14,842 papers · page 17 of 743
Sparse matrix-vector multiplication (SpMV) is a fundamental operation in scientific computing, machine learning, and graph analytics, demanding efficient execution on modern hardware. Recent advances in hardware accelerators, such as Tensor Cores, have significantly improved the …
As transformer sequence lengths grow, existing pipeline parallelisms incur suboptimal performance due to the quadratic attention computation and the substantial memory overhead. To relieve these challenges, we propose HelixPipe, a novel pipeline parallelism for long sequence tran…
This paper proposes FastAlign, a faster, cheaper, and practical end-to-end solution for sequence alignment using commercial CPUs. It introduces two key innovations: a multi-stage seeding algorithm that improves search performance while maintaining low memory consumption, and an i…
Mixed precision quantization has been adopted to accelerate large language models (LLMs) serving by leveraging high-throughput low-precision compute units in GPUs while preserving outliers in higher precision to maintain model accuracy. However, existing methods focus on mitigati…
Space-partitioning indexes are widely used for managing multi-dimensional data, but their throughput is often memory-bottlenecked. Processing-in-memory (PIM), an emerging architectural paradigm, mitigates memory bottlenecks by embedding processing cores directly within memory mod…
This tool paper introduces AGTix, an extension to reference attribute grammars that allows constraints, inspired by Visser et al.'s Statix system, that construct and query scope graphs to be written alongside traditional attribute equations. Previous work has shown that a restric…
For over 20 years, the Software Language Engineering community has pushed languages forward. More expressive. More formal. Better tooled. Metamodeling frameworks, transformations, graphical and textual notations, language workbenches. At Obeo, we have contributed to this journey …
Large Language Models (LLMs) are increasingly used for code generation, yet their support for programming languages is uneven, particularly for low-resource languages and infrequently used language constructs. Existing evaluation methodologies primarily assess functional correctn…
Effect handlers and dynamic wind provide mechanisms for implementing language constructs, such as generators and dynamically scoped variables, as libraries or embedded domain-specific languages (EDSLs). This paper presents a library implementation of dynamic wind in OCaml that is…
Resource Description Framework (RDF) is a widespread W3C standard defining a domain-specific language for knowledge representation, providing a platform to build higher-level languages such as OWL or SHACL. RDF concrete syntaxes define an external representation of its abstract m…
Context-awareness is crucial for developing cyber-physical systems to enable dynamic behavior. Context-oriented programming (COP) aims to improve the definition of context-dependent behavioral variations. However, existing layer-based COP approaches have two major limitations: la…
Maintaining consistency across heterogeneous textual DSLs that describe the same system is a recurring challenge in software language engineering. We advocate a separation of concerns: (i) language-specific model-to-text synchronizations that relate the concrete syntax to its abs…
Model-driven transformations play a central role in MDE processes, yet their correctness is most often validated through testing or simulation. However, in contexts where critical scenarios may compromise system integrity or certification objectives, stronger guarantees are neces…
Just-in-time (JIT) compilation is a core feature of many modern programming languages, combining the portability of interpreted languages with the performance benefits of ahead-of-time compilation. Originally introduced to improve the runtime performance of interpreted languages,…
Industrial cyber-physical systems generate vast amounts of semi-structured time-series data that require careful preprocessing before they can be effectively used for machine learning applications such as fault detection and identification. Raw sensor datasets are often corrupted…
This tool paper presents the Ztrategic framework, which integrates strategic term rewriting and attributes grammars through a unifying navigation abstraction based on functional zippers. The resulting zipper-based embedding is both concise and expressive, integrating the strength…
Datalog is a lightweight logic programming language, based on the logic of Horn clauses. Lean, on the other hand, is a proof assistant system and language based on the Calculus of Inductive Constructions (CIC). Datalog is more constrained and less expressive than Lean but has a l…
Many widely adopted programming languages (e.g., C++, Java, JavaScript, WebAssembly) have specification documents that provide authoritative answers about the language and its semantics. However, such specification documents are oftentimes demanding to read and navigate. This esp…
The Software Language Engineering (SLE) community has accumulated nearly two decades of content. This content comprises not only papers as such, but also other tangible artefacts like conference descriptions, calls for papers, tools, repositories. Yet, the field's self-descriptio…
Abstract elided by the publisher.