Research in Signal-Processing On-Demand Architecture (SODA) for Software
Defined Radio (SDR) was selected for inclusion in IEEE Micro’s annual
Special Issue (Jan/Feb 2007) on Micro’s Top Picks from Computer Architecture
The paper, “SODA: A High-Performance DSP Architecture for
Software-Defined Radio,” was authored by U-M graduate students Yuan Lin,
Hyunseok Lee, and Mark Woh, Yoav Harel from Intel, U-M faculty
Scott Mahlke and Bredt
Family Professor of Engineering
Trevor Mudge, Arizona State U. faculty Chaitali Chakrabarti, and U-M
alumnus Krisztián Flautner, now at ARM Ltd.
This year’s issue included eleven papers from the past year’s most
significant research publications in computer architecture based on both
novelty and the potential for short or long term impact on industry.
Signal-Processing On-Demand Architecture (SODA) is a fully programmable
architecture that was developed to accommodate the requirements of
software-defined radio (SDR), which the authors describe as belonging to an
emerging class of applications that combine the processing requirements of a
supercomputer with the power requirements of a mobile terminal. The paper
describes how two widely differing wireless protocols, W-CDMA and 802.11a,
are able to meet power-performance requirements by separating control and
data processing and by employing ultrawide SIMD execution.