PHPackages krakjoe/ort - PHPackages - PHPackages [Skip to content](#main-content)[PHPackages](/)[Directory](/)[Categories](/categories)[Trending](/trending)[Leaderboard](/leaderboard)[Changelog](/changelog)[Analyze](/analyze)[Collections](/collections)[Log in](/login)[Sign up](/register) 1. [Directory](/) 2. / 3. krakjoe/ort ActivePhp-ext-zend krakjoe/ort =========== High performance tensor mathematics for PHP. 136139[1 issues](https://github.com/krakjoe/ort/issues)[1 PRs](https://github.com/krakjoe/ort/pulls)C Since Feb 3Pushed 3mo ago5 watchersCompare [ Source](https://github.com/krakjoe/ort)[ Packagist](https://packagist.org/packages/krakjoe/ort)[ RSS](/packages/krakjoe-ort/feed)WikiDiscussions develop Synced 1mo ago READMEChangelogDependenciesVersions (1)Used By (0) PHP-ORT ======= [](#php-ort) [![Testing](https://github.com/krakjoe/ort/actions/workflows/testing.yml/badge.svg)](https://github.com/krakjoe/ort/actions/workflows/testing.yml)[![Coverage](https://camo.githubusercontent.com/cfce2ef6d89f7b2b2361143ed0fd1a161345906ee3ae23004f44cef03de5cfba/68747470733a2f2f636f766572616c6c732e696f2f7265706f732f6769746875622f6b72616b6a6f652f6f72742f62616467652e7376673f6272616e63683d646576656c6f70)](https://coveralls.io/github/krakjoe/ort?branch=develop) *PHP - Oh Really, Tensors?* This extension provides high performance tensor mathematics. It supports a wide range of mathematical operations on tensors, including element-wise computations, reductions, linear algebra functions, and more. This extension also includes optional integration with Microsoft's ONNX Runtime for efficient model execution. Features -------- [](#features) - **High-Performance Mathematics**: Hardware Acceleration - CPU (SIMD): WASM, NEON, RISCV64, AVX512, AVX2, SSE4.1, SSE2 - GPU: CUDA - **Hybrid Parallelism**: Automatic work distribution across CPU/GPU - **Comprehensive Type Support**: 10 data types with automatic promotion - **Memory Efficient**: Reference-counted tensors with zero-copy slicing - **ONNX Integration**: Optional and Direct ONNX model loading and inference Quick Start ----------- [](#quick-start) ### Installation [](#installation) ``` # Build extension phpize ./configure --enable-ort make sudo make install ``` ### Installation with ONNX Runtime [](#installation-with-onnx-runtime) ``` # Fetch onnxruntime from github wget https://github.com/microsoft/onnxruntime/releases/download/v1.22.0/onnxruntime-linux-x64-1.22.0.tgz # Install onnxruntime tar -C /usr/local -xvf onnxruntime-linux-x64-1.22.0.tgz --strip-components 1 # Build and install extension phpize ./configure --with-ort-onnx=/usr/local --enable-ort make -j$(nproc) sudo make install ``` *Note: Use microsoft provided releases of ONNX, not brew or ppa or apt provided releases.* ### Installation with CUDA [](#installation-with-cuda) See [docs/gpu.md](docs/gpu.md) for more information about GPU acceleration. #### Post Installation [](#post-installation) Add to your `php.ini`: ``` extension=ort.so ``` ### Basic Usage [](#basic-usage) ``` use ORT\Tensor\Transient; use ORT\Math; // Create tensors $a = new Transient([1000, 1000], $matrix_data, ORT\Tensor::FLOAT); $b = new Transient([1000, 1000], $matrix_data, ORT\Tensor::FLOAT); // High-performance matrix multiplication $result = Math\matmul($a, $b); // Element-wise operations $sum = Math\add($a, $b); $scaled = Math\multiply($a, 2.5); // Scalar operations // Reductions $total = Math\reduce\tensor\sum($a); $row_sums = Math\reduce\axis\sum($a, 1); ``` Performance ----------- [](#performance) - **SIMD Acceleration**: Up to 8x faster with AVX2 vectorization - **Multi-Core Scaling**: Automatic scaling to available CPU cores - **Memory Optimization**: Aligned memory allocation for optimal performance Check your system configuration: ``` echo "Backend: " . (ORT\Math\backend() ?: "scalar") . "\n"; echo "Cores: " . ORT\Math\cores() . "\n"; ``` API Reference (not exhaustive) ------------------------------ [](#api-reference-not-exhaustive) ### Tensor Types [](#tensor-types) - `ORT\Tensor\Transient` - Temporary tensors - `ORT\Tensor\Persistent` - Named, persistent tensors ### Mathematical Operations [](#mathematical-operations) - **Binary**: `add`, `subtract`, `multiply`, `divide`, `pow`, `mod` - **Unary**: `sin`, `cos`, `exp`, `log`, `sqrt`, `abs`, `neg` - **Matrix**: `matmul`, `dot`, `transpose` - **Reduction**: `sum`, `mean`, `min`, `max`, `argmax`, `argmin` - **Transformation**: `softmax` ### ONNX Integration [](#onnx-integration) - `ORT\Model` - Load and manage ONNX models - `ORT\Runtime` - Execute ONNX model inference See [docs/api.md](docs/api.md) for more extensive API reference. Configuration ------------- [](#configuration) ### Environment Variables [](#environment-variables) - `ORT_SCALE_CORES` - Set thread pool size (default: CPU cores) ### Build Options [](#build-options) - `--enable-ort-backend=type` - Enable SIMD optimizations (default: auto) - `--with-ort-onnx` - Link against ONNX Runtime (default: no) ### Backend Types [](#backend-types) - `auto` - detect backend type for platform - `wasm` - use wasm backend - `neon` - use neon backend - `riscv64` - use riscv64 backend - `avx2` - use avx2 backend - `sse41` - use sse41 backend - `sse2` - use sse2 backend - `cuda` - use cuda backend - `none` - disable backend Technical Details ----------------- [](#technical-details) See [docs](docs) for detailed technical documentation. Requirements ------------ [](#requirements) - **PHP**: 7.4+ or 8.0+ - **ONNX Runtime**: 1.16+ (optional) - **Compiler**: GCC 4.8+ or Clang 3.8+ - **CPU**: x86\_64, aarch64, riscv64, wasm License ------- [](#license) This project is licensed under the PHP License 3.01. Contributing ------------ [](#contributing) Contributions are welcome! Please ensure: - Code follows existing style conventions - SIMD implementations include proper fallbacks - Thread safety is maintained - Tests pass across supported PHP versions Support ------- [](#support) For issues and feature requests, please use the project's issue tracker. ### Health Score 26 — LowBetter than 43% of packages Maintenance54 Moderate activity, may be stable Popularity22 Limited adoption so far Community16 Small or concentrated contributor base Maturity12 Early-stage or recently created project Bus Factor1 Top contributor holds 99.2% of commits — single point of failure How is this calculated?**Maintenance (25%)** — Last commit recency, latest release date, and issue-to-star ratio. Uses a 2-year decay window. **Popularity (30%)** — Total and monthly downloads, GitHub stars, and forks. Logarithmic scaling prevents top-heavy scores. **Community (15%)** — Contributors, dependents, forks, watchers, and maintainers. Measures real ecosystem engagement. **Maturity (30%)** — Project age, version count, PHP version support, and release stability. ### Community Maintainers ![](https://www.gravatar.com/avatar/a53e08d9e444763eaca5611e49ff1403e8e0803a648b6cbd0e3a0d33430a2ff4?d=identicon)[krakjoe](/maintainers/krakjoe) --- Top Contributors [![krakjoe](https://avatars.githubusercontent.com/u/2236138?v=4)](https://github.com/krakjoe "krakjoe (393 commits)")[![pronskiy](https://avatars.githubusercontent.com/u/1196825?v=4)](https://github.com/pronskiy "pronskiy (2 commits)")[![ndossche](https://avatars.githubusercontent.com/u/7771979?v=4)](https://github.com/ndossche "ndossche (1 commits)") ### Embed Badge ![Health badge](/badges/krakjoe-ort/health.svg) ``` [![Health](https://phpackages.com/badges/krakjoe-ort/health.svg)](https://phpackages.com/packages/krakjoe-ort) ``` PHPackages © 2026 [Directory](/)[Categories](/categories)[Trending](/trending)[Changelog](/changelog)[Analyze](/analyze)