%0 Conference Proceedings
%T Fine Grain Precision Scaling for Datapath Approximations in Digital Signal Processing Systems
%+ University of Texas at Austin [Austin]
%A Lee, Seogoo
%A Gerstlauer, Andreas
%< avec comité de lecture
%( IFIP Advances in Information and Communication Technology
%B 21th IFIP/IEEE International Conference on Very Large Scale Integration - System on a Chip (VLSI-SoC)
%C Istanbul, Turkey
%Y Alex Orailoglu
%Y H. Fatih Ugurdag
%Y Luís Miguel Silveira
%Y Martin Margala
%Y Ricardo Reis
%3 VLSI-SoC: At the Crossroads of Emerging Trends
%V AICT-461
%P 119-143
%8 2013-10-06
%D 2013
%R 10.1007/978-3-319-23799-2_6
%K Power reduction
%K Approximate computing
%K Word length optimization
%K Digital signal processing
%Z Computer Science [cs]Conference papers
%X Finding optimal word lengths in digital signal processing systems has been one of the primary mechanisms for reducing complexity. Recently, this topic has been explored in a broader approximate computing context, where architectures allowing for fine-grain control of hardware or software accuracy have been proposed. One of the obstacles for adoption of fine-grain scaling techniques is that they require determining the precision of all intermediate values at all possible operation points, making simulation-based optimization infeasible. In this chapter, we study efficient analytical heuristics to find optimal sets of word lengths for all variables and operations in a dataflow graph constrained by mean squared error type of metrics. We apply our method to several industrial-strength examples. Our results show a more than 5,000x improvement in optimization time compared to an efficient simulation-based word length optimization method with less than 10 % estimation error across a range of target quality metrics.
%G English
%Z TC 10
%Z WG 10.5
%2 https://inria.hal.science/hal-01380301/document
%2 https://inria.hal.science/hal-01380301/file/367527_1_En_6_Chapter.pdf
%L hal-01380301
%U https://inria.hal.science/hal-01380301
%~ IFIP
%~ IFIP-AICT
%~ IFIP-TC
%~ IFIP-WG
%~ IFIP-HINC
%~ IFIP-AICT-461
%~ IFIP-TC10
%~ IFIP-WG10-5