Babak Falsafi

File:Prof. Babak Falsafi.png

Babak Falsafi is a Swiss computer scientist specializing in computer architecture and digital platform design. He is the founding director of EcoCloud at EPFL, an industrial/academic consortium investigating efficient and intelligent data-centric technologies. He is a professor in the School of Computer and Communication Sciences at EPFL. Prior to that he was a professor of electrical and computer engineering at Carnegie Mellon University, and an assistant professor of electrical and computer engineering at Purdue University. He holds a bachelor's degree in computer science, a bachelor's degree in electrical and computer engineering (both summa cum laude) with distinctions from SUNY Buffalo, and a master's degree and PhD in computer science from University Wisconsin - Madison.

He has made numerous contributions to computer system design and evaluation including a server architecture{{cite journal |last1=Falsafi |first1=Babak |last2=Wood |first2=David A. |title=Reactive NUMA: A design for unifying S-COMA and CC-NUMA |journal=ACM SIGARCH Computer Architecture News |volume=25 |issue=2 |date=1997 |pages=229–240 |doi=10.1145/384286.264205 |url=http://digital.library.wisc.edu/1793/9214 }} which laid the foundation for Sun Microsystems' NUMA machines,{{cite book |last1=Hagersten |first1=Erik |last2=Koster |first2=Michael |title=Proceedings Fifth International Symposium on High-Performance Computer Architecture |chapter=WildFire: A scalable path for SMPS |date=1999 |pages=172–181 |chapter-url=https://dl.acm.org/citation.cfm?id=822785|isbn=9780769500041 |series=Hpca '99 |doi=10.1109/HPCA.1999.744361 |citeseerx=10.1.1.21.8882 |s2cid=13915228 }} technologies to minimize (leakage) power in the memory system in the absence of activity (Supply Gating){{cite book |last1=Powell |first1=Michael |last2=Yang |first2=Se-Hyun |last3=Falsafi |first3=Babak |last4=Roy |first4=Kaushik |last5=Vijaykumar |first5=T. N. |title=Proceedings of the 2000 international symposium on Low power electronics and design - ISLPED '00 |chapter=Gated-V_dd |date=2000 |pages=90–95 |chapter-url=https://dl.acm.org/citation.cfm?id=344526|doi=10.1145/344166.344526 |isbn=9781581131901 |s2cid=3241652 }} and in shared memory (Snoop Filtering){{cite book |last1=Moshovos |first1=Andreas |last2=Memik |first2=Gokhan |last3=Falsafi |first3=Babak |last4=Choudhary |first4=Alok |title=Proceedings HPCA Seventh International Symposium on High-Performance Computer Architecture |chapter=JETTY: Filtering snoops for reduced energy consumption in SMP servers |pages=85–96 |date=2001 |doi=10.1109/HPCA.2001.903254 |isbn=0-7695-1019-1 |s2cid=6811066 |chapter-url=http://infoscience.epfl.ch/record/135568/files/hpca01_jetty.pdf }} prevalent in modern CPUs and multi-socket servers, and memory system accelerators {{cite journal |last1=Somogyi |first1=Stephen |last2=Wenisch |first2=Thomas F. |last3=Ailamaki |first3=Anastasia |last4=Falsafi |first4=Babak |last5=Moshovos |first5=Andreas |title=Spatial Memory Streaming |journal=ACM SIGARCH Computer Architecture News |volume=34 |issue=2 |date=2006 |pages=252–263 |doi=10.1145/1150019.1136508 |s2cid=3552110 |url=http://infoscience.epfl.ch/record/112674 }} in modern (ARM) CPUs in mobile platforms. He has shown that hardware memory consistency models are neither necessary (in the 90's){{cite journal |last1=Gniady |first1=Chris |title=Is SC + ILP = RC? |last2=Falsafi |first2=Babak |last3=Vijaykumar |first3=T. N. |journal=ACM SIGARCH Computer Architecture News |volume=27 |issue=2 |date=1999 |pages=162–171 |doi=10.1145/307338.300993 |doi-access=free }} nor sufficient (a decade later){{cite book |last1=Wenisch |first1=Thomas F. |last2=Ailamaki |first2=Anastasia |last3=Falsafi |first3=Babak |last4=Moshovos |first4=Andreas |title=Proceedings of the 34th annual international symposium on Computer architecture |chapter=Mechanisms for store-wait-free multiprocessors |pages=266–277 |chapter-url=https://dl.acm.org/citation.cfm?id=1250696|doi=10.1145/1250662.1250696 |year=2007 |isbn=9781595937063 |series=Isca '07 |volume=35 |issue=2 |s2cid=1024059 |url=https://infoscience.epfl.ch/record/112672/files/store-wait.pdf }} to achieve high performance in multiprocessor systems. These results eventually led to fence speculation in modern (x86) CPUs. He argued and demonstrated that the slowdown in silicon efficiency (Dennard's Law) and density scaling (Moore's Law) would lead to Dark Silicon and specialization in servers.{{cite journal |last1=Hardavellas |first1=Nikos |last2=Ferdman |first2=Michael |last3=Falsafi |first3=Babak |last4=Ailamaki |first4=Anastasia |title=Toward Dark Silicon in Servers |journal=IEEE Micro |date=2011 |volume=31 |issue=4 |pages=6–15 |doi=10.1109/MM.2011.77 |s2cid=2765349 |url=http://infoscience.epfl.ch/record/168285 }} These results led to a follow-on study on careful characterization of scale-out workloads on server platforms {{cite journal |last1=Ferdman |first1=Michael |last2=Adileh |first2=Almutaz |last3=Kocberber |first3=Onur |last4=Volos |first4=Stavros |last5=Alisafaee |first5=Mohammad |last6=Jevdjic |first6=Djordje |last7=Kaynak |first7=Cansu |last8=Popescu |first8=Adrian Daniel |last9=Ailamaki |first9=Anastasia |last10=Falsafi |first10=Babak |title=Clearing the clouds: A study of emerging scale-out workloads on modern hardware |journal=ACM SIGARCH Computer Architecture News |volume=40 |issue=1 |pages=37–48 |doi=10.1145/2189750.2150982 |date=March 2012 }} which laid the foundation for the first generation of Cavium ARM server CPUs, ThunderX.

He is a recipient of an Alfred P. Sloan Research Fellowship, was named an ACM Fellow in 2015 {{cite web |last1=Falsafi |first1=Babak |title=ACM Fellows |url=https://awards.acm.org/award-winners/FALSAFI_3889037}} for contributions to multiprocessor and memory architecture design and evaluation and a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) in 2012{{Cite web|url=https://www.ieee.org/documents/fellows_class_2012.pdf|archive-url=https://web.archive.org/web/20120215201114/http://www.ieee.org/documents/fellows_class_2012.pdf|url-status=dead|archive-date=February 15, 2012|title=2012 elevated fellow|website=IEEE Fellows Directory}} for contributions to multiprocessor architecture and memory systems.