BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//cfp.scipy.org//2023//speaker//VQNJX7
BEGIN:VTIMEZONE
TZID:America/Chicago
BEGIN:DAYLIGHT
DTSTART:20220712T000000
TZNAME:CDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0500
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20221106T020000
RDATE:20231105T020000
TZNAME:CST
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20230312T030000
RDATE:20240310T030000
TZNAME:CDT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Can There Be Too Much Parallelism? - Thomas J. Fan
DTSTART;TZID=America/Chicago:20230712T112500
DTEND;TZID=America/Chicago:20230712T115500
DTSTAMP:20261007T025431Z
UID:pretalx-2023-VUFGS8@cfp.scipy.org
DESCRIPTION:Numerical Python libraries can run computations on many CPU co
 res with various parallel interfaces. When we simultaneously use multiple 
 levels of parallelism\, it may result in oversubscription and degraded per
 formance. This talk explores the programming interfaces used to control pa
 rallelism exposed by libraries such as NumPy\, SciPy\, and scikit-learn. W
 e will learn about parallel primitives used in these libraries\, such as O
 penMP and Python's multiprocessing module. We will see how to control para
 llelism in these libraries to avoid oversubscription. Finally\, we will lo
 ok at the overall landscape for configuring parallelism and highlight path
 s for improving the user experience.
LOCATION:Amphitheater 204
URL:https://cfp.scipy.org/2023/talk/VUFGS8/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Python Array API Standard: Toward Array Interoperability in the Sc
 ientific Python Ecosystem - Ralf Gommers\, Stephannie Jimenez Gacha\, Leo 
 Fang\, Saul Shanabrook\, Travis Oliphant\, Matthew Barber\, Aaron Meurer\,
  Thomas J. Fan\, John Kirkham\, Stephan Hoyer\, Tyler Reddy\, Andreas Muel
 ler\, Athan Reines\, Mario\, Alexandre Passos
DTSTART;TZID=America/Chicago:20230714T112500
DTEND;TZID=America/Chicago:20230714T115500
DTSTAMP:20261007T025431Z
UID:pretalx-2023-T7DTX8@cfp.scipy.org
DESCRIPTION:The array API standard (https://data-apis.org/array-api/) is a
  common specification for Python array libraries\, such as NumPy\, PyTorch
 \, CuPy\, Dask\, and JAX. \n\nThis standard will make it straightforward f
 or array-consuming libraries\, like scikit-learn and SciPy\, to write code
  that uniformly supports all of these libraries. This will allow\, for ins
 tance\, running the same code on the CPU and GPU.\n\nThis talk will cover 
 the scope of the array API standard\, supporting tooling which includes a 
 library-independent test suite and compatibility layer\, what work has bee
 n completed so far\, and the plans going forward.
LOCATION:Amphitheater 204
URL:https://cfp.scipy.org/2023/talk/T7DTX8/
END:VEVENT
END:VCALENDAR