//! Layer 3 complete compute engines.
//!
//! Each engine is a self-contained GPU compute pipeline: structured input in,
//! compute passes on a real GPU backend, typed output back.
/// Adapter-limit helper for BFS queue sizing.
/// Generic graph dataflow engines.
/// Recursive decode pipeline (base64, hex, url, unicode) with region management.
/// Block decompression pipeline (deflate, LZ4, zstd) over IR compositions.
/// DFA pattern matching engine.