The Intel® Trace Analyzer and Collector for Windows* is a low-overhead scalable event-tracing library with graphical analysis that reduces the time it takes an application developer to enable maximum performance of cluster applications. This package is for users who develop on and build for Intel® 64 architectures on Windows*. You must have a valid license to download, install and use this product.
The Intel® MPI Library is a high-performance interconnect-independent multi-fabric library implementation of the industry-standard Message Passing Interface, v3.0 (MPI-3.0) specification. This package is for MPI users who develop on and build for Intel® 64 architectures on Windows*. You must have a valid license to download, install, and use this product.
What is the best DAPL provider to used with IntelMPI 5.1 when running with a 10Gb RDMA chipsset from Mellanox?
Currently I'm using I_MPI_DAPL_PROVIDER=ofa-v2-scm-roe-mlx4_0-1
I added some others extra parameters to workaround the resource limitation that we have on the chip
I am finding that MPI_Abort is not killing processes on remote nodes when I_MPI_FABRICS=shm:tmi. The attached PBS job works correctly when the default fabric (shm:dapl) is used, but fails to abort cleanly when shm:tmi is used. Any help with overcoming this problem would be much appreciated.
I would like to manually start my application (paraview's pvserver) on each node of a windows cluster without using mpiexec. The reason is that the application needs to make visible windows so that it can drive a CAVE display system. mpiexec will not work for me, because it uses smpd, which is a service, and since Vista windows services can not interact with the user. (Alternatively, is there a way to run smpd.exe as a startup application instead of a service?)