BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20220812T074335Z
LOCATION:Singapore Room
DTSTART;TZID=Europe/Stockholm:20220628T150000
DTEND;TZID=Europe/Stockholm:20220628T153000
UID:submissions.pasc-conference.org_PASC22_sess177_pap111@linklings.com
SUMMARY:Scalable Low-Rank Factorization Using a Task-Based Runtime System 
 with Distributed Memory
DESCRIPTION:Paper\n\nScalable Low-Rank Factorization Using a Task-Based Ru
 ntime System with Distributed Memory\n\nGnanasekaran, Darve\n\nWe present 
 a new parallel task-based algorithm for randomized low-rank factorizations
  of a matrix and its application to fast hierarchical solvers. TaskTorrent
  is a lightweight, distributed, task-based runtime in C++. We explain how 
 the randomized factorization can be expressed as a task graph in TaskTorre
 nt. We compare the key steps in the factorization with similar implementat
 ions using StarPU, PaRSEC, and in ScaLAPACK. The overall randomized algori
 thm in TaskTorrent outperforms the column pivoted QR routine from ScaLAPAC
 K by far. The weak and strong scaling experiments demonstrate excellent sc
 alability up to our largest available configuration with 1,024 cores. Fina
 lly, we incorporate our distributed low-rank factorization into spaND, a f
 ast hierarchical solver for solving large sparse linear systems. We demons
 trate gains in performance and scalability of the spaND solver.\n\nDomain:
  Computer Science and Applied Mathematics
END:VEVENT
END:VCALENDAR