BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260522T150124Z
LOCATION:D173
DTSTART;TZID=America/Chicago:20181112T140000
DTEND;TZID=America/Chicago:20181112T142000
UID:submissions.supercomputing.org_SC18_sess175_pec423@linklings.com
SUMMARY:HPC PowerStack: a community-wide open collaboration for enabling s
 ystem-wide power efficiency
DESCRIPTION:Siddhartha Jana (Intel Corporation)\n\nOne of the challenges t
 o Open Supercomputing in the exascale era is the design of software soluti
 ons that drive Power Management across the system stack. In order to tackl
 e the power challenges, the HPC community has designed various open source
  and vendor-specific solutions to meet power management goals at various l
 evels of granularity - from system level to node-level. These components a
 re built under certain assumptions of the system behavior and remain compl
 etely agnostic of each other’s techniques. For example, an open source wor
 kload manager attempting to reduce the system-level power consumption may 
 remain completely oblivious of the power demands by the application-level 
 runtime and node-level hardware components. Attempting to integrate them i
 nto a unified system stack often lead to them overstepping on each other’s
  solutions which, in turn leads to unreliable system performance. To avoid
  this, system integrators end up designing their stack with tightly-couple
 d cherry-picked vendor-specific solutions. This lack of coordination betwe
 en the vendors and the open solution community leads to underutilization o
 f system Watts and FLOPS. As a result, there is an urgent need for the HPC
  community to (A) identify the key software actors needed in a system powe
 r stack: job-schedulers, application-level runtime, hardware knobs (B) arr
 ive at a consensus on the roles and responsibilities of these actors, (C) 
 design communication protocols for bidirectional control and feedback sign
 als among the actors to enable scalable closed-loop coordination at differ
 ent granularities, and (E) study and combine existing standalone engineeri
 ng and development prototypes and build a community that actively particip
 ates in open development and engineering efforts.\n \nThis realization led
  to the formation of the PowerStack Community, in 2016. The participants o
 f this community include global-wide members from academia, government lab
 s, and vendors working on different layers of the system software stack. T
 his talk is intended to spread awareness among the workshop attendees and 
 solicit participation. The hope is to facilitate integrating open power ma
 nagement solutions into future PowerStack-compatible systems.\n\nTag: Arch
 itectures, Collaborative Environments, Parallel Programming Languages, Lib
 raries, and Models, Simulation, Workflows\n\nRegistration Category: Worksh
 op Reg Pass\n\nSession Chairs: David Donofrio (Tactical Computing Laborato
 ries LLC, Lawrence Berkeley National Laboratory (LBNL)); Farzad Fatollahi-
 Fard (Lawrence Berkeley National Laboratory); and John Leidel (Tactical Co
 mputing Laboratories LLC, Texas Tech University)\n\n
END:VEVENT
END:VCALENDAR