CGO 2020
18th ACM/IEEE International Symposium on Code Generation and Optimization (CGO)
Powered by
Conference Publishing Consulting

18th ACM/IEEE International Symposium on Code Generation and Optimization (CGO), February 22–26, 2020, San Diego, CA, USA

CGO 2020 – Proceedings

Contents - Abstracts - Authors

Frontmatter

Title Page
Welcome form the Steering Committee
Message from the Program Chairs
CGO 2020 Organization
Report from the Artifact Evaluation Committee
Sponsors

Dynamic Languages

Efficient Nursery Sizing for Managed Languages on Multi-core Processors with Shared Caches
Mohamed Ismail and G. Edward Suh ORCID logo
(Cornell University, USA)
Publisher's Version Artifacts Reusable Results Replicated
Type Freezing: Exploiting Attribute Type Monomorphism in Tracing JIT Compilers
Lin Cheng ORCID logo, Berkin Ilbeyi, Carl Friedrich Bolz-Tereick, and Christopher Batten ORCID logo
(Cornell University, USA; University of Düsseldorf, Germany)
Publisher's Version Artifacts Reusable Results Replicated

Safety and Reliability

Low-Cost Prediction-Based Fault Protection Strategy
Sunghyun Park ORCID logo, Shikai Li, Ze Zhang, and Scott Mahlke
(University of Michigan, USA)
Publisher's Version
Secure Automatic Bounds Checking: Prevention Is Simpler Than Cure
Ejebagom John Ojogbo, Mithuna Thottethodi, and T. N. Vijaykumar
(Purdue University, USA)
Publisher's Version
Aloe: Verifying Reliability of Approximate Programs in the Presence of Recovery Mechanisms
Keyur Joshi, Vimuth Fernando, and Sasa MisailovicORCID logo
(University of Illinois at Urbana-Champaign, USA)
Publisher's Version
Interactive Debugging of Concurrent Programs under Relaxed Memory Models
Aakanksha Verma, Pankaj Kumar Kalita ORCID logo, Awanish Pandey, and Subhajit Roy
(IIT Kanpur, India)
Publisher's Version

Best Paper Finalists

Testing Static Analyses for Precision and Soundness
Jubi Taneja, Zhengyang Liu, and John Regehr ORCID logo
(University of Utah, USA)
Publisher's Version Artifacts Reusable Results Replicated
HALO: Post-Link Heap-Layout Optimisation
Joe Savage and Timothy M. JonesORCID logo
(University of Cambridge, UK)
Publisher's Version Artifacts Reusable Results Replicated
Efficient and Scalable Cross-ISA Virtualization of Hardware Transactional Memory
Wenwen Wang, Pen-Chung Yew, Antonia Zhai, and Stephen McCamant
(University of Georgia, USA; University of Minnesota, USA)
Publisher's Version

GPUs

Speculative Reconvergence for Improved SIMT Efficiency
Sana Damani, Daniel R. Johnson, Mark Stephenson ORCID logo, Stephen W. Keckler ORCID logo, Eddie Yan, Michael McKeown, and Olivier Giroux
(Georgia Institute of Technology, USA; NVIDIA, USA; University of Washington, USA; Esperanto Technologies, USA)
Publisher's Version
Optimizing Occupancy and ILP on the GPU using a Combinatorial Approach
Ghassan ShobakiORCID logo, Austin Kerbow ORCID logo, and Stanislav Mekhanoshin
(California State University at Sacramento, USA; Advanced Micro Devices, USA)
Publisher's Version

Compilation for Specialized Domains

Multi-layer Optimizations for End-to-End Data Analytics
Amir Shaikhha, Maximilian Schleich, Alexandru Ghita, and Dan Olteanu
(University of Oxford, UK)
Publisher's Version
Optimizing Ordered Graph Algorithms with GraphIt
Yunming Zhang, Ajay BrahmakshatriyaORCID logo, Xinyi Chen, Laxman Dhulipala, Shoaib Kamil, Saman AmarasingheORCID logo, and Julian Shun
(Massachusetts Institute of Technology, USA; Carnegie Mellon University, USA; Adobe Research, USA)
Publisher's Version Artifacts Functional Results Replicated
A Performance-Optimizing Compiler for Cyber-Physical Digital Microfluidic Biochips
Tyson Loveless, Jason Ott, and Philip Brisk ORCID logo
(University of California at Riverside, USA)
Publisher's Version
CogniCryptGEN: Generating Code for the Secure Usage of Crypto APIs
Stefan Krüger, Karim Ali, and Eric BoddenORCID logo
(University of Paderborn, Germany; University of Alberta, Canada; Fraunhofer IEM, Germany)
Publisher's Version Artifacts Functional Results Replicated

Tool and Practical Experience Papers

AN5D: Automated Stencil Framework for High-Degree Temporal Blocking on GPUs
Kazuaki Matsumura, Hamid Reza Zohouri, Mohamed Wahib, Toshio Endo ORCID logo, and Satoshi Matsuoka
(Barcelona Supercomputing Center, Spain; Edgecortix, Japan; AIST, Japan; Tokyo Institute of Technology, Japan; RIKEN CCS, Japan)
Publisher's Version
The Design and Implementation of the Wolfram Language Compiler
Abdul Dakkak, Tom Wickham-Jones, and Wen-mei Hwu ORCID logo
(University of Illinois at Urbana-Champaign, USA; Wolfram Research, UK)
Publisher's Version Artifacts Functional
SIMD Support in .NET: Abstract and Concrete Vector Types and Operations
Carol Eidt and Tanner Gooding
(Microsoft, USA)
Publisher's Version Artifacts Reusable Results Replicated

Code Optimization

NeuroVectorizer: End-to-End Vectorization with Deep Reinforcement Learning
Ameer Haj-Ali, Nesreen K. Ahmed, Ted Willke, Yakun Sophia Shao, Krste Asanovic, and Ion Stoica
(University of California at Berkeley, USA; Intel Labs, USA)
Publisher's Version Info Artifacts Reusable Results Replicated
Introducing the Pseudorandom Value Generator Selection in the Compilation Toolchain
Michael Leonard and Simone Campanoni ORCID logo
(Northwestern University, USA)
Publisher's Version

Heterogeneity and Parallelism

COLAB: A Collaborative Multi-factor Scheduler for Asymmetric Multicore Processors
Teng Yu, Pavlos Petoumenos, Vladimir Janjic, Hugh Leather, and John Thomson
(University of St. Andrews, UK; University of Manchester, UK; University of Edinburgh, UK)
Publisher's Version
PreScaler: An Efficient System-Aware Precision Scaling Framework on Heterogeneous Systems
Seokwon Kang, Kyunghwan Choi, and Yongjun Park
(Hanyang University, South Korea)
Publisher's Version Artifacts Functional Results Replicated
ATMem: Adaptive Data Placement in Graph Applications on Heterogeneous Memories
Yu Chen, Ivy B. Peng, Zhen Peng, Xu Liu, and Bin Ren
(College of William and Mary, USA; Lawrence Livermore National Laboratory, USA)
Publisher's Version

Code Generation and Transformation

Automatic Generation of High-Performance Quantized Machine Learning Kernels
Meghan Cowan, Thierry Moreau, Tianqi Chen, James BornholtORCID logo, and Luis Ceze ORCID logo
(University of Washington, USA; University of Texas at Austin, USA)
Publisher's Version Artifacts Functional Results Replicated
Deriving Parametric Multi-way Recursive Divide-and-Conquer Dynamic Programming Algorithms using Polyhedral Compilers
Mohammad Mahdi Javanmard, Zafar Ahmad, Martin Kong, Louis-Noël Pouchet, Rezaul Chowdhury, and Robert Harrison
(Stony Brook University, USA; University of Oklahoma, USA; Colorado State University, USA)
Publisher's Version

proc time: 5.76