.1}, URL = {https://doi.org/10.1190/geo2019-0087.1}, eprint = { https://doi.org/10.1190/geo2019-0087.1}, abstract = { ABSTRACTThe spectral element method (SEM) has gained tremendous popularity within the seismological community to solve the wave equation at all scales. Classic SEM applications mostly rely on degrees 4–8 elements in each tensorial direction. Higher degrees are usually not considered due to two main reasons. First, high degrees imply large elements, which make the meshing of mechanical discontinuities difficult. Second, the SEM’s collocation points cluster toward the edge of the elements with the degree, degrading the time-marching stability criteria and imposing a small time step and a high numerical cost. Recently, the homogenization method has been introduced in seismology. This method can be seen as a preprocessing step before solving the wave equation that smooths out the internal mechanical discontinuities of the elastic model. It releases the meshing constraint and makes use of very high degree elements more attractive. Thus, we address the question of memory and computing time efficiency of very high degree elements in SEM, up to degree 40. Numerical analyses reveal that, for a fixed accuracy, very high degree elements require less computer memory than low-degree elements. With minimum sampling points per minimum wavelength of 2.5, the memory needed for a degree 20 is about a quarter that of the one necessary for a degree 4 in two dimensions and about one-eighth in three dimensions. Moreover, for the SEM codes tested in this work, the computation time with degrees 12–24 can be up to twice faster than the classic degree 4. This makes SEM with very high degrees attractive and competitive for solving the wave equation in many situations. } } @unpublished {witte2019RHPCssi, title = {Serverless seismic imaging in the cloud}, year = {2019}, note = {Submitted to Rice Oil and Gas High Performance Computing Conference 2020 on November 27, 2019}, abstract = {This talk presents a serverless approach to seismic imaging in the cloud based on high-throughput containerized batch processing, event-driven computations and a domain-specific language compiler for solving the underlying wave equations. A 3D case study on Azure demonstrates that this approach allows reducing the operating cost of up to a factor of 6, making the cloud a viable alternative to on-premise HPC clusters for seismic imaging.}, keywords = {cloud, HPC, reverse-time migration, serverless}, url = {https://slim.gatech.edu/Publications/Public/Submitted/2019/witte2019RHPCssi/witte2019RHPCssi.html}, author = {Philipp A. Witte and Mathias Louboutin and Charles Jones and Felix J. Herrmann} } @InProceedings{dask, author = { Matthew Rocklin }, title = { Dask: Parallel Computation with Blocked algorithms and Task Scheduling }, booktitle = { Proceedings of the 14th Python in Science Conference }, pages = { 130 - 136 }, year = { 2015 }, editor = { Kathryn Huff and James Bergstra } }