Configuring Intel® Xeon Phi™ coprocessors inside a cluster

Power Configuration Part 0: Introduction: Yikes, there is a lot that is not documented

I was hoping to write a brief two part overview of how to configure the various power settings for the Intel® Xeon Phi™ coprocessor. It was going to be concise and brief, allowing me to get on to the next topic. Unfortunately, as I dug into the topic further, I discovered that much of it is not very well documented. I found myself essentially writing quite a bit of explanation.

As usual, I am starting off writing this as a series of blogs. At a later point, I will reformat the blogs into a more formal article with any semblance of humor removed.

Symmetric Mode MPI Performance without InfiniBand*

Symptom

Slow host-coprocessor MPI communications in systems with no InfiniBand* HCA.  If running with I_MPI_DEBUG=2 or higher, you will see one of the following messages indicating that the TCP fabric has been selected:

[0] MPI startup(): tcp data transfer mode
[0] MPI startup(): shm and tcp data transfer modes
  • Linux*
  • 服务器
  • 中级
  • Intel® MPI Library
  • tips and tricks
  • Configuring Intel® Xeon Phi™ coprocessors inside a cluster
  • Intel® MPSS
  • ofed
  • 消息传递接口
  • Intel® Many Integrated Core Architecture
  • Configuring Intel® Xeon Phi™ coprocessors inside a cluster

    Author: Michael Hebenstreit
    Contributions: Romain Dolbeau, Jeremy C. Siadal
    Version: 0.81, 20130110

    Abstract

    This paper is intended to provide readers a blueprint of how to set up and configure a cluster with systems containing the Intel® Xeon Phi™ Coprocessor, based on how Intel configured its own Endeavor cluster.  Along the way, specific information about how to compile tools, configure filesystems, and setting up network interfaces is shared in great detail to help understand how this can be done en masse.    

  • 开发人员
  • 服务器
  • Configuring Intel® Xeon Phi™ coprocessors inside a cluster
  • MIC
  • Intel® Xeon Phi™ Coprocessor
  • Xeon Phi
  • 订阅 Configuring Intel® Xeon Phi™ coprocessors inside a cluster