Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/32500
Title: POWER-AWARE TECHNOLOGY MAPPING AND ROUTING FOR DUAL-VT FPGAS
Authors: LOKE WEI TING
Keywords: Technology Mapping, Routing, Reverse Back Bias, Dual VT, FPGA, EDA
Issue Date: 20-Jan-2012
Citation: LOKE WEI TING (2012-01-20). POWER-AWARE TECHNOLOGY MAPPING AND ROUTING FOR DUAL-VT FPGAS. ScholarBank@NUS Repository.
Abstract: In this thesis, we present a technology mapping and clustering scheme, as well as a novel interconnect routing architecture, for modern FPGAs with programmable dual-VT fabrics. The use of Reverse Back Bias (RBB) in circuit design is recognized today as a feasible strategy for mitigating leakage power, a critical issue as process technologies shrink relentlessly towards sub-nanometre proportions. FPGAs with the ability to adjust fabric VT at configuration time offers the ability to reduce leakage power reduction with minimal or no sacrifice to circuit speed. Today, Altera¿s Stratix-III/IV line of FPGAs already demonstrate the feasibility of a similar architecture, but with dual-VT optimization limited to post-P&R. We explore the limitations of such an approach. We also discuss why a dual-VT solution is superior to a dual-VDD one, an architecture adopted by some of the existing works in academia. Together, these form the basis for the contributions presented in this thesis. The first work presented is RBBMap, a power-aware, dual-VT technology mapping tool, and RBBPack, a dual-VT logic clustering tool. Using an existing baseline tool Emap, the combined use of RBBMap and RBBPack yields an average of 70.95% and 28.30% savings in logic block leakage and total power respectively. The second work explores a completely new domain: a programmable, dual-VT switch box routing architecture. This work holds promise in mitigating leakage power in the interconnect - the largest constituent component of the FPGA, yielding an average of 53.69% and 28.23% savings in leakage power savings on the routing network and in total power respectively.
URI: http://scholarbank.nus.edu.sg/handle/10635/32500
Appears in Collections:Master's Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
rbbthesis_final.pdf2.7 MBAdobe PDF

OPEN

NoneView/Download

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.