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:20190719T085744Z
LOCATION:HG F 3
DTSTART;TZID=Europe/Stockholm:20190614T090000
DTEND;TZID=Europe/Stockholm:20190614T093000
UID:submissions.pasc-conference.org_PASC19_sess181_pap_jan121@linklings.co
m
SUMMARY:Prediction of Time-to-Solution in Material Science Simulations usi
ng Deep Learning
DESCRIPTION:Paper\nComputer Science and Applied Mathematics, Chemistry and
Materials\n\nPrediction of Time-to-Solution in Material Science Simulatio
ns using Deep Learning\n\nPittino, Bonfa', Bartolini, Affinito, Benini...\
n\nPredicting the time to solution for massively parallel scientific codes
is a complex task. The reason for this is the presence of multiple, stron
gly interconnected algorithms that possibly react differently to the chang
es in compute power, vectorization length, memory and network bandwidth an
d latency and I/O throughput.A reliable prediction of execution time is ho
wever of great importance to the user who wants to plan on large scale sim
ulations or virtual screening procedures characteristic of high throughput
computing.In this article we present a practical approach based on machin
e learning techniques to achieve very accurate predictions of the time to
solution for a DFT-based material science code.We compare our results with
the predictions provided by a parametrized analytical performance model s
howing that deep learning solutions allow for a greater accuracy without t
he need of domain knowledge to introduce an explicit description of the al
gorithms implemented in the code.
Full paper: https://doi.org/1
0.1145/3324989.3325720
END:VEVENT
END:VCALENDAR