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:20211207T054758Z
LOCATION:220-221
DTSTART;TZID=America/Chicago:20211116T113000
DTEND;TZID=America/Chicago:20211116T120000
UID:submissions.supercomputing.org_SC21_sess154_pap129@linklings.com
SUMMARY:Scalable FBP Decomposition for Cone-Beam CT Reconstruction
DESCRIPTION:Paper\n\nScalable FBP Decomposition for Cone-Beam CT Reconstru
 ction\n\nChen, Wahib, Wang, Hirofuchi, Ogawa...\n\nFiltered Back-Projectio
 n (FBP) is a fundamental compute-intense algorithm used in tomographic ima
 ge reconstruction. Cone-Beam Computed Tomography (CBCT) devices use a cone
 -shaped X-ray beam, in comparison to the parallel beam used in older CT ge
 nerations. Distributed image reconstruction of cone-beam datasets typicall
 y relies on dividing batches of images into different nodes. This simple i
 nput decomposition, however, introduces limits on input/output sizes and s
 calability.\n\nWe propose a novel decomposition scheme and reconstruction 
 algorithm for distributed FPB. This scheme enables arbitrarily large input
 /output sizes, eliminates the redundancy arising in the end-to-end pipelin
 e and improves the scalability by replacing two communication collectives 
 with only one segmented reduction. Finally, we implement the proposed deco
 mposition scheme in a framework that is useful for all current-generation 
 CT devices (7th gen). In our experiments using up to 1024 GPUs, our framew
 ork can construct 4096^3 volumes, for real-world datasets, in under 16 sec
 onds (including I/O).\n\nTag: Applications, Computational Science\n\nRegis
 tration Category: Tech Program Reg Pass
END:VEVENT
END:VCALENDAR