2021/11/30 by Garvit Dewan, Dewan, Garvit, Louis Jenkins +1
Computer Science · #Caching and Content Delivery #Data Structures and Algorithms (cs.DS) #Distributed #Distributed and Parallel Computing Systems #Distributed systems and fault tolerance #FOS: Computer and information sciences #Network Packet Processing and Optimization #Parallel #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.2112.00068
openalex publication_date 2021/11/30 · openalex created_date 2022/11/09 · openalex updated_date 2026/07/28
The Partitioned Global Address Space (PGAS), a memory model in which the\nglobal address space is explicitly partitioned across compute nodes in a\ncluster, strives to bridge the gap between shared-memory and distributed-memory\nprogramming. To further bridge this gap, there has been an adoption of\nglobal-view distributed data structures, such as 'global arrays' or\n'distributed arrays'. This work demonstrates how shared-memory data structures\ncan be modified to scale in distributed memory. Presented in this work is the\nDistributed Interlocked Hash Table (DIHT), a global-view distributed map data\nstructure inpired by the Interlocked Hash Table (IHT). At 64 nodes with 44\ncores per node, DIHT provides upto 110x the performance of the Chapel\nstandard-library HashedDist.\n