Mapple: A Domain-Specific Language for Mapping Distributed Heterogeneous Parallel Programs

2025-07-22

AI Summary

To address the complexity of parallel program mapping on distributed heterogeneous systems and the low-level, unwieldy interfaces of existing task-based systems, this paper proposes Mapple—a domain-specific language (DSL) that declaratively decouples application logic from performance optimization. Mapple introduces the novel *decompose* primitive to resolve dimensional mismatches between iteration spaces and processor topologies, thereby significantly reducing communication overhead. Built atop the Legion runtime, Mapple employs high-order transformation primitives to automatically compile mapper specifications into efficient C++ code, supporting diverse applications including matrix multiplication and scientific computing. Evaluation across nine benchmarks shows that Mapple reduces mapper code size by 14× compared to conventional approaches; achieves up to 1.34× speedup over expert-written C++ mappers; and accelerates mapping via the *decompose* primitive by up to 1.83× relative to state-of-the-art alternatives.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Distributed Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Optimizing parallel programs for distributed heterogeneous systems remains a complex task, often requiring significant code modifications. Task-based programming systems improve modularity by separating performance decisions from core application logic, but their mapping interfaces are often too low-level. In this work, we introduce Mapple, a high-level, declarative programming interface for mapping distributed applications. Mapple provides transformation primitives to resolve dimensionality mismatches between iteration and processor spaces, including a key primitive, decompose, that helps minimize communication volume. We implement Mapple on top of the Legion runtime by translating Mapple mappers into its low-level C++ interface. Across nine applications, including six matrix multiplication algorithms and three scientific computing workloads, Mapple reduces mapper code size by 14X and enables performance improvements of up to 1.34X over expert-written C++ mappers. In addition, the decompose primitive achieves up to 1.83X improvement over existing dimensionality-resolution heuristics. These results demonstrate that Mapple simplifies the development of high-performance mappers for distributed applications.

Problem

• Simplifying optimization of distributed heterogeneous parallel programs
• Providing high-level declarative mapping for distributed applications
• Reducing code size and improving performance of mappers

Innovation

• Declarative interface for distributed application mapping
• Transformation primitives resolve dimensionality mismatches
• Decompose primitive minimizes communication volume