BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20141118T193000Z
DTEND:20141118T201500Z
LOCATION:386-87
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: The early 1990s was a time of tremendous change and uncertainty in high performance computing.  Multiple visions for future architectures and programming models were competing for market-share, and key abstractions and algorithms were still being formulated.  Today, after nearly 20 years of remarkable stability in core HPC concepts, we are entering another era of uncertainty.  This talk will explore lessons from that long-ago era that might be relevant for the challenges of today.  More specifically, the talk will recount the role of graph theoretic abstractions in helping to mask the complexity of underlying hardware (as highlighted in an impactful paper from SC’95).  Looking forward, how can we protect application scientists from the vastly greater complexity of upcoming platforms?
SUMMARY:Graphs and HPC in the '90s: A Distant Mirror or Merely Distant?
PRIORITY:3
END:VEVENT
END:VCALENDAR