BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:10 years of meataxe development. - Richard Parker (None / Other)
DTSTART:20200128T160500Z
DTEND:20200128T165500Z
UID:TALK137911@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Myself\, Steve Linton and Jon Thackray have been working for n
 early 10 years on a fairly major overhaul of matrix multiplication and Gau
 ssian elimination over finite fields of order (mainly) up to 1\,000 or so\
 , aiming to make good use of modern processors - specifically the ubiquito
 us x86-64 from Intel and AMD.  With clock speeds approaching a plateau we 
 now need to use multiple cores\, utilize the various levels of cache to re
 duce memory bandwidth demands\, use the vector registers and avoid unpredi
 ctable branches\, but by doing all of these\, speed improvements in excess
  of a factor of 100 are readily obtained over the methods of a couple of d
 ecades ago.<br>This talk will explain some of the changes in technique tha
 t are needed to achieve this - both algorithmic and technological - that s
 eem quite radical at the moment\, but which I expect to become more mainst
 ream in future.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR