SparseLNR: Accelerating Sparse Tensor Computations Using Loop Nest Restructuring

TL;DR

SparseLNR introduces loop nest restructuring techniques to optimize sparse tensor computations, significantly enhancing performance by supporting kernel distribution and loop fusion within the TACO compiler framework.

Contribution

It extends TACO's scheduling space with kernel distribution and loop fusion, enabling more efficient sparse tensor algebra code generation.

Findings

Performance improvements on real-world tensor computations

Enhanced data locality and reduced asymptotic complexity

Effective optimization of sparse tensor kernels

Abstract

Sparse tensor algebra computations have become important in many real-world applications like machine learning, scientific simulations, and data mining. Hence, automated code generation and performance optimizations for tensor algebra kernels are paramount. Recent advancements such as the Tensor Algebra Compiler (TACO) greatly generalize and automate the code generation for tensor algebra expressions. However, the code generated by TACO for many important tensor computations remains suboptimal due to the absence of a scheduling directive to support transformations such as distribution/fusion. This paper extends TACO's scheduling space to support kernel distribution/loop fusion in order to reduce asymptotic time complexity and improve locality of complex tensor algebra computations. We develop an intermediate representation (IR) for tensor operations called branched iteration graph which specifies breakdown of the computation into smaller ones (kernel distribution) and then fuse (loop fusion) outermost dimensions of the loop nests, while the innermost dimensions are distributed, to increase data locality. We describe exchanges of intermediate results between space iteration spaces, transformation in the IR, and its programmatic invocation. Finally, we show that the transformation can be used to optimize sparse tensor kernels. Our results show that this new transformation significantly improves the performance of several real-world tensor algebra computations compared to TACO-generated code.