BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Zero-copy programming - Stephen Kell (University of Cambridge)
DTSTART:20100216T130000Z
DTEND:20100216T134500Z
UID:TALK23257@talks.cam.ac.uk
CONTACT:Stephen Kell
DESCRIPTION:In this "crazy idea" talklet\, I will begin by claiming that m
 emory remains a scarce resource\, in contrast to the relative plenty of CP
 U\, disk space etc. on modern machines. Many programs contain redundant co
 pies of information\, which might be collapsible by OS-level mechanisms (s
 hared memory\, memory-mapped I/O). Sharing-recovery approaches (like Sator
 i) rely on copies being bitwise-verbatim\, whereas I estimate that more co
 pies are found in somewhat transcoded form\, where there is some small com
 putational distance between different copies' representations. I will desc
 ribe a few examples of such programs\, a few ideas that might help save me
 mory in these cases\, and a few foreseeable challenges.
LOCATION:Computer Laboratory\, William Gates Building\, Room FW11
END:VEVENT
END:VCALENDAR