BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Computer Laboratory Programming Research Group Sem
 inar
SUMMARY:Improving Implicit Parallelism - Jose Calderon\, U
 niversity of York
DTSTART;TZID=Europe/London:20150612T140000
DTEND;TZID=Europe/London:20150612T150000
UID:TALK59544AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/59544
DESCRIPTION:Using static analysis techniques compilers for laz
 y functional languages can identify parts of a pro
 gram that can be legitimately evaluated in paralle
 l with the main thread of execution. These techniq
 ues can produce improvements in the runtime perfor
 mance of a program\, but are limited by the static
  analyses’ poor prediction of runtime performance.
  This talk outlines the development of a system th
 at uses iterative compilation in addition to well-
 studied static analysis techniques. Our representa
 tion of the parallel programs allows us to use tra
 ditional 'blind' search techniques or profile-dire
 cted improvement. We compare the results of differ
 ent search strategies and discuss the pitfalls and
  limitations of our technique. Overall\, the use o
 f iterative feedback allows us to achieve higher p
 erformance than through static analysis alone.
LOCATION:SS03
CONTACT:Raphael Proust
END:VEVENT
END:VCALENDAR