Research led by Assistant Professor Kevin Angstadt ‘14 was not too long ago printed in the September/October 2022 particular difficulty of IEEE Micro on Compiling for Accelerators.
Demands by enterprise leaders for real-time analyses of information and the growing technical challenges of designing sooner general-purpose processors have led to the event and adoption of specialised, application-specific laptop {hardware}, generally known as accelerators.
Angstadt’s paper, Synthesizing Legacy String Code for FPGAs Using Bounded Automata Learning, focuses on mechanically changing present software program to run on these new accelerators. Existing programming methods require specialised coaching and data, however the brand new strategy offered by Angstadt and his collaborators can cut back this burden considerably for sure sorts of packages used throughout many domains, together with virus scanning, community safety, social community evaluation, machine studying, and bioinformatics. The paper printed in IEEE Micro additionally consists of experimental outcomes demonstrating that Angstadt’s new strategy additionally improves the velocity and reduces the {hardware} necessities of ensuing packages compared with the present trade commonplace instruments.
This publication was the results of collaboration between Kevin Angstadt at St. Lawrence University, Tommy Tracy II and Kevin Skadron on the University of Virginia, and Jean-Baptiste Jeannin and Westley Weimer on the University of Michigan. The work was supported in half by the National Science Foundation, Air Force Research Labs, Jefferson Scholars Foundation, and the Center for Research in Intelligent Storage and Processing in Memory (CRISP), one in all six facilities in the Joint University Microelectronics Program (JUMP), a Semiconductor Research Corporation (SRC) program sponsored by the Defense Advanced Research Projects Agency (DARPA).
https://news.google.com/__i/rss/rd/articles/CBMiX2h0dHBzOi8vd3d3LnN0bGF3dS5lZHUvbmV3cy9jb21wdXRlci1zY2llbmNlLXByb2Zlc3Nvci1wdWJsaXNoZXMtcmVzZWFyY2gtam91cm5hbC1zcGVjaWFsLWlzc3Vl0gEA?oc=5
