Energy Efficient Computing in FPGA Through Embedded RAM Blocks

Energy Efficient Computing in FPGA Through Embedded RAM Blocks

Bibliographic Details
Main Author: Ghosh, Anandaroop
Language:English
Published: Case Western Reserve University School of Graduate Studies / OhioLINK 2013
Subjects:
Computer Engineering
FPGA
Memory Based Computing
Energy Efficient Computing
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=case1365198486
id ndltd-OhioLink-oai-etd.ohiolink.edu-case1365198486
record_format oai_dc
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-case13651984862021-08-03T05:21:47Z Energy Efficient Computing in FPGA Through Embedded RAM Blocks Ghosh, Anandaroop Computer Engineering FPGA Memory Based Computing Energy Efficient Computing FPGAs have emerged as the preferred prototyping and accelerator platform for diverseapplication domains like digital signal processing (DSP), security, and multimedia having real time performance requirements. Applications in these domains are often dominated by complex compute-intensive operations requiring implementation of complex datapaths or functions e.g. transcendental functions. Conventional spatial mapping of theseoperations to the configurable logic blocks (CLBs) or embedded DSP blocks of a FPGA device imposes a major bottleneck in energy efficiency. In this thesis, we propose to use embedded memory blocks (EMBs) in FPGA for energy-efficient mapping of these operations. We select appropriate parts of an application for mapping into embedded memory blocks in a heterogeneous mapping framework that aims at maximizing energy efficiency. Complex operations are decomposed / fused into large multi-input multi-output look-up tables, mapped into EMBs and evaluated through sequential access of them. Optimal energy configuration of the embedded memory blocks are determined and effectiveness of the proposed methodology is evaluated for a set of common applications using a commercial state-of-the-art FPGA system (Altera Stratix IV). The proposed work also builds a strategy for energy-accuracy trade-off for multimedia applications and leverages the effectiveness of memory based computing in FPGA for such approximate computations. 2013-08-16 English text Case Western Reserve University School of Graduate Studies / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=case1365198486 http://rave.ohiolink.edu/etdc/view?acc_num=case1365198486 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.
collection NDLTD
language English
sources NDLTD
topic Computer Engineering
FPGA
Memory Based Computing
Energy Efficient Computing
spellingShingle Computer Engineering
FPGA
Memory Based Computing
Energy Efficient Computing
Ghosh, Anandaroop
Energy Efficient Computing in FPGA Through Embedded RAM Blocks
author Ghosh, Anandaroop
author_facet Ghosh, Anandaroop
author_sort Ghosh, Anandaroop
title Energy Efficient Computing in FPGA Through Embedded RAM Blocks
title_short Energy Efficient Computing in FPGA Through Embedded RAM Blocks
title_full Energy Efficient Computing in FPGA Through Embedded RAM Blocks
title_fullStr Energy Efficient Computing in FPGA Through Embedded RAM Blocks
title_full_unstemmed Energy Efficient Computing in FPGA Through Embedded RAM Blocks
title_sort energy efficient computing in fpga through embedded ram blocks
publisher Case Western Reserve University School of Graduate Studies / OhioLINK
publishDate 2013
url http://rave.ohiolink.edu/etdc/view?acc_num=case1365198486
work_keys_str_mv AT ghoshanandaroop energyefficientcomputinginfpgathroughembeddedramblocks
_version_ 1719419001586581504

Similar Items