From KTIR to SpyreCode: An MLIR Dataflow Backend for IBM Spyre

发生了什么

TL;DR Torch-spyre’s Inductor decides how each kernel’s work is divided across Spyre’s cores and manages tensor residency between on-chip scratchpad and off-chip memory, and the backend compiler turns those decisions into something the hardware can actually run. The contract between them is represented using KTIR (Kernel Tile IR), an MLIR representation built on a data-parallel view of the accelerator, in which every…

摘要按规则整理自下方来源原文

来源