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:275
DTSTART;TZID=America/Chicago:20251119T143700
DTEND;TZID=America/Chicago:20251119T150000
UID:submissions.supercomputing.org_SC25_sess283_pap449@linklings.com
SUMMARY:MANS: Efficient and Portable ANS Encoding for Multi-Byte Integer D
 ata on CPUs and GPUs
DESCRIPTION:Wenjing Huang and Jinwu Yang (Institute of Computing Technolog
 y, Chinese Academy of Sciences; University of Chinese Academy of Sciences,
  Beijing); Shengquan Yin (Institute of Computing Technology, Chinese Acade
 my of Sciences; University of Science and Technology of China); Haoxu Li a
 nd Yida Gu (Institute of Computing Technology, Chinese Academy of Sciences
 ; University of Chinese Academy of Sciences, Beijing); Zedong Liu (Institu
 te of Computing Technology, Chinese Academy of Sciences; University of Ele
 ctronic Science and Technology of China); Xing Jing and Zheng Wei (Institu
 te of Computing Technology, Chinese Academy of Sciences); Shiyuan Fu and H
 ao Hu (Institute of High Energy Physics, Chinese Academy of Sciences); and
  Guangming Tan and Dingwen Tao (Institute of Computing Technology, Chinese
  Academy of Sciences)\n\nLossless compression is a classic technique for r
 educing data storage and transmission requirements. Asymmetric numeral sys
 tems (ANS) is a high-throughput, high-ratio lossless compression algorithm
 , but it lacks effective support for multi-byte data and cross-platform co
 mpatibility.\n\nTo address this issue, we propose an adaptive data mapping
  (ADM) scheme, which maps multi-byte integer data into single-byte space b
 ased on the data's characteristics, improving the compression ratio of ANS
  while maintaining low encoding redundancy. We also optimize the ADM algor
 ithm and the ANS encoder for GPU and CPU architectures, respectively, and 
 combine them to create an efficient and portable ANS encoding method for m
 ulti-byte integer data, called MANS.\n\nExperimental results show that MAN
 S improves compression ratios by an average of 1.24$\times$, achieves 870.
 27MB/s throughput on CPUs, and delivers up to 288.45$\times$ and 135.86$\t
 imes$ speedups on an NVIDIA A100 and an AMD MI210 GPU compared to the CPU 
 version—demonstrating its efficiency and portability.\n\nTag: Data Analyti
 cs, Visualization & Storage\n\nRecording: Livestreamed, Recorded\n\nRegist
 ration Category: Technical Program Reg Pass\n\nSession Chair: Ana Kupresan
 ian (Lawrence Berkeley National Laboratory (LBNL))\n\n
END:VEVENT
END:VCALENDAR