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:20230124T170804Z
LOCATION:C1-2-3
DTSTART;TZID=America/Chicago:20221116T083000
DTEND;TZID=America/Chicago:20221116T170000
UID:submissions.supercomputing.org_SC22_sess249_drs104@linklings.com
SUMMARY:A Data-Centric Optimization Workflow for the Python Language
DESCRIPTION:Doctoral Showcase, Posters\n\nA Data-Centric Optimization Work
 flow for the Python Language\n\nZiogas\n\nPython's extensive software ecos
 ystem leads to high productivity, rendering it the language of choice for 
 scientific computing. However, executing Python code is often slow or impo
 ssible in emerging architectures and accelerators. To complement Python's 
 productivity with the performance and portability required in high-perform
 ance computing (HPC), we introduce a workflow based on data-centric (DaCe)
  parallel programming. Python code with HPC-oriented extensions is parsed 
 into a dataflow-based intermediate representation, facilitating analysis o
 f the program's data movement. The representation is optimized via graph t
 ransformations driven by the users, performance models, and automatic heur
 istics. Subsequently, hardware-specific code is generated for supported ar
 chitectures, including CPU, GPU, and FPGA. We evaluate the above workflow 
 through three case studies. First, to compare our work to other Python-acc
 elerating solutions, we introduce NPBench, a collection of over 50 Python 
 microbenchmarks across a wide range of scientific domains. We show perform
 ance results and scaling across CPU, GPU, FPGA, and the Piz Daint supercom
 puter. DaCe runs 10x faster than the reference Python execution and achiev
 es 2.47x and 3.75x speedups over previous-best solutions and up to 93.16% 
 scaling efficiency. Second, we re-implement in Python and optimize the Qua
 ntum Transport Simulator OMEN. The application's DaCe version executes one
  to two orders of magnitude faster than the original code written in C++, 
 achieving 42.55% of the Summit supercomputer's peak performance. Last, we 
 utilize our workflow to build Deinsum, an automated framework for distribu
 ted multilinear algebra computations expressed in Einstein notation. Deins
 um performs up to 19x faster over state-of-the-art solutions on the Piz Da
 int supercomputer.\n\nRegistration Category: Tech Program Reg Pass, Exhibi
 ts Reg Pass
END:VEVENT
END:VCALENDAR