Mat2Stencil: A Modular Matrix-Based DSL for Explicit and Implicit Matrix-Free PDE Solvers on Structured Grid
Partial differential equation (PDE) solvers are extensively utilized across numerous scientific and engineering fields. However, achieving high performance and scalability often necessitates intricate and low-level programming, particularly when leveraging deterministic sparsity patterns in structured grids.
In this paper, we propose an innovative domain-specific language (DSL), Mat2Stencil, with its compiler, for PDE solvers on structured grids.
Mat2Stencil introduces a structured sparse matrix abstraction, facilitating modular, flexible, and easy-to-use expression of solvers across a broad spectrum, encompassing components such as Jacobi or Gauss-Seidel preconditioners, incomplete LU or Cholesky decompositions, and multigrid methods built upon them.
Our DSL compiler subsequently generates matrix-free code consisting of generalized stencils through multi-stage programming. The code allows spatial loop-carried dependence in the form of quasi-affine loops, in addition to the Jacobi-style stencil's embarrassingly parallel on spatial dimensions.
We further propose a novel automatic parallelization technique for the spatially dependent loops, which offers a compile-time deterministic task partitioning for threading, calculates necessary inter-thread synchronization automatically, and generates an efficient multi-threaded implementation with fine-grained synchronization.
Implementing 4 benchmarking programs, 3 of them being the pseudo-applications in NAS Parallel Benchmarks with $6.3%$ lines of code and 1 being matrix-free High Performance Conjugate Gradients with $16.4%$ lines of code, we achieve up to $1.67\times$ and on average $1.03\times$ performance compared to manual implementations.
Wed 25 OctDisplayed time zone: Lisbon change
16:00 - 17:48
|Fluent APIs in Functional Languages
|A Pretty Expressive Printer
Sorawee Porncharoenwase University of Washington, Justin Pombrio Unaffiliated, Emina Torlak Amazon Web Services, USADOI Pre-print
|How Domain Experts Use an Embedded DSL
Lisa Rennels University of California at Berkeley, Sarah E. Chasins University of California at BerkeleyDOI
|Saggitarius: A DSL for Specifying Grammatical Domains
Anders Miltner Simon Fraser University, Devon Loehr Princeton University, Arnold Mong Princeton University, Kathleen Fisher Tufts University, David Walker Princeton UniversityDOI
|Mat2Stencil: A Modular Matrix-Based DSL for Explicit and Implicit Matrix-Free PDE Solvers on Structured Grid
Huanqi Cao Tsinghua University, Shizhi Tang Tsinghua University, Qianchao Zhu Peking University, Bowen Yu Tsinghua University, Wenguang Chen Tsinghua University; Pengcheng LaboratoryDOI
|Translating canonical SQL to imperative code in Coq
Véronique Benzaken Université Paris-Saclay - Laboratoire de Méthodes Formelles , Évelyne Contejean CNRS, ENS Paris-Saclay & Université Paris-Saclay, Houssem Hachmaoui , Chantal Keller Université Paris Saclay, Louis Mandel IBM Research, USA, Avraham Shinnar IBM Research, Jerome Simeon DocuSign, Inc.Link to publication DOI