BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pydata-amsterdam2026//speaker//FMBDC8
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-pydata-amsterdam2026-EKZMWT@pretalx.com
DTSTART;TZID=CET:20260911T141000
DTEND;TZID=CET:20260911T144000
DESCRIPTION:1. Why free-threaded Python matters for native extensions (4 mi
 n)\n\n- What Python 3.14 changes\n- Why existing C/C++ extensions do not b
 enefit automatically\n\n2. Where our embedded engine still hit the GIL (5 
 min)\n\n- Python UDF calls\n- String-heavy columns\n- DataFrame integratio
 n\n\n3. Three changes that unlocked parallelism (12 min)\n\n- Running Pyth
 on UDFs in-process\n- Declaring the module GIL-independent and fixing shar
 ed state\n- Reworking string hot paths\n\n4. What broke and what surprised
  us (4 min)\n\n- Hidden GIL assumptions\n- Stable ABI trade-offs\n- Places
  where no-GIL did not help\n\n5. Benchmarks and a migration checklist (5 m
 in)\n\n- Real workloads on 16 cores\n- A practical checklist for extension
  authors
DTSTAMP:20260710T150432Z
LOCATION:Room 1 (170)
SUMMARY:No GIL\, Real Gains: Porting a C++ Python Extension to Free-Threade
 d Python 3.14 - Auxten Wang
URL:https://pretalx.com/pydata-amsterdam2026/talk/EKZMWT/
END:VEVENT
END:VCALENDAR