3D-MAPS: 3D Massively parallel processor with stacked memory

Dae Hyun Kim, Krit Athikulwongse, Michael Healy, Mohammad Hossain, Moongon Jung, Ilya Khorosh, Gokul Kumar, Young Joon Lee, Dean Lewis, Tzu Wei Lin, Chang Liu, Shreepad Panth, Mohit Pathak, Minzhen Ren, Guanhao Shen, Taigon Song, Dong Hyuk Woo, Xin Zhao, Joungho Kim, Ho ChoiGabriel Loh, Hsien Hsin Lee, Sung Kyu Lim

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

140 Scopus citations

Abstract

Several recent works have demonstrated the benefits of through-silicon-via (TSV) based 3D integration [1-4], but none of them involves a fully functioning multicore processor and memory stacking. 3D-MAPS (3D Massively Parallel Processor with Stacked Memory) is a two-tier 3D IC, where the logic die consists of 64 general-purpose processor cores running at 277MHz, and the memory die contains 256KB SRAM (see Fig. 10.6.1). Fabrication is done using 130nm GlobalFoundries device technology and Tezzaron TSV and bonding technology. Packaging is done by Amkor. This processor contains 33M transistors, 50K TSVs, and 50K face-to-face connections in 5x5mm 2 footprint. The chip runs at 1.5V and consumes up to 4W, resulting in 16W/cm 2 power density. The core architecture is developed from scratch to benefit from single-cycle access to SRAM.

Original languageEnglish
Title of host publication2012 IEEE International Solid-State Circuits Conference, ISSCC 2012 - Digest of Technical Papers
Pages188-189
Number of pages2
DOIs
StatePublished - 2012
Event59th International Solid-State Circuits Conference, ISSCC 2012 - San Francisco, CA, United States
Duration: 19 Feb 201223 Feb 2012

Publication series

NameDigest of Technical Papers - IEEE International Solid-State Circuits Conference
Volume55
ISSN (Print)0193-6530

Conference

Conference59th International Solid-State Circuits Conference, ISSCC 2012
Country/TerritoryUnited States
CitySan Francisco, CA
Period19/02/1223/02/12

Fingerprint

Dive into the research topics of '3D-MAPS: 3D Massively parallel processor with stacked memory'. Together they form a unique fingerprint.

Cite this