rav1e Supported Operating Systems and Platforms

This article provides a quick overview of the operating systems, hardware architectures, and platforms officially supported by rav1e, the fastest and safest AV1 video encoder. Because rav1e is written in Rust and exposes a C-compatible API (librav1e), it enjoys broad compatibility across major desktop, server, and mobile environments, utilizing platform-specific assembly optimizations to maximize encoding speeds.

Supported Operating Systems

rav1e is designed to be highly portable and officially supports the following operating systems:

• Linux: Fully supported with native packages available for most major distributions (including Debian, Ubuntu, Fedora, Arch Linux, and Gentoo).
• Windows: Fully supported. It can be compiled natively using MSVC (Microsoft Visual C++) or MinGW, and pre-compiled binaries are regularly provided.
• macOS: Fully supported on both Intel-based Macs and Apple Silicon (M1/M2/M3) Macs. It is easily installable via package managers like Homebrew.
• BSD: Supported on FreeBSD, OpenBSD, and NetBSD.
• Android: Supported via cross-compilation using the Android NDK.
• iOS: Supported via cross-compilation for integration into iOS applications.

Supported Processor Architectures

While rav1e can run on any platform where the Rust compiler (rustc) is supported, optimal performance is achieved on platforms with hand-written assembly optimizations. The officially supported architectures include:

• x86_64 (64-bit Intel/AMD): Highly optimized. Includes assembly speedups for instruction sets up to AVX2 and AVX-512.
• aarch64 / ARM64 (64-bit ARM): Highly optimized. Features extensive NEON assembly optimizations for Apple Silicon, modern Android devices, and ARM-based servers.
• x86 (32-bit Intel/AMD): Supported, with basic SSE2 acceleration.
• ARMv7 (32-bit ARM): Supported, with limited NEON optimizations.

If rav1e is compiled on an unsupported CPU architecture (such as RISC-V or MIPS), it will fall back to a fully functional, platform-independent C/Rust implementation, though encoding speeds will be significantly slower without assembly acceleration.