awesome-repositories.com
博客
awesome-repositories.com

通过 AI 驱动的搜索,发现最优秀的开源仓库。

探索精选搜索开源替代品自托管软件博客网站地图
项目关于排名机制媒体报道MCP 服务器
法律隐私政策服务条款
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
·

19 个仓库

Awesome GitHub RepositoriesIntermediate Representations

Generation of target-neutral intermediate formats that can be optimized for various machine-code back-ends.

Distinct from Machine Code Generation: Distinct from Machine Code Generation: focuses on the creation of the intermediate representation rather than the final emission of native instructions.

Explore 19 awesome GitHub repositories matching programming languages & runtimes · Intermediate Representations. Refine with filters or upvote what's useful.

Awesome Intermediate Representations GitHub Repositories

用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
  • google/protobufgoogle 的头像

    google/protobuf

    71,412在 GitHub 上查看↗

    Protocol Buffers is a binary serialization framework used to encode structured information into compact payloads to reduce network bandwidth and storage. It functions as a cross-language data interchange standard that enables different platforms and languages to exchange structured data using a shared schema. The project includes an interface definition language compiler that transforms schema definitions into type-safe source code for multiple target programming languages. This mechanism decouples data structures from specific language memory layouts and ensures consistent data handling acro

    Uses a common definition format to decouple data structures from specific language memory layouts.

    C++
    在 GitHub 上查看↗71,412
  • chrismaltby/gb-studiochrismaltby 的头像

    chrismaltby/gb-studio

    9,315在 GitHub 上查看↗

    GB Studio 是一个可视化集成开发环境和游戏引擎,用于为 Game Boy 硬件创建 8 位游戏。它充当复古硬件 ROM 编译器,将图形逻辑转译为与原始掌机和模拟器兼容的原生 Z80 汇编和二进制镜像。 该项目作为跨平台构建工具,从单个项目源生成原生硬件 ROM 和 Web 兼容版本。它利用拖放界面进行游戏逻辑和场景设计,允许在不编写底层机器码的情况下创建游戏机制和资产放置。 该工具集包括一个将原始图像和音频处理为硬件特定平铺格式和调色板的流水线。它还提供性能分析功能,以分析时序统计信息和执行日志,从而优化软件在受限硬件上的响应速度。

    Translates project source files into a portable intermediate representation before generating final machine code for target platforms.

    TypeScript
    在 GitHub 上查看↗9,315
  • red/redred 的头像

    red/red

    6,009在 GitHub 上查看↗

    Red is a programming language with a native compiler that translates high-level source code into standalone executables for Windows, macOS, and Linux without external runtime dependencies. It combines a cross-platform GUI development framework that renders native operating system widgets from a single codebase with a reactive data binding system that automatically synchronizes UI state with data sources. The language also includes an embedded DSL and parsing engine based on PEG grammar rules for defining and processing domain-specific languages within the language itself. The project distingu

    Translates high-level Red code into a C-like intermediate representation before generating native machine code.

    Red
    在 GitHub 上查看↗6,009
  • microsoft/mmdnnMicrosoft 的头像

    Microsoft/MMdnn

    5,804在 GitHub 上查看↗

    MMdnn 是一个深度学习模型转换与迁移工具,旨在实现 TensorFlow、PyTorch 和 Keras 等不同框架之间的神经网络架构与权重转换。它利用标准化的中间表示(IR)将网络结构和权重与特定框架的实现解耦,从而支持在不同环境中转换预训练模型。 该项目的特色在于能从中间表示生成原生的 Python 重构代码,允许在目标环境中重建和微调模型。它还包含用于移动端模型部署的专用工具,可将深度学习模型转换为 CoreML 和 TensorFlow Lite 等移动兼容格式。 该系统提供了一套更广泛的功能,包括用于检查图结构和元数据的神经网络架构可视化,以及用于验证转换后模型是否保持原始行为和精度的模型推理执行。其他实用程序还可处理从远程仓库获取预训练权重以及组装可部署的模型检查点。

    Generates native framework architecture and weight files from intermediate model representations via Python code snippets.

    Python
    在 GitHub 上查看↗5,804
  • chai2010/go-ast-bookchai2010 的头像

    chai2010/go-ast-book

    5,510在 GitHub 上查看↗

    go-ast-book 是一系列专注于抽象语法树(AST)分析、编译器开发和静态代码验证的教育和技术资源。它提供了解析、遍历和分析 Go 源代码以提取语义含义的指南和手册。 该项目作为构建编译器前端的参考,涵盖了将高级代码转换为中间表示和静态单赋值形式的过程。它还提供了使用这些技术开发语言工具和执行静态代码分析的说明。 这些资源涵盖了广泛的静态分析能力,包括词法标记化、表达式和声明的结构化解析,以及源文件的坐标跟踪。它还详细介绍了语义分析过程,如标识符解析、类型正确性验证,以及针对并发和延迟执行的控制流分析。

    Details the transformation of high-level syntax trees into single static assignment form for optimized execution.

    Go
    在 GitHub 上查看↗5,510
  • rust-gpu/rust-cudaRust-GPU 的头像

    Rust-GPU/rust-cuda

    5,245在 GitHub 上查看↗

    rust-cuda 是一个 GPU 编程框架和设备编译器,允许使用 Rust 在 NVIDIA 硬件上开发和执行高性能内核。它提供了一个驱动程序包装器来管理设备内存分配和内核启动,有效地作为一个无需依赖 C++ 即可编写 GPU 计算逻辑的系统。 该项目包含一个计算库,其中包含用于神经网络加速和硬件加速光线追踪的硬件优化原语。它利用一个编译工具链,将源代码转换为用于在图形处理器上执行的低级中间表示。 该框架涵盖了设备资源管理、内核开发以及高精度整数运算的模拟。它还支持设备端随机数生成和特定目标的计算优化。 提供预配置的容器镜像,以简化跨不同硬件架构的编译器工具链和开发环境的配置。

    Compiles Rust source code into a PTX-based intermediate representation for GPU execution.

    Rustcudacuda-kernelscuda-programming
    在 GitHub 上查看↗5,245
  • microsoft/bosquelanguagemicrosoft 的头像

    microsoft/BosqueLanguage

    5,215在 GitHub 上查看↗

    Bosque is an experimental programming language and development platform designed for machine-assisted software construction. It combines functional programming semantics with imperative syntax to enforce logic correctness and runtime safety, providing a type-safe environment that utilizes structured data models to maintain information integrity throughout the application lifecycle. The platform distinguishes itself through deep integration with formal verification tools, including automated theorem provers and symbolic execution engines. By transforming source code into a regularized intermed

    Transforms source code into a standardized intermediate representation to enable automated analysis and formal verification of program logic.

    TypeScript
    在 GitHub 上查看↗5,215
  • airbnb/lottieairbnb 的头像

    airbnb/lottie

    4,922在 GitHub 上查看↗

    Lottie 是一个 JSON 矢量动画引擎和跨平台播放器,用于渲染分辨率无关的动态图形。它作为一个运行时渲染器,使用基于 JSON 的中间表示将 Adobe After Effects 合成转换为实时动画。 该项目通过在运行时实现对矢量资产的动态控制而脱颖而出。它允许通过分层键路径和动态主题对动画属性、颜色和播放状态进行编程修改。这使得动画能够响应实时应用状态变化和用户手势。 该引擎涵盖了广泛的功能,包括用于高性能的硬件加速渲染和模型缓存。它为帧搜索、段循环和事件触发提供了全面的播放控制,以及支持 3D 摄像机、形状变形和程序化矢量生成的渲染系统。资产管理通过多源数据导入进行处理,支持本地包、网络 URL 和各种矢量文件格式。

    Uses a JSON-based intermediate representation to decouple design tool compositions from cross-platform runtime playback.

    HTML
    在 GitHub 上查看↗4,922
  • llvm-mirror/llvmllvm-mirror 的头像

    llvm-mirror/llvm

    4,580在 GitHub 上查看↗

    LLVM 是一个模块化的编译器基础设施和框架,由一系列可重用库组成,用于构建语言前端和优化工具。它作为一个底层虚拟机,提供中间表示和优化流水线,以实现与语言无关的代码生成。 该项目提供了一套即时编译(JIT)工具包,可在运行时将中间表示翻译为机器码以供立即执行。其模块化设计允许开发者在多种目标架构上创建自定义语言前端和高性能运行时环境。 该基础设施涵盖了高性能代码优化和跨平台工具链的协调,包括编译器、运行时和调试器的集成。

    Implements an intermediate representation using static single assignment form to simplify data flow analysis and optimization.

    LLVMcode-generationintermediate-representationllvm
    在 GitHub 上查看↗4,580
  • yosyshq/yosysYosysHQ 的头像

    YosysHQ/yosys

    4,543在 GitHub 上查看↗

    Yosys 是一个数字逻辑综合器和 RTL 综合工具,可将硬件描述语言代码转换为优化的门级表示。它作为硬件逻辑优化器和形式验证工具,提供了一个将通用电路表示转换为特定技术库单元的框架,以适配目标硬件平台。 该项目通过可脚本化的环境充当硬件设计自动化工具,用于执行综合命令序列。这种自动化允许管理复杂的数字硬件工作流,将 RTL 设计综合与技术单元映射集成在一起。 其验证和分析能力包括符号模型检查和设计等价性验证。这些工具通过检查安全性和活性属性并生成仿真测试平台,支持形式化硬件验证,确保设计符合其原始规范。

    Uses directed graphs of cells and nets as an intermediate representation for hardware design analysis and transformation.

    C++
    在 GitHub 上查看↗4,543
  • openvinotoolkit/open_model_zooopenvinotoolkit 的头像

    openvinotoolkit/open_model_zoo

    4,408在 GitHub 上查看↗

    Open Model Zoo 是一个精选的预训练和优化深度学习模型集合,专为使用 OpenVINO 进行高性能推理而设计。它作为一个模型库和部署框架,简化了神经网络到生产环境的集成。 该项目利用中心化清单和版本化注册表来自动化模型权重和元数据的下载与组织。它包括用于对推理性能进行基准测试的工具,并通过将输出与真值张量进行比较来验证模型准确性,从而量化精度损失。 该生态系统提供参考实现和模块化演示架构,以将推理引擎与应用逻辑解耦。这些工具通过抽象预处理和后处理要求的标准化封装器,支持计算机视觉任务的实现。

    Translates diverse deep learning framework formats into a unified intermediate representation for hardware-independent execution.

    Pythoncaffemodelcnn-modeldeep-learning-models
    在 GitHub 上查看↗4,408
  • phpdocumentor/phpdocumentorphpDocumentor 的头像

    phpDocumentor/phpDocumentor

    4,332在 GitHub 上查看↗

    phpDocumentor 是一个 PHP API 文档生成器和源代码分析器,将 PHP 文件和 DocBlocks 转换为结构化的 HTML API 参考。它作为一个静态站点生成器和自动化文档工具,旨在使技术文档与代码变更保持同步。 该项目的独特之处在于它充当 UML 图生成器,基于源代码分析通过 PlantUML 生成类图和架构图。它还支持技术手册编写,在自动生成的 API 参考旁边渲染手写的 Markdown 和 ReStructuredText 指南。 该工具提供了广泛的源代码分析能力,包括抽象语法树的生成、元数据标记以及文档继承的管理。它包括用于代码架构、元素可见性控制和版本化文档管理的可视化工具。 构建可以通过 XML 配置文件进行管理,以在持续集成管道中实现自动化生成。

    Transforms source code into structured descriptors to decouple analysis from final documentation rendering.

    PHPdocblockdocblocksgraphviz
    在 GitHub 上查看↗4,332
  • hyperai/tvm-cnhyperai 的头像

    hyperai/tvm-cn

    3,813在 GitHub 上查看↗

    This project is a collection of technical guides and manuals for the Apache TVM compiler stack translated into Simplified Chinese. It provides translated documentation focusing on deep learning compilation and the transformation of machine learning models into optimized executable code. The documentation covers the use of hardware backend guides for deploying models across CPUs, GPUs, and specialized accelerators. It also includes references for intermediate representations and graph-level optimizations used to compile tensor programs.

    Describes the transformation of high-level graph operators into low-level primitives matching target hardware specifications.

    TypeScriptapachechinese-simplifieddeep-learning
    在 GitHub 上查看↗3,813
  • ghdl/ghdlghdl 的头像

    ghdl/ghdl

    2,759在 GitHub 上查看↗

    GHDL is a compiler and simulator for VHDL hardware descriptions. It functions as a multi-pass analysis elaborator that resolves design hierarchies and dependencies to prepare hardware descriptions for simulation or synthesis. The project transforms VHDL source code into executable binaries for high-speed digital design verification and serves as a synthesis tool that converts descriptions into structural netlists compatible with vendor or open-source flows. It also implements the Language Server Protocol to provide static analysis, autocomplete, and code navigation for VHDL files. The toolse

    Implements backend-agnostic intermediate representations to support multiple target machine-code generators.

    VHDLcompilergccghdl
    在 GitHub 上查看↗2,759
  • flix/flixflix 的头像

    flix/flix

    2,585在 GitHub 上查看↗

    Flix is a statically typed multi-paradigm programming language that combines functional, imperative, and logic programming paradigms within a single unified environment. It utilizes a constraint-driven execution environment to derive results by satisfying logical predicates rather than following linear sequential instructions. The language features a unification-based logic engine designed to solve complex computational problems through term unification and backtracking. This is supported by a core library of persistent immutable data structures that ensure side-effect free state management a

    Uses a statically typed intermediate representation to maintain safety during translation to machine instructions.

    Flixflixfunctionalhacktoberfest
    在 GitHub 上查看↗2,585
  • foonathan/type_safefoonathan 的头像

    foonathan/type_safe

    1,641在 GitHub 上查看↗

    This library is a collection of zero-overhead C++ abstractions designed to replace primitive types with strict wrappers. By leveraging template metaprogramming, it enforces compile-time type safety and validates data integrity, ensuring that logic errors are caught during the build process rather than at runtime. The project distinguishes itself by providing a zero-cost abstraction layer that maps high-level type safety wrappers directly to machine instructions. It utilizes static type identity enforcement to treat logically distinct data as incompatible types, even when they share the same u

    Treats logically distinct data as incompatible types at compile time to prevent accidental mixing.

    C++c-plus-plustype-safety
    在 GitHub 上查看↗1,641
  • webassembly/wasi-sdkWebAssembly 的头像

    WebAssembly/wasi-sdk

    1,504在 GitHub 上查看↗

    The WASI SDK is a comprehensive compiler toolchain designed to transform C and C++ source code into portable, sandboxed binary modules. It provides the necessary utilities to target the WebAssembly System Interface, enabling native code to execute across diverse hardware and operating system environments through a standardized interface. The toolkit distinguishes itself by providing a complete, sysroot-isolated environment that ensures build consistency regardless of the host operating system. By bundling compilers, system headers, and libraries into unified, portable archives, it facilitates

    Utilizes LLVM intermediate representation to perform cross-platform optimizations during the compilation process.

    CMakellvmsysrootwasi-libc
    在 GitHub 上查看↗1,504
  • grin-compiler/gringrin-compiler 的头像

    grin-compiler/grin

    1,062在 GitHub 上查看↗

    Grin is a compiler infrastructure designed for the translation and optimization of high-level functional languages. It functions as a graph-reduction compiler, transforming source code into a standardized intermediate representation that facilitates whole-program analysis and efficient execution across diverse hardware architectures. The project distinguishes itself through a modular backend infrastructure that separates intermediate representation management from target-specific code generation. By utilizing graph-based transformations, the system performs advanced optimizations such as valu

    Generates standardized intermediate representations to enable back-end optimizations and translation for diverse hardware architectures.

    Haskellcompilerdata-flow-analysisfunctional-programming
    在 GitHub 上查看↗1,062
  • moonbitlang/moonbit-compilermoonbitlang 的头像

    moonbitlang/moonbit-compiler

    665在 GitHub 上查看↗

    MoonBit is a programming language compiler and build system designed to translate source code into high-performance binaries for web, browser, and native environments. It provides a unified toolchain that manages the entire development lifecycle, from source code compilation and dependency management to the final bundling of applications. The platform distinguishes itself through a modular architecture that enables cross-environment logic sharing, allowing developers to maintain a single codebase that targets both WebAssembly and native machine code. By utilizing a strongly-typed intermediate

    Uses a rigorous internal type system to validate code correctness and perform cross-platform optimizations before final binary generation.

    OCaml
    在 GitHub 上查看↗665
  1. Home
  2. Programming Languages & Runtimes
  3. Machine Code Generation
  4. Intermediate Representations

探索子标签

  • AnimationTarget-neutral schemas that translate design tool compositions into runtime-ready animation specifications. **Distinct from Intermediate Representations:** Specializes general intermediate representations (f6_mt3) by focusing on design-to-playback animation mapping rather than machine code generation.
  • C-Like Intermediate LanguagesIntermediate representations that use C-like syntax and semantics as a stepping stone to native code generation. **Distinct from Intermediate Representations:** Distinct from general Intermediate Representations: specifically uses a C-like syntax and semantics as the intermediate format, not an abstract or target-neutral IR.
  • Data Structure RepresentationsIntermediate formats used to define data structures independently of a target language's memory layout. **Distinct from Intermediate Representations:** Distinct from Intermediate Representations: focuses on data structure layouts for serialization rather than instruction sets for execution.
  • Documentation DescriptorsIntermediate tree structures specifically designed for decoupled API documentation analysis and rendering. **Distinct from Intermediate Representations:** Distinct from general machine-code intermediate representations: focuses on documentation metadata rather than executable instructions.
  • Hardware IR GraphsDirected graphs representing hardware cells and nets used for logic analysis and transformation. **Distinct from Intermediate Representations:** Distinct from general machine-code IRs by representing structural hardware nets and cells.
  • IR LoweringThe process of transforming high-level intermediate representations into lower-level primitive functions. **Distinct from Intermediate Representations:** Focuses on the transition from high-level graphs to low-level primitives, rather than just the definition of the IR.
  • Model Architecture Code GeneratorsTools that generate native framework code from intermediate model representations. **Distinct from Intermediate Representations:** Distinct from general Intermediate Representations by focusing on the generation of executable source code for neural networks.
  • Static Single Assignment FormsIntermediate representations where each variable is assigned exactly once to simplify data flow analysis. **Distinct from Intermediate Representations:** Specializes general intermediate representations into the specific Static Single Assignment (SSA) form.
  • Statically Typed1 个子标签Internal formats that preserve type safety during translation from source to machine code. **Distinct from Intermediate Representations:** Specifically emphasizes the preservation of static types within the IR, unlike general target-neutral IRs.