CGO 2021
2021 IEEE/ACM International Symposium on Code Generation and Optimization (CGO)
Powered by
Conference Publishing Consulting

2021 IEEE/ACM International Symposium on Code Generation and Optimization (CGO), February 27 – March 3, 2021, Virtual, Republic of Korea

CGO 2021 – Proceedings

Contents - Abstracts - Authors

Frontmatter

Title Page
Message from the General Chair
Message from the Program Chairs
CGO 2021 Organization
Report from the Artifact Evaluation Committee
CGO 2021 Sponsors

Keynote

Data Layout and Data Representation Optimizations to Reduce Data Movement (Keynote)
Mary Hall ORCID logo
(University of Utah, USA)

Compiler Infrastructure
(Chair: Michael Kruse, Argonne National Laboratory, USA)

MLIR: Scaling Compiler Infrastructure for Domain Specific Computation
Chris Lattner, Mehdi Amini ORCID logo, Uday Bondhugula ORCID logo, Albert Cohen ORCID logo, Andy Davis ORCID logo, Jacques Pienaar ORCID logo, River Riddle, Tatiana Shpeisman, Nicolas Vasilache ORCID logo, and Oleksandr Zinenko ORCID logo
(Google, USA; Indian Institute of Science, India; Google, France)
Artifacts Available Artifacts Reusable
Progressive Raising in Multi-level IR
Lorenzo Chelini, Andi Drebes, Oleksandr Zinenko ORCID logo, Albert Cohen ORCID logo, Nicolas Vasilache ORCID logo, Tobias Grosser, and Henk Corporaal
(Eindhoven University of Technology, Netherlands; Inria, France; ENS Paris, France; Google, France; Google, Switzerland; University of Edinburgh, UK)
Artifacts Available Artifacts Functional Results Reproduced
Towards a Domain-Extensible Compiler: Optimizing an Image Processing Pipeline on Mobile CPUs
Thomas KoehlerORCID logo and Michel SteuwerORCID logo
(University of Glasgow, UK; University of Edinburgh, UK)
Artifacts Available Artifacts Reusable Results Reproduced
BuildIt: A Type-Based Multi-stage Programming Framework for Code Generation in C++
Ajay BrahmakshatriyaORCID logo and Saman AmarasingheORCID logo
(Massachusetts Institute of Technology, USA)
Artifacts Available Artifacts Reusable Results Reproduced

Dealing with Precision
(Chair: Uma Srinivasan, Twitter)

An Interval Compiler for Sound Floating-Point Computations
Joao Rivera ORCID logo, Franz Franchetti ORCID logo, and Markus Püschel ORCID logo
(ETH Zurich, Switzerland; Carnegie Mellon University, USA)
Artifacts Available Artifacts Reusable Results Reproduced
Seamless Compiler Integration of Variable Precision Floating-Point Arithmetic
Tiago Trevisan Jost ORCID logo, Yves Durand ORCID logo, Christian Fabre ORCID logo, Albert Cohen ORCID logo, and Frédéric Pétrot ORCID logo
(Université Grenoble Alpes, France; CEA LIST, France; Google, France; CNRS, France; Grenoble INP, France; TIMA, France)
Artifacts Functional Results Reproduced
UNIT: Unifying Tensorized Instruction Compilation
Jian WengORCID logo, Animesh Jain, Jie Wang, Leyuan Wang, Yida Wang ORCID logo, and Tony NowatzkiORCID logo
(University of California at Los Angeles, USA; Amazon, USA)
Artifacts Available Artifacts Functional
Unleashing the Low-Precision Computation Potential of Tensor Cores on GPUs
Guangli Li, Jingling Xue ORCID logo, Lei Liu, Xueying Wang, Xiu Ma, Xiao Dong, Jiansong Li, and Xiaobing Feng ORCID logo
(Institute of Computing Technology at Chinese Academy of Sciences, China; University of Chinese Academy of Sciences, China; UNSW, Australia; Jilin University, China)

Binary Profiling, Tracing, Sampling
(Chair: Wei Wang, University of Texas at San Antonio, USA)

Cinnamon: A Domain-Specific Language for Binary Profiling and Monitoring
Mahwish Arif, Ruoyu Zhou, Hsi-Ming Ho, and Timothy M. JonesORCID logo
(University of Cambridge, UK; University of Sussex, UK)
GPA: A GPU Performance Advisor Based on Instruction Sampling
Keren Zhou ORCID logo, Xiaozhu Meng, Ryuichi Sai ORCID logo, and John Mellor-Crummey
(Rice University, USA)
Artifacts Available Artifacts Reusable Results Reproduced
ELFies: Executable Region Checkpoints for Performance Analysis and Simulation
Harish Patil ORCID logo, Alexander Isaev, Wim Heirman ORCID logo, Alen Sabu ORCID logo, Ali Hajiabadi ORCID logo, and Trevor E. Carlson ORCID logo
(Intel Corporation, USA; Intel Corporation, Belgium; National University of Singapore, Singapore)
Artifacts Functional Results Reproduced
Vulkan Vision: Ray Tracing Workload Characterization using Automatic Graphics Instrumentation
David Pankratz, Tyler Nowicki, Ahmed Eltantawy, and José Nelson Amaral ORCID logo
(University of Alberta, Canada; Huawei Technologies, Canada)
Artifacts Functional Results Reproduced

Parallelism - Optimizing, Modeling, Testing
(Chair: Michael O'Boyle, University of Edinburgh, UK)

Loop Parallelization using Dynamic Commutativity Analysis
Christos Vasiladiotis ORCID logo, Roberto Castañeda Lozano, Murray Cole, and Björn Franke ORCID logo
(University of Edinburgh, UK)
Fine-Grained Pipeline Parallelization for Network Function Programs
Seungbin Song, Heelim Choi, and Hanjun Kim
(Yonsei University, South Korea)
YaskSite: Stencil Optimization Techniques Applied to Explicit ODE Methods on Modern Architectures
Christie L. Alappat, Johannes Seiferth, Georg Hager, Matthias Korch, Thomas Rauber, and Gerhard Wellein
(University of Erlangen-Nuremberg, Germany; University of Bayreuth, Germany)
Artifacts Available Artifacts Functional Results Reproduced
GoBench: A Benchmark Suite of Real-World Go Concurrency Bugs
Ting Yuan, Guangwei Li, Jie Lu, Chen Liu, Lian Li, and Jingling Xue ORCID logo
(Institute of Computing Technology at Chinese Academy of Sciences, China; University of Chinese Academy of Sciences, China; UNSW, Australia)
Artifacts Available Artifacts Functional Results Reproduced

Memory Optimization and Safeness
(Chair: Eunjung Park, Los Alamos National Laboratory, USA)

Memory-Safe Elimination of Side Channels
Luigi Soares and Fernando Magno Quintão Pereira ORCID logo
(Federal University of Minas Gerais, Brazil)
Info Artifacts Available Artifacts Reusable Results Reproduced
Variable-Sized Blocks for Locality-Aware SpMV
Naveen NamashivayamORCID logo, Sanyam Mehta, and Pen-Chung Yew
(HPE, USA; University of Minnesota at Twin Cities, USA)
Object Versioning for Flow-Sensitive Pointer Analysis
Mohamad Barbar, Yulei SuiORCID logo, and Shiping Chen
(University of Technology Sydney, Australia; CSIRO's Data61, Australia)
Artifacts Available Artifacts Functional Results Reproduced
Scaling Up the IFDS Algorithm with Efficient Disk-Assisted Computing
Haofeng Li, Haining Meng, Hengjie Zheng, Liqing Cao, Jie Lu, Lian Li, and Lin Gao
(Institute of Computing Technology at Chinese Academy of Sciences, China; University of Chinese Academy of Sciences, China; TianqiSoft, China)
Artifacts Available Artifacts Functional

Compiling Graph Algorithms, Compiling for GPU's
(Chair: Maria Garzaran, Intel Corporation and University of Illinois at Urbana-Champaign, USA)

Compiling Graph Applications for GPUs with GraphIt
Ajay BrahmakshatriyaORCID logo, Yunming Zhang, Changwan Hong, Shoaib Kamil, Julian Shun, and Saman AmarasingheORCID logo
(Massachusetts Institute of Technology, USA; Adobe, USA)
Artifacts Available Artifacts Reusable Results Reproduced
Efficient Execution of Graph Algorithms on CPU with SIMD Extensions
Ruohuang Zheng and Sreepathi Pai ORCID logo
(University of Rochester, USA)
Artifacts Available Artifacts Reusable Results Reproduced
r3d3: Optimized Query Compilation on GPUs
Alexander Krolik, Clark Verbrugge ORCID logo, and Laurie Hendren ORCID logo
(McGill University, Canada)
Info Artifacts Available Artifacts Functional Results Reproduced
C-for-Metal: High Performance SIMD Programming on Intel GPUs
Guei-Yuan Lueh, Kaiyu Chen, Gang Chen, Joel Fuentes ORCID logo, Wei-Yu Chen, Fangwen Fu, Hong Jiang, Hongzheng Li, and Daniel Rhee
(Intel Corporation, USA)
Artifacts Available Artifacts Functional Results Reproduced

Compiling for Spatial, Quantum, and Embedded Devices
(Chair: Wei-Fen Lin, National Cheng Kung University, Taiwan)

Relaxed Peephole Optimization: A Novel Compiler Optimization for Quantum Circuits
Ji Liu ORCID logo, Luciano Bello ORCID logo, and Huiyang Zhou ORCID logo
(North Carolina State University, USA; IBM Research, USA)
Artifacts Available Artifacts Reusable Results Reproduced
StencilFlow: Mapping Large Stencil Programs to Distributed Spatial Computing Systems
Johannes de Fine Licht ORCID logo, Andreas Kuster ORCID logo, Tiziano De MatteisORCID logo, Tal Ben-NunORCID logo, Dominic Hofer ORCID logo, and Torsten Hoefler ORCID logo
(ETH Zurich, Switzerland; MeteoSwiss, Switzerland)
Artifacts Available Artifacts Reusable Results Reproduced
Thread-Aware Area-Efficient High-Level Synthesis Compiler for Embedded Devices
Changsu Kim, Shinnung Jeong, Sungjun Cho ORCID logo, Yongwoo Lee, William Song, Youngsok Kim ORCID logo, and Hanjun Kim
(POSTECH, South Korea; Yonsei University, South Korea)

JIT and Binary Translation; Optimizing for Code-Size
(Chair: Probir Roy, University of Michigan at Dearborn, USA)

HHVM Jump-Start: Boosting Both Warmup and Steady-State Performance at Scale
Guilherme Ottoni and Bin Liu
(Facebook, USA)
Enhancing Atomic Instruction Emulation for Cross-ISA Dynamic Binary Translation
Ziyi ZhaoORCID logo, Zhang Jiang, Ying Chen ORCID logo, Xiaoli Gong ORCID logo, Wenwen Wang, and Pen-Chung Yew
(Nankai University, China; University of Georgia, USA; University of Minnesota at Twin Cities, USA)
Info Artifacts Available Artifacts Reusable Results Reproduced
An Experience with Code-Size Optimization for Production iOS Mobile Applications
Milind Chabbi ORCID logo, Jin Lin, and Raj Barik
(Uber Technologies, USA)
Info Artifacts Available Artifacts Reusable
AnghaBench: A Suite with One Million Compilable C Benchmarks for Code-Size Reduction
Anderson Faustino da Silva ORCID logo, Bruno Conde Kind, José Wesley de Souza Magalhães, Jerônimo Nunes Rocha, Breno Campos Ferreira Guimarães, and Fernando Magno Quintão Pereira ORCID logo
(State University of Maringá, Brazil; Federal University of Minas Gerais, Brazil)
Info

proc time: 0.3