BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.coscup.org//coscup-2026//talk//PKYW3X
BEGIN:VTIMEZONE
TZID:CST
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:CST
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-coscup-2026-PKYW3X@pretalx.coscup.org
DTSTART;TZID=CST:20260808T153000
DTEND;TZID=CST:20260808T160000
DESCRIPTION:Running unmodified x86-64 binaries on ARM64 and RISC-V\nLinux s
 ystems remains important\, but general-purpose emulators such as\nQEMU oft
 en impose too much overhead for demanding workloads\, while\nhardware-assi
 sted translators such as Rosetta 2 depend on platform-\nspecific silicon. 
 Box64 addresses this gap with a wrapping-first architec-\nture that routes
  calls into native host libraries through ABI translation\nwrappers and re
 serves JIT compilation for guest code.\nThis paper describes three parts o
 f the design: a native wrapping layer\nspanning 278 libraries\, including 
 entry-point interception and callback\nbridging\; a four-pass JIT compiler
  with flag liveness analysis\, deferred\nflag computation\, and native fla
 g reuse\; and system-level support for\ndeployment\, including Steam conta
 iner bypass\, Wine mixed-bitness co-\nordination\, and TSO emulation on we
 akly ordered hosts. Evaluation on\nRaspberry Pi 5 (ARM64)\, HiFive™ Prem
 ier P550 and Andes AX45MP \n(RISC-V) shows a median of 67% native CPU perf
 ormance on ARM64 \nmicrobenchmarks\, 63–94% native throughput for favora
 ble wrapped \nlibrary calls\, and a median 4× speedup over QEMU\, reachin
 g \n40×when wrapping removes translation from the hot path.
DTSTAMP:20260713T142502Z
LOCATION:TR213
SUMMARY:Box64 in Action: Playing x86-64 Steam Games on ARM64 / RISC-V with 
 Dynamic Binary Translation - Chi-Kuan Chiu
URL:https://pretalx.coscup.org/coscup-2026/talk/PKYW3X/
END:VEVENT
END:VCALENDAR