skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Can many-task models make data movement and fault-tolerance easier to express?.


Abstract not provided.

Publication Date:
Research Org.:
Sandia National Lab. (SNL-CA), Livermore, CA (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
Report Number(s):
DOE Contract Number:
Resource Type:
Resource Relation:
Conference: Proposed for presentation at the CIS ERB in Albuquerque, NM.
Country of Publication:
United States

Citation Formats

Wilke, Jeremiah J. Can many-task models make data movement and fault-tolerance easier to express?.. United States: N. p., 2015. Web.
Wilke, Jeremiah J. Can many-task models make data movement and fault-tolerance easier to express?.. United States.
Wilke, Jeremiah J. 2015. "Can many-task models make data movement and fault-tolerance easier to express?.". United States. doi:.
title = {Can many-task models make data movement and fault-tolerance easier to express?.},
author = {Wilke, Jeremiah J},
abstractNote = {Abstract not provided.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = 2015,
month = 4

Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • Recent trends in high-performance computing point towards increasingly large machines with millions of processing, storage, and networking elements. Unfortunately, the reliability of these machines is inversely proportional to their size, resulting in a system-wide mean-time-between-failures (MTBF) ranging from a few days to a few hours. As such, for long-running applications, the ability to efficiently recover from frequent failures is essential. Traditional forms of fault tolerance, such as checkpoint/restart, suffer from performance issues related to limited I/O and memory bandwidth. In this paper, we present a fault-tolerance mechanism that reduces the cost of failure recovery by maintaining shadow data structures andmore » performing redundant remote memory accesses. We present results from a computational chemistry application running at scale to show that our techniques provide applications with a high degree of fault tolerance and low (2%--4%) overhead for 2048 processors.« less
  • Hybrid or pure permanent magnet undulators (PMU) are widely used because they have high field quality, allow easy field correction and do not consume power. Their main drawback is the necessity of moving one half of the magnet relative to the other to change field value, which requires a high precision, remotely controlled (and thus costly) driving system On the other hand, electromagnetic undulatory (EMU) have no problem with field variation, but consume too much power (100 - 400 kW) for high fields. Adding permanent magnets to EMU results in a considerable decrease of power consumption, while retaining the advantagemore » of easily changing field level. A model of a CW combined EM+PM plane undulator having a 4.8 cm period and 8 periods long is described. It is simple in design and cheap in manufacturing: magnet yokes are made of soft steel rings in which 1.6 cm air gaps were cut to form pole faces. Odd yokes are placed to one side of the undulator axis and even yokes to the other with the air gaps on the axis. Each set of yokes is excited by its own separate winding of simple racetrack shape. Undulator deflection parameter K = 1.1 (B = 2.4 kG) can be reached at a 0.78kW power level, i.e., less than 100 W per period, while without PM only a maximum K = 0.8 can be obtained and requires 4 kW power. No water cooling is needed, which greatly simplifies undulator design. The undulator was not optimized relative to the axial-air-gap to ring-width ratio: one might expect some increase in field level for thinner rings. Field amplitude depends also on relative transverse position of odd and even pole faces.« less
  • Abstract not provided.
  • Abstract not provided.
  • Abstract not provided.