High Performance Code Generation in MLIR: An Early Case Study with GEMM

03/01/2020
by   Uday Bondhugula, et al.
0

This article is primarily meant to present an early case study on using MLIR, a new compiler intermediate representation infrastructure, for high-performance code generation. Aspects of MLIR covered in particular include memrefs, the affine dialect, and polyhedral utilities and pass infrastructure surrounding those. This article is also aimed at showing the role compiler infrastructure could play in generating code that is competitive with highly tuned manually developed libraries, albeit in a more modular, reusable, and automatable way.

READ FULL TEXT

page 17

page 21

page 22

research
08/23/2021

High Performance GPU Code Generation for Matrix-Matrix Multiplication using MLIR: Some Early Results

This report presents some early results on code generation targeting ten...
research
11/08/2017

DLVM: A modern compiler infrastructure for deep learning systems

Deep learning software demands reliability and performance. However, man...
research
03/25/2022

WEcharge: democratizing EV charging infrastructure

The sustainable growth of EVs will have to be met with proportional grow...
research
12/11/2010

On the Implementation of GNU Prolog

GNU Prolog is a general-purpose implementation of the Prolog language, w...
research
07/09/2018

Architecture and performance of Devito, a system for automated stencil computation

Stencil computations are a key part of many high-performance computing a...
research
02/07/2022

Composable and Modular Code Generation in MLIR: A Structured and Retargetable Approach to Tensor Compiler Construction

Despite significant investment in software infrastructure, machine learn...
research
03/19/2018

Beyond correct and fast: Inspection Testing

Occasionally, developers need to ensure that the compiler treats their c...

Please sign up or login with your details

Forgot password? Click here to reset