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:20230124T171519Z
LOCATION:C140-142
DTSTART;TZID=America/Chicago:20221118T113000
DTEND;TZID=America/Chicago:20221118T115000
UID:submissions.supercomputing.org_SC22_sess444_ws_hipar104@linklings.com
SUMMARY:CAMP: a Synthetic Micro-Benchmark for Assessing Deep Memory Hierar
 chies
DESCRIPTION:Workshop\n\nCAMP: a Synthetic Micro-Benchmark for Assessing De
 ep Memory Hierarchies\n\nBelikov\n\nWe present the open-source CAMP tool f
 or assessing deep memory hierarchies through performance measurements of s
 ynthetic kernels. CAMP provides different access patterns and allows to va
 ry kernels' operational intensities. We describe the tool's design and imp
 lementation, and analyse measurements on a compute node of ARCHER2, the UK
  national supercomputer and compare it to measurements on a compute node o
 n NEXTGenIO. We report results of a strong scaling study of contiguous, st
 rided and stencil access patterns for various operational intensities and 
 explore thread placement options and data sizes. The results confirm that 
 bandwidth saturation can be achieved with a relatively small number of thr
 eads on AMD Rome and that underpopulation may be beneficial as performance
  drops when the node is fully populated for configurations with lower oper
 ational intensity, whilst the effect is less pronounced on the less hierar
 chical Intel Cascade Lake. Finally we discuss sub-NUMA-node awareness and 
 directions for extending CAMP.\n\nSession Format: Recorded\n\nTag: Algorit
 hms, Architectures, Compilers, Computational Science, Exascale Computing, 
 Heterogeneous Systems, Hierarchical Parallelism, Memory Systems, Parallel 
 Programming Languages and Models, Parallel Programming Systems, Resource M
 anagement and Scheduling\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR