从循环嵌套到硅片:使用MLIR-AIR将AI工作负载映射到AMD NPU
From Loop Nests to Silicon: Mapping AI Workloads onto AMD NPUs with MLIR-AIR
AI总结:
本文提出MLIR-AIR编译器栈,通过AIR方言将AI工作负载映射到AMD NPU,实现高达78.7%计算效率,并以约150行代码支持多头注意力融合实现。
摘要:
通用编译器抽象了并行性、局部性和同步性,限制了它们在现代空间架构上的有效性。随着现代计算架构越来越依赖对数据移动、执行顺序和计算放置的细粒度控制来提升性能,编译器基础设施必须提供显式机制来编排计算和数据,以充分利用此类架构。我们介绍了MLIR-AIR,一个基于MLIR构建的新型开源编译器栈,它弥合了高层工作负载与AMD NPU等细粒度空间架构之间的语义鸿沟。MLIR-AIR定义了AIR方言,为跨计算和内存资源的异步及分层操作提供了结构化表示。AIR原语使编译器能够编排空间调度、跨硬件区域分布计算,并重叠通信与计算,而无需依赖临时的运行时协调或手动调度。我们通过两个案例研究展示了MLIR-AIR的能力:矩阵乘法和LLaMA 2模型中的多头注意力模块。对于矩阵乘法,MLIR-AIR实现了高达78.7%的计算效率,并生成了与使用底层、接近硬件的MLIR-AIE框架编写的最先进的手工优化矩阵乘法性能几乎相同的实现。对于多头注意力,我们证明了AIR接口支持使用约150行代码的融合实现,使得复杂工作负载能够以可处理的方式表达并高效映射到空间硬件。MLIR-AIR将高层结构化控制流转换为空间程序,通过编译器管理的调度,利用异步执行、分块和通信重叠,高效利用NPU的计算结构和内存层次结构。
英文摘要:
General-purpose compilers abstract away parallelism, locality, and synchronization, limiting their effectiveness on modern spatial architectures. As modern computing architectures increasingly rely on fine-grained control over data movement, execution order, and compute placement for performance, compiler infrastructure must provide explicit mechanisms for orchestrating compute and data to fully exploit such architectures. We introduce MLIR-AIR, a novel, open-source compiler stack built on MLIR that bridges the semantic gap between high-level workloads and fine-grained spatial architectures such as AMD's NPUs. MLIR-AIR defines the AIR dialect, which provides structured representations for asynchronous and hierarchical operations across compute and memory resources. AIR primitives allow the compiler to orchestrate spatial scheduling, distribute computation across hardware regions, and overlap communication with computation without relying on ad hoc runtime coordination or manual scheduling. We demonstrate MLIR-AIR's capabilities through two case studies: matrix multiplication and the multi-head attention block from the LLaMA 2 model. For matrix multiplication, MLIR-AIR achieves up to 78.7% compute efficiency and generates implementations with performance almost identical to state-of-the-art, hand-optimized matrix multiplication written using the lower-level, close-to-metal MLIR-AIE framework. For multi-head attention, we demonstrate that the AIR interface supports fused implementations using approximately 150 lines of code, enabling tractable expression of complex workloads with efficient mapping to spatial hardware. MLIR-AIR transforms high-level structured control flow into spatial programs that efficiently utilize the compute fabric and memory hierarchy of an NPU, leveraging asynchronous execution, tiling, and communication overlap through compiler-managed scheduling.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。