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:20260202T201806Z
LOCATION:240
DTSTART;TZID=America/Chicago:20251116T113000
DTEND;TZID=America/Chicago:20251116T115500
UID:submissions.supercomputing.org_SC25_sess119_ws_xloop101@linklings.com
SUMMARY:Adapting scientific streaming inference workflows for a determinis
 tic tensor processing unit
DESCRIPTION:Samantha Fowler, Kazutomo Yoshii, Antonino Miceli, Senthil Gna
 nasekaran, Tao Zhou, and Nicholas Contini (Argonne National Laboratory (AN
 L))\n\nThe realization of real-time data processing near X-ray detectors p
 resents ongoing challenges due to long ASIC development cycles and the lim
 ited computational capacity of near-detector FPGAs. We propose a hybrid so
 lution that streams data directly to a deterministic tensor processing uni
 t (i.e., Groq AI accelerator), enabling low-latency, high-throughput infer
 ence. This paper describes the system architecture, supporting software st
 ack, and performance projections, demonstrating the advantages of this hyb
 rid platform for future X-ray imaging systems. This integration shows prom
 ise for advancing real-time edge computing and enabling intelligent contro
 l in photon science experiments. A single inference on a 128 × 128 image, 
 including image transfer time, completes in 156.06 𝜇s, enabling approximat
 ely 6.4 kHz processing with the edgePtychoNN model and improving experimen
 tal-in-the-loop computing. Using this system, we achieve a 3.6× speedup ov
 er previous systems, highlighting the potential of this approach\n\nRecord
 ing: Livestreamed, Recorded\n\nRegistration Category: Technical Program Re
 g Pass, Workshop Reg Pass\n\nSession Chairs: Justin Wozniak (Argonne Natio
 nal Laboratory (ANL), University of Chicago); Nicholas Schwarz (Argonne Na
 tional Laboratory (ANL)); and Hannah Parraga (Argonne National Laboratory 
 (ANL))\n\n
END:VEVENT
END:VCALENDAR