DISTRIBUTED COMPUTING SEMINAR (TELE-SEMINAR) Black-box Concurrent Data Structures for NUMA Architectures Dr. Irina Calciu VMware Research 10:30 am November 17, 2017 (Friday) 239 CSL ABSTRACT: Modern data centers increasingly employ non-uniform memory access (NUMA) machines. To fully leverage these machines, programmers need efficient concurrent data structures that are NUMA-aware to avoid the performance pitfalls of NUMA. Unfortunately, designing concurrent data structures is hard, and NUMA-awareness makes it even harder. In this work, we show how to obtain NUMA-aware concurrent data structures automatically. We propose an algorithm called Node Replication (NR), which transforms any sequential data structure into a NUMA-aware concurrent data structure. NR draws ideas from two disciplines: shared-memory algorithms and distributed systems. Briefly, NR implements a NUMA-aware shared log, and then uses the log to replicate data structures consistently across NUMA nodes. BIO: Irina Calciu is a researcher at VMware Research working on concurrent algorithms and data structures, and on systems for non-volatile memory and rack-scale computing. Prior to joining VMware, she completed a PhD at Brown University, working with Maurice Herlihy and Justin Gottschlich (Intel Labs) on algorithms for non-uniform memory access (NUMA) architectures and hybrid transactional memory. Irina has co-authored papers at top conferences, obtaining Best Paper Awards at ASPLOS and TRANSACT. She has been awarded a Kanellakis Fellowship in 2014 and has been a research intern at Microsoft Research, Intel Labs, Oracle Labs, Google and Mozilla. Seminar website: disc.ece.illinois.edu/disc-seminar.php