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:20260202T201804Z
LOCATION:275
DTSTART;TZID=America/Chicago:20251119T135200
DTEND;TZID=America/Chicago:20251119T141500
UID:submissions.supercomputing.org_SC25_sess283_pap664@linklings.com
SUMMARY:Phoenix: A Refactored I/O Stack for GPU Direct Storage Without Pho
 ny Buffers
DESCRIPTION:Jianqin Yan, Shi Qiu, Yina Lv, Yifan Hu, Hao Chen, and Zhirong
  Shen (Xiamen University); Xin Yao and Renhai Chen (Huawei Theory Lab); Ji
 wu Shu (Xiamen University); Gong Zhang (Huawei Theory Lab); and Yiming Zha
 ng (Shanghai Jiao Tong University, Xiamen University)\n\nGPU Direct Storag
 e (GDS) plays a vital role in GPU storage systems, utilizing P2P-DMA techn
 ology to establish a direct data transfer path between the GPU and storage
  devices. This direct path reduces storage access latency and CPU overhead
 , thus improving data transfer efficiency. Currently, however, GDS employs
  a phony buffer in host memory to interact with the Linux kernel, resultin
 g in suboptimal performance, additional resource consumption, and deployme
 nt complexity.\n\nIn this paper, we propose Phoenix, a refactored GDS soft
 ware stack without phony buffers. Phoenix employs the memory mapping servi
 ce of ZONE_DEVICE to map GPU memory into the page table at system startup.
  The kernel module of Phoenix stores the returned address information, all
 ocates user-space virtual memory, and establishes a mapping with the desig
 nated GPU memory. Extensive evaluation shows that, compared to the existin
 g GDS software stack, Phoenix reduces software overhead along the critical
  I/O path and improves end-to-end performance.\n\nTag: Data Analytics, Vis
 ualization & Storage\n\nRecording: Livestreamed, Recorded\n\nRegistration 
 Category: Technical Program Reg Pass\n\nSession Chair: Ana Kupresanian (La
 wrence Berkeley National Laboratory (LBNL))\n\n
END:VEVENT
END:VCALENDAR