JSPIM: A Skew-Aware PIM Accelerator for High-Performance Databases Join and Select Operations

Sabiha Tajdari, Anastasia Ailamaki, Sandhya Dwarkadas · 2025

Read JSPIM: A Skew-Aware PIM Accelerator for High-Performance Databases Join and Select Operations by Sabiha Tajdari, Anastasia Ailamaki, Sandhya Dwarkadas (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-08503.txt.

Description

Sabiha Tajdari, Anastasia Ailamaki, and Sandhya Dwarkadas design a memory-based accelerator for database join queries. Their work appeared in twenty twenty-five. Joins, which match rows across tables, are starved by the slow road between memory and processor.

Their module, JSPIM, computes inside memory itself. Parallel search engines sit in each memory subarray, and redesigned hash tables answer lookups in one step. Skewed data, where a few values dominate, is handled by splitting work within and across memory ranks. That split also cuts wasteful off-chip transfers. Against a modern database engine, joins ran hundreds of times faster. On a full benchmark the gain was two and a half times, at small added cost in space and chip area.

Catalog record

CREATORSabiha Tajdari, Anastasia Ailamaki, Sandhya Dwarkadas
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-08503

Browse the full library · About the Tower of Babel