Intel® Cluster Ready

How to handle mismatched InfiniBand HCA and switch speed with Intel® Cluster Checker v3.0

Intel® Cluster Checker v3.0 can detect the type of the InfiniBand Host Channel Adapter (HCA) that is installed in a cluster node. It can also determine what the expected performance for a given HCA is and use that performance to compare measured performance values with.

Intel® Cluster Checker v3.0 does not have the ability to detect the InfiniBand switch type and can, therefore, not determine what the maximum possible performance of the overall InfiniBand fabric can be.

  • Partner
  • Linux*
  • Server
  • Intel® Cluster Checker
  • Intel Cluster Ready
  • Intel Cluster Checker
  • Intel Cluster Checker v3
  • Intel Cluster Checker v3.0
  • Cluster Checker
  • Intel® Cluster Ready
  • Elaborazione basata su cluster
  • How to achieve an Intel® Cluster Ready Certification with Intel® Cluster Checker v3.0 on clusters with non-identical configuration of head and compute nodes

    This article describes the steps that are necessary to achieve an Intel® Cluster Ready Certification using Intel® Cluster Checker v3.0 on a cluster where the head node and compute nodes are using different hardware or software configurations. The compute nodes use identical hardware and software configurations.
  • Partner
  • Linux*
  • Server
  • Intel® Cluster Checker
  • Intel Cluster Ready
  • Intel Cluster Checker
  • Cluster Checker
  • Intel Cluster Checker v3
  • Intel Cluster Checker v3.0
  • Intel® Cluster Ready
  • Elaborazione basata su cluster
  • How do I get ITAC to show routine names from my code?

    By default ITAC only displays "MPI" and "APPLICATION" in the event display.

     

    • I've set VT_PCTRACE, but that only seems to allow me to pop up a source dialog, and even that does not show a back trace so that's fairly useless.
    • I put VT_initialize/VT_finalize in my code, and create a bunch of state handles which as pass to VT_begin / VT_end.

    None to any avail. What am I missing? What file should contain my routine names?

     

    Victor.

    Trying to make MPI work in a QuickWin application

    Hi,

    This is my first try to enter MPI into a simple QuickWin program under Visual Studio 2010

    I first installed Cluster Studio 2015 update 4 on my i7 4910MQ Dell laptop as an x64 project. The simple code compiled, linked and run directly.

    Then I just added INCLUDE *mpif.h' on a new line after the IMPLICIT NONE line and got the compiler message error #5102 cannot open include file

    'mpif.h'.

    I then tried a number of ways to inform the compiler about the location of the mpif.h file but instead got a number of warnings and errors. 

    Parallel Universe link slightly mangled

    I don't know exactly where to submit this, but one of the links on the Parallel Universe magazine page is mangled.  On the page https://software.intel.com/en-us/intel-parallel-universe-magazine the displayed Issue 19 actually points to issue 20.  Maybe someone at Intel would like to fix this.  (Of course you can still get issue 19 if you really want to, but there is an annoyance factor.)

    How to achieve an Intel Cluster Ready Certification with Intel® Cluster Checker v3.0 on a Cluster using the Intel® True Scale Fabric

    This article describes the steps that are necessary to achieve an Intel® Cluster Ready Certification using Intel® Cluster Checker v3.0 on a cluster using the Intel® True Scale fabric. In order to comply with the requirements of the Intel® Cluster Ready architecture specification 1.3.1, certain aspects of the cluster configuration have to be checked. Those requirements may not necessarily be checked for when checking for cluster health only. For clusters based on the Intel® True Scale fabric this requires testing the DAPL provider interface in addition to the native TMI message fabric interface and the TCP based fabric.
  • Partner
  • Linux*
  • Server
  • Intel® Cluster Checker
  • Intel Cluster Ready
  • Intel Cluster Checker
  • Cluster Checker
  • Intel Cluster Checker v3
  • Intel® Cluster Ready
  • Elaborazione basata su cluster
  • Problem with Intel Trace Collector

    I'm trying to use for the first time the intel trace collector on a cluster machine (with the intel xe 2013 and the itac 8.1.2.033).

    I built my program in the standard production mode and in the bash script submitted to the PBS scheduler there were the following commands:

    #PBS -l select=14:ncpus=16:mem=120gb:mpiprocs=16

    module load intel/cs-xe-2013

    source [path]/cs-xe-2013/none/itac/8.1.2.033/bin/itacvars.sh

    mpirun -trace [path]/my_program [arguments]

    Deadlock with MPI_Win_fence going from Intel MPI 4.1.3.049 to 5.0.3.048

    We encountered a problem when migrating a code from Intel MPI 4.1.3.049 to 5.0.3.048. The code in question is a complex simulation that first reads global input state from disk into several parts in memory and then accesses this memory in a hard to predict fashion to create a new decomposition. We use active target RMA for this (on machines which support this like BG/Q we also use passive target) since a rank might need data from the part that is at another rank to form its halo.

    Iscriversi a Intel® Cluster Ready