FusedKernelLibrary

Build the FusedKernelLibrary tests and run them. Covers CMake configuration for Linux

Write user code with the Fused Kernel Library (FKL) — compose fused GPU/CPU pipelines with executeOperations, manage Ptr2D/Tensor data, streams, and runtime parameters. Use when writing an application or library that calls FKL, when porting OpenCV-style image pipelines to fused kernels, or when you need the canonical pipeline patterns (DNN preprocessing, multi-ROI crop, color conversion).

Wrap FKL from another language (Python, Rust, Julia...) using an AST/graph approach and JIT compilation. Covers why NVRTC fails, the Pointer Array ABI (void** for aligned parameters and I/O), generating C++ source strings for IOps and static DPP execution, zero-copy interop, and cross-platform shared library generation (.so/.dll). Use when building or extending a language binding for FKL.

Implement a new Operation for the Fused Kernel Library — Unary, Binary, Ternary, Read, Write, ReadBack or IncompleteReadBack structs, with Parent aliases, DECLARE_*_PARENT macros, build() overloads and unit tests. Use when adding an algorithm to FKL (new arithmetic, color conversion, geometric transform, memory pattern) or when a FusedOperation/alias fails to compile.

Use this skill to classify any FKL work as an Operation (Op), Data Parallel Pattern (DPP), or constexpr_lib. Trigger this before writing a new algorithm, when adapting an existing kernel to FKL standards, or when addressing PR review comments like 'should be a DPP' or 'not FKL'.

FKL data structures — Ptr2D, Tensor, TensorT, RawPtr, PtrDims, MemType, constructors and memory layouts (packed, planar CHW, transposed T3D). Use when allocating or wrapping GPU memory for FKL pipelines, when interfacing external pointers (torch/cupy buffers), or when a Tensor/TensorT constructor or pitch issue appears.

Choose and combine FKL's four fusion techniques — Vertical Fusion (VF), Backwards Vertical Fusion (BVF), Horizontal Fusion (HF) and Divergent Horizontal Fusion (DHF) — as described in the paper (arXiv:2508.07071). Use when deciding how to structure a pipeline for maximum fusion, when batching ROIs or images, or when different data planes need different processing in one kernel.

Implement a new Data Parallel Pattern (DPP) for the Fused Kernel Library. Covers the anatomy of a DPP struct, device-side invocation of IOps (the IOp-form table, reading inputs, fusing epilogues into the write), compile-time vs runtime types, vector types, and testing. Use when adding an algorithm to FKL that requires thread coordination, or when debugging a DPP's internal device execution.

Invoke and compose FKL Operations and Instantiable Operations (IOps) from the CONSUMER side (Host API). Covers how to create inputs, compute chains, and outputs via ::build(), and how to pass them sequentially to the Executor. Use when writing host code to launch FKL pipelines.