Network Design for Wafer-Scale Systems with Wafer-on-Wafer Hybrid Bonding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Iff, Patrick, Bonato, Tommaso, Besta, Maciej, Benini, Luca, Hoefler, Torsten
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918405460721664
author Iff, Patrick
Bonato, Tommaso
Besta, Maciej
Benini, Luca
Hoefler, Torsten
author_facet Iff, Patrick
Bonato, Tommaso
Besta, Maciej
Benini, Luca
Hoefler, Torsten
contents Transformer-based large language models are increasingly constrained by data movement as communication bandwidth drops sharply beyond the chip boundary. Wafer-scale integration using wafer-on-wafer hybrid bonding alleviates this limitation by providing ultra-high bandwidth between reticles on bonded wafers. In this paper, we investigate how the physical placement of reticles on wafers influences the achievable network topology and the resulting communication performance. Starting from a 2D mesh-like baseline, we propose four reticle placements (Aligned, Interleaved, Rotated, and Contoured) that improve throughput by up to 250%, reduce latency by up to 36%, and decrease energy per transmitted byte by up to 38%.
format Preprint
id arxiv_https___arxiv_org_abs_2603_05266
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Network Design for Wafer-Scale Systems with Wafer-on-Wafer Hybrid Bonding
Iff, Patrick
Bonato, Tommaso
Besta, Maciej
Benini, Luca
Hoefler, Torsten
Hardware Architecture
Transformer-based large language models are increasingly constrained by data movement as communication bandwidth drops sharply beyond the chip boundary. Wafer-scale integration using wafer-on-wafer hybrid bonding alleviates this limitation by providing ultra-high bandwidth between reticles on bonded wafers. In this paper, we investigate how the physical placement of reticles on wafers influences the achievable network topology and the resulting communication performance. Starting from a 2D mesh-like baseline, we propose four reticle placements (Aligned, Interleaved, Rotated, and Contoured) that improve throughput by up to 250%, reduce latency by up to 36%, and decrease energy per transmitted byte by up to 38%.
title Network Design for Wafer-Scale Systems with Wafer-on-Wafer Hybrid Bonding
topic Hardware Architecture
url https://arxiv.org/abs/2603.05266