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:20220812T074334Z
LOCATION:Foyer 2nd Floor
DTSTART;TZID=Europe/Stockholm:20220628T090000
DTEND;TZID=Europe/Stockholm:20220628T110000
UID:submissions.pasc-conference.org_PASC22_sess181_pos147@linklings.com
SUMMARY:P29 - Benchmarking Memory-Bound Computational Physics Codes with I
 n-House Developed Cloud-Bursting Solution
DESCRIPTION:Poster\n\nP29 - Benchmarking Memory-Bound Computational Physic
 s Codes with In-House Developed Cloud-Bursting Solution\n\nLlopis Sanmilla
 n, Álvarez, Józefiak, Miroslaw\n\nAt CERN, the Theoretical Physics departm
 ent (TH) and the Accelerator Technology Sector (ATS) rely heavily on HPC u
 sage for developing next-generation LHC technology. Due to the varying com
 putational needs of each department, we observe a bursty compute demand pa
 ttern that depends on our HPC project deadlines and conference schedules. 
 In order to tackle compute demand periods that exceed our on-premise compu
 te capacity, we have implemented a cloud-bursting solution leveraging Micr
 osoft Azure. This cloud-bursting solution is integrated into our Slurm-bas
 ed on-premise infrastructure to allow users to submit their jobs to the cl
 oud without any major modifications to their MPI programs or job submit sc
 ripts. It provides the elasticity to grow our HPC resources in the cloud w
 hen more compute capacity is needed, and scale down these cloud-associated
  costs when the on-premise capacity is sufficient to satisfy demand.In add
 ition, we have carried out a cost-effectiveness evaluation of different VM
  sizes, leveraging common workloads from our ATS and TH use cases. More pr
 ecisely, we evaluate different Azure HPC VM sizes equipped with AMD EPYC 7
 551 (Naples), 7742 (Rome), 7003 (Milan), targeting memory-bound codes. For
  each of these we run the following workloads: FDS (Fire Dynamics Simulato
 r), Ansys Fluent, and OpenQCD (open Quantum Chromodynamics).
END:VEVENT
END:VCALENDAR