prefetch-data-load

What problem does it solve?

This Skill optimizes GPU kernel performance by overlapping data loading latency with computation, significantly reducing execution time for memory-bound loops.

Core Features & Use Cases

• Software Prefetching: Implements double-buffering to preload data for the next iteration while the current one is computing.
• Latency Hiding: Effectively hides global memory load latency behind compute instructions in Triton/Gluon kernels.
• Use Case: Accelerate deep learning inference by optimizing the data loading pipeline for matrix multiplication kernels (MFMA) in attention mechanisms or other compute-intensive operations.

Quick Start

Apply prefetch optimization to the provided Triton/Gluon kernel loop.