Crafty: Efficient, HTM-Compatible Persistent Transactions
Byte-addressable persistent memory, such as Intel/Micron 3D XPoint, is an emerging technology that bridges the gap between volatile memory and persistent storage.
Data in persistent memory survives crashes and restarts; however, it is challenging to ensure that this data is consistent after failures. Existing approaches incur significant performance costs to ensure crash consistency.
This paper introduces <em>Crafty</em>, a new approach for ensuring consistency and atomicity on persistent memory operations using <em>commodity hardware</em> with existing hardware transactional memory (HTM) capabilities, while incurring low overhead. Crafty employs a novel technique called <em>nondestructive undo logging</em> that leverages commodity HTM to control persist ordering.
Our evaluation shows that Crafty outperforms state-of-the-art prior work
under low contention, and performs competitively under high
contention.
Fri 22 OctDisplayed time zone: Central Time (US & Canada) change
10:50 - 12:10 | PLDI 2020 Papers 4SIGPLAN Papers at Zurich G Chair(s): Muhammad Usman University of Texas at Austin, USA | ||
10:50 15mTalk | BlankIt Library Debloating: Getting What You Want Instead of Cutting What You Don’t SIGPLAN Papers Chris Porter Georgia Institute of Technology, USA, Girish Mururu Georgia Institute of Technology, Prithayan Barua Georgia Institute of Technology, USA, Santosh Pande Georgia Institute of Technology | ||
11:05 15mTalk | Crafty: Efficient, HTM-Compatible Persistent Transactions SIGPLAN Papers Kaan Genç Ohio State University, USA, Michael D. Bond Ohio State University, USA, Guoqing Harry Xu University of California at Los Angeles | ||
11:20 15mTalk | SCAF: A Speculation-Aware Collaborative Dependence Analysis Framework SIGPLAN Papers Sotiris Apostolakis Google, Ziyang Xu Princeton University, Zujun Tan Princeton University, USA, Greg Chan Princeton University, USA, Simone Campanoni Northwestern University, USA, David I. August Princeton University DOI Pre-print | ||
11:35 15mTalk | Towards a Verified Range Analysis for JavaScript JITs SIGPLAN Papers Fraser Brown Stanford University, USA, John Renner University of California at San Diego, USA, Andres Nötzli Stanford University, USA, Sorin Lerner University of California at San Diego, Hovav Shacham University of Texas at Austin, Deian Stefan University of California at San Diego, USA Link to publication DOI | ||
11:50 20mLive Q&A | Discussion, Questions and Answers SIGPLAN Papers | ||