21 dépôts
Techniques used by compilers to improve the execution speed and efficiency of generated code.
Distinct from Tool Use and Function Calling: None of the candidates address compiler-level function inlining.
Explore 21 awesome GitHub repositories matching software engineering & architecture · Compiler Optimizations. Refine with filters or upvote what's useful.
Tailwind CSS is a utility-first CSS framework and design system framework used to build custom user interfaces. It provides a set of low-level utility classes that are applied directly in HTML, allowing for the creation of custom web designs without writing traditional custom CSS files. The framework functions as a PostCSS plugin and a responsive web design tool. It utilizes a mobile-first system of modifiers to create layouts that adapt to different screen sizes and establishes consistent visual languages through a centralized configuration of colors, spacing, and typography. The system sup
Features a JIT engine that scans files to generate only the CSS rules actually used in the markup.
Compose Multiplatform is a declarative user interface framework that enables the development of native graphical interfaces across desktop, mobile, and web platforms using a single Kotlin codebase. It functions as a multiplatform library that allows developers to share both application logic and UI components while maintaining access to platform-specific features through abstraction layers. The framework utilizes a reactive state management system that automatically tracks data mutations to trigger granular interface updates. By employing a cross-platform rendering engine based on a hardware-
Transforms functional UI code into efficient imperative instructions to optimize state tracking and execution.
CS-Base is a comprehensive educational platform and technical repository designed to support software engineers in mastering backend architecture, artificial intelligence engineering, and career development. It functions as a centralized knowledge hub that combines illustrated theoretical tutorials with practical, project-based learning to bridge the gap between foundational computer science concepts and professional industry requirements. The project distinguishes itself by integrating a robust career mentorship framework with advanced AI engineering resources. It provides users with tools f
Details compiler optimization techniques including function inlining for performance.
CoreCLR is a runtime environment that manages the execution, memory, and basic types for applications built on the .NET platform. It serves as a managed execution environment that handles low-level system interactions and provides primitive data types for high-level application code. The project functions as a JIT compilation engine and a garbage collected runtime. It translates intermediate language into machine code at runtime for execution on specific hardware and automatically reclaims unused memory to prevent leaks. The system covers broad capability areas including managed code executi
Balances startup speed and steady-state performance by recompiling frequently used methods with aggressive optimizations.
Numba est un compilateur juste-à-temps (JIT) qui traduit des fonctions Python de haut niveau en code machine optimisé lors de l'exécution. En tirant parti de l'infrastructure de compilation LLVM, il fournit un framework pour accélérer le traitement des données numériques et les calculs mathématiques, permettant des niveaux de performance comparables aux langages compilés statiquement. Le projet se distingue par sa capacité à effectuer une spécialisation basée sur l'inférence de type, qui génère des instructions machine adaptées aux types de données spécifiques utilisés lors de l'exécution. Il emploie un pipeline de compilation paresseuse qui diffère la traduction jusqu'au moment de l'invocation, minimisant la surcharge au démarrage tout en maintenant des performances constantes sur diverses architectures de processeurs et systèmes d'exploitation. Au-delà de la compilation de base, le toolkit offre un support étendu pour l'accélération matérielle en distribuant les opérations itératives et les expressions de tableaux sur plusieurs cœurs CPU et unités de traitement graphique. Il utilise des stratégies de vectorisation et de parallélisation pour maximiser le débit pour les grands jeux de données numériques, permettant aux développeurs de cibler du matériel spécialisé directement depuis du code standard.
Defers the translation of functions into machine code until the moment of first invocation to minimize startup overhead and optimize execution paths.
ChakraCore is an embeddable JavaScript engine and high-performance scripting runtime. It provides a just-in-time compiler that converts JavaScript source code into optimized machine code during runtime to increase execution speed and throughput. The engine utilizes a multi-tiered compilation pipeline and tiered machine code generation to balance startup time with execution speed. Memory is managed via a concurrent garbage collector that reclaims unreachable objects on background threads to minimize application pauses. The project provides capabilities for embedded JavaScript execution and au
Uses a multi-tiered compilation strategy to balance initial startup time with steady-state execution speed.
Magic est une bibliothèque d'animation CSS et un framework d'effets visuels fournissant une collection de styles de mouvement prédéfinis pour déplacer, faire pivoter et transformer des éléments web. Il sert d'optimiseur CSS au moment du build qui compile uniquement les styles d'animation spécifiques utilisés dans un projet pour minimiser la taille finale de la feuille de style. Le framework permet le développement d'interfaces utilisateur interactives en permettant aux animations d'être déclenchées programmatiquement via des scripts, des événements de survol ou des délais temporels. Il utilise un système de classes de timing et de mouvement pour gérer la durée et la vitesse des transitions visuelles. Le projet couvre le stylage visuel frontend et la gestion des animations, se concentrant sur l'application d'effets de mouvement et l'optimisation des performances web via des bundles d'animation personnalisés.
Includes an engine that generates a minimal CSS stylesheet by scanning source files for used animation classes.
This project is a JavaScript optimizer, minifier, module bundler, transpiler, and static analysis tool. It provides a compilation pipeline designed to shrink file sizes and improve runtime performance. The system utilizes a multi-pass compilation process to perform dead code elimination, global name mangling, and static type inference. It identifies unreachable functions and unused variables to reduce the final output size and detects potential runtime errors without executing the code. The tool manages assets through dependency resolution, code chunking, and bundle management. It ensures co
Applies advanced compilation techniques to reduce processing overhead and improve execution speed.
Warp is a Python framework that JIT-compiles Python functions into CUDA kernels for GPU-accelerated parallel computation, with built-in automatic differentiation and multi-framework array interoperability. At its core, it provides a GPU kernel compilation system that enables writing and executing custom GPU kernels directly from Python, while supporting automatic gradient computation through those kernels for integration with machine learning pipelines. The framework also includes tile-based cooperative computing, where thread blocks partition into tiles for shared-memory and tensor-core opera
Selects the optimization level applied to GPU kernels, trading compile time for execution speed.
Applies link-time optimization to select the best GPU kernels for a given configuration without manual tuning.
Torchtune is a PyTorch-native library for fine-tuning, aligning, and quantizing large language models. It provides a configurable training pipeline orchestrated through YAML recipes, with CLI overrides and component swapping, distributed training via FSDP2, memory optimizations, and parameter-efficient fine-tuning methods like LoRA, DoRA, and QLoRA. The library distinguishes itself through its YAML-driven configuration system that defines all training parameters and instantiates components from config files, with full CLI override capability for any field or component at launch time. It suppo
Fuses the optimizer step into the backward pass to eliminate gradient buffers and reduce memory for stateful optimizers.
Torchtune is a PyTorch-native library for fine-tuning, aligning, and quantizing large language models. It provides a config-driven system for instantiating components, orchestrating distributed training, and managing parameter-efficient fine-tuning with quantization support, all through YAML-based configurations and command-line overrides. The library distinguishes itself through its comprehensive post-training workflow orchestration, combining supervised fine-tuning, preference optimization (DPO, PPO, GRPO), knowledge distillation, and quantization-aware training in a single configurable pip
Fuses optimizer updates into the backward pass to eliminate gradient buffer memory during training.
TileLang is a Python-embedded domain-specific language compiler that JIT-compiles and autotunes GPU kernels. It uses a tile-based DSL, automatic software pipelining, and parallel autotuning to generate optimized GPU kernels at runtime. It supports tensor core operations with Pythonic syntax, automatic memory management, and thread mapping. The compiler searches over tile sizes, thread counts, and scheduling policies, compiling and benchmarking candidates in parallel to find the fastest kernel. It also caches compiled binaries and tuning results to disk for reuse across sessions. TileLang inc
Allows customization of the compiler pass pipeline to fine-tune generated kernel code.
c3c is the compiler for the C3 programming language, transforming source code into executable binaries, static libraries, or dynamic libraries using an LLVM backend. It implements a system based on result-based error handling, scoped memory pooling, and a semantic macro system. The compiler provides first-class support for hardware-backed SIMD vectors that map directly to processor instructions and enables runtime polymorphism through interface-based dynamic dispatch. The project covers a broad set of low-level capabilities, including manual and pooled memory management, inline assembly inte
Forces or prevents function inlining and disables sanitizer instrumentation for specific functions.
Ce projet est un compilateur d'obfuscation LLVM et un obfuscateur de code machine conçu pour protéger les applications C++ contre la décompilation et l'analyse statique. Il fonctionne comme un framework de passe LLVM qui transforme la représentation intermédiaire en code machine complexe pour empêcher l'ingénierie inverse de la logique du programme. L'outil implémente le durcissement de la sécurité binaire et des mesures anti-ingénierie inverse pour protéger la propriété intellectuelle. Il y parvient en modifiant le pipeline de compilation pour masquer le flux de contrôle original et les structures de données. Le processus de transformation inclut l'aplatissement du flux de contrôle, l'injection de prédicats opaques et la substitution d'instructions. Il complique davantage le binaire par l'insertion de code mort et la modification des instructions machine tout en préservant la sémantique d'exécution originale.
Utilizes a framework of transformation passes to modify the optimization pipeline for code obfuscation.
phpDocumentor est un générateur de documentation API PHP et un analyseur de code source qui transforme les fichiers PHP et les DocBlocks en références API HTML structurées. Il fonctionne comme un générateur de site statique et un outil de documentation automatique conçu pour synchroniser la documentation technique avec les changements de code. Le projet se distingue en agissant comme un générateur de diagrammes UML, produisant des graphiques de classes et d'architecture via PlantUML basés sur l'analyse du code source. Il prend également en charge la rédaction de manuels techniques, rendant des guides écrits à la main en Markdown et ReStructuredText aux côtés des références API générées automatiquement. L'outil fournit de larges capacités pour l'analyse du code source, y compris la génération d'arbres de syntaxe abstraits, le marquage de métadonnées et la gestion de l'héritage de la documentation. Il inclut des outils de visualisation pour l'architecture du code, le contrôle de la visibilité des éléments et la gestion de la documentation versionnée. Les builds peuvent être gérés via des fichiers de configuration XML pour automatiser la génération au sein des pipelines d'intégration continue.
Runs specific business rules or transformations on extracted source data during the compilation process.
ExecuTorch is a lightweight C++ runtime for deploying PyTorch models on mobile, embedded, and edge hardware. It provides an ahead-of-time compilation pipeline that exports, quantizes, and lowers model graphs into compact serialized programs, then executes them through a minimal runtime with hardware acceleration and on-device large language model inference capabilities. The project distinguishes itself through a hardware accelerator delegate system that partitions model subgraphs and offloads computation to specialized backends including NPUs, GPUs, and DSPs from Apple, Arm, Intel, MediaTek,
Ships a custom pass system for applying user-defined transformations to model graphs during compilation.
Clean-CSS est un optimiseur CSS Node.js qui fonctionne comme un minificateur, un bundler et un post-processeur. Il est conçu pour diminuer le volume total des feuilles de style en supprimant les espaces blancs, les commentaires et le code redondant. Le projet fournit un pipeline pour appliquer des transformations personnalisées et des ajustements de compatibilité des navigateurs. Il permet la modification programmatique des règles et valeurs CSS via un système de plugin et l'utilisation de plugins d'optimisation personnalisés. L'outil couvre un large éventail de capacités d'optimisation d'actifs, incluant le regroupement de feuilles de style, l'intégration des règles d'importation et le rebasage des URL relatives. Il prend également en charge la génération de source maps pour le débogage et le formatage de sortie personnalisable pour l'embellissement.
Provides selectable optimization levels to control the aggressiveness of CSS code reduction.
IREE is an MLIR-based compiler toolchain and runtime designed to translate machine learning models from various frameworks into optimized binaries for execution across diverse hardware targets. It provides a unified pipeline to ingest models from PyTorch, TensorFlow, JAX, and ONNX, lowering them into a common intermediate representation for deployment on CPUs, GPUs, and bare-metal embedded systems. The project distinguishes itself through a bytecode virtual machine and a hardware abstraction layer that decouple high-level model logic from specific hardware instruction sets. It supports sophis
Adjusts LLVM optimization levels for generated code to isolate bugs or identify race conditions.
This project serves as the core integration layer for building full-stack web applications in PHP. It provides a unified foundation by connecting essential infrastructure components, including a dependency injection container, an event dispatcher, and a request routing engine. The framework operates through a kernel-based request lifecycle that transforms incoming HTTP requests into responses via prioritized middleware stages. The system distinguishes itself through a highly modular architecture that relies on configuration-driven service wiring and metadata-driven attribute mapping. During t
Performs compiler-pass service optimization to analyze and transform the dependency graph for efficient service resolution.