The Seventh Annual Workshop on the LLVM Compiler Infrastructure in HPC

Workshop held in conjunction with SC21 - Sunday, November 14, 2021 - St. Louis, Missouri, USA

SC21Logo Held in conjunction with SC21: The International Conference for High
Performance Computing, Networking, Storage, and Analysis, and in
cooperation with the IEEE Technical Consortium On High Performance Computing (TCHPC).

SC21Logo SC21Logo



Agenda

Time (CST) Speaker Title
9:00 John Leidel Welcome
9:02 Michael Wong Invited Talk
10:00 Coffee Break
10:30 Utpal Bora OpenMP aware MHP Analysis for Improved Static Data-Race Detection
11:10 Valentin Clement Flacc: Towards OpenACC support for Fortran in the LLVM Ecosystem
11:50 Dounia Khaldi Extending LLVM IR for DPC++ Matrix Support: A Case Study with Intel® Advanced Matrix Extensions (Intel® AMX)
12:30 Lunch Break
2:00 John Leidel Welcome Back - Lightning Talk Intro
2:02 Sandya Mannarswamy Neural Instruction Combiner for LLVM
2:09 Alexey Bader Implementation of SYCL™ Specialization Constants for Intel GPUs
2:16 Mahesh Lakshminarasimhan Optimizing Data Layout Transformations in MLIR
2:23 Alister Johnson Automatic and Customizable Code Rewriting and Refactoring with Clang
2:30 Michael Kruse Loop Transformations using Clang's Abstract Syntax Tree
2:37 William Moses Enzyme: Fast, Language Agnostic, Differentiation of Parallel Programs in LLVM
2:44 Jan Hueckelheim ORAQL: Optimistic Responses to Alias Queries in LLVM
2:51 Atmn Patel Remote OpenMP Offloading
3:00 Coffee Break
3:30 Ruiqin Tian A High Performance Sparse Tensor Algebra Compiler in MLIR
4:10 Ryan Kabrick Toward an Automated Hardware Pipelining LLVM Pass Infrastructure
4:50 Tan Nguyen Facilitating CoDesign with Automatic Code Similarity Learning
5:30 The End

Overview

LLVM has become an integral part of the software-development ecosystem for optimizing compilers, dynamic-language execution engines, source-code analysis and transformation tools, debuggers and linkers, and a whole host of programming-language and toolchain-related components. Now heavily used in both academia and industry, where it allows for rapid development of production-quality tools, LLVM is increasingly used in work targeted at high-performance computing. Research in, and implementation of, program analysis, compilation, execution, and profiling has clearly benefited from the availability of a high-quality, freely-available infrastructure on which to build. This workshop will focus on recent developments, from both academia and industry, that build on LLVM to advance the state of the art in high-performance computing.

Format

This workshop will feature contributed papers, lightning talks, and invited talks focusing on recent developments that build on LLVM to advance the state of the art in high-performance computing.

Topics of interest include, but are not limited to:

Deadlines

Please see the SC21 home page for registration deadlines and other information associated with the parent event. Pending the acceptance of the final workshop proceedings, the selected papers will be will be published by TCHPC.

Submissions

Please submit papers using the SC21 Submissions system by selecting the "SC21 Workshop: LLVM-HPC2021 Full Papers" form. Papers must be in IEEE conference format (templates are available). Papers should be no more than 12 pages (including references and figures) and must be at least eight pages long. Please also note IEEE's Article-Posting Policy.

To submit a lightning talk, please use the "SC21 Workshop: LLVM-HPC2021 Lightning Talks" form. Deadlines for lightning talks will be the same as the full paper dates provided above.

Organizers

Program Committee

Contact Information

Alexis Perry-Holby (aperry@lanl.gov)