diff --git a/docs/installing-open-mpi/configure-cli-options/networking.rst b/docs/installing-open-mpi/configure-cli-options/networking.rst index dc6bffb7c5a..cabb723a482 100644 --- a/docs/installing-open-mpi/configure-cli-options/networking.rst +++ b/docs/installing-open-mpi/configure-cli-options/networking.rst @@ -73,7 +73,7 @@ can be used with ``configure``: only necessary if the PSM headers and libraries are not in default compiler/linker search paths. - PSM is the support library for QLogic InfiniPath and Intel TrueScale + PSM is the support library for QLogic InfiniPath and Intel True Scale network adapters. * ``--with-psm-libdir=DIR``: diff --git a/docs/tuning-apps/networking/cuda.rst b/docs/tuning-apps/networking/cuda.rst index 6f925085b58..1b75e6655dd 100644 --- a/docs/tuning-apps/networking/cuda.rst +++ b/docs/tuning-apps/networking/cuda.rst @@ -160,7 +160,7 @@ PSM2 support for CUDA --------------------- CUDA-aware support is present in PSM2 MTL. When running CUDA-aware -Open MPI on Intel Omni-path, the PSM2 MTL will automatically set +Open MPI on Cornelis Networks Omni-Path, the PSM2 MTL will automatically set ``PSM2_CUDA`` environment variable which enables PSM2 to handle GPU buffers. If the user wants to use host buffers with a CUDA-aware Open MPI, it is recommended to set ``PSM2_CUDA`` to ``0`` in the execution @@ -168,13 +168,13 @@ environment. PSM2 also has support for the NVIDIA GPUDirect support feature. To enable this, users will need to set ``PSM2_GPUDIRECT`` to ``1`` in the execution environment. -Note: The PSM2 library and ``hfi1`` driver with CUDA support are -requirements to use GPUDirect support on Intel Omni-Path. The minimum +Note: The PSM2 library and ``hfi1`` driver with CUDA support are requirements +to use GPUDirect support on Cornelis Networks Omni-Path. The minimum PSM2 build version required is `PSM2 10.2.175 `_. -For more information refer to the `Intel Omni-Path documentation -`_. +For more information refer to the `Cornelis Networks Customer Center +`_. ///////////////////////////////////////////////////////////////////////// @@ -629,10 +629,9 @@ limit is reached: What are some guidelines for using CUDA and Open MPI with Omni-Path? -------------------------------------------------------------------- -When developing CUDA-aware Open MPI applications for OPA-based -fabrics, the PSM2 transport is preferred and a CUDA-aware version of -PSM2 is provided with all versions of the Intel Omni-Path IFS software -suite. +When developing CUDA-aware Open MPI applications for OPA-based fabrics, the +PSM2 transport is preferred and a CUDA-aware version of PSM2 is provided with +all versions of the Cornelis Networks Omni-Path OPXS software suite. .. error:: TODO Are Intel/OPA references still correct? @@ -655,10 +654,10 @@ processes do not move between NUMA nodes. See the section on :ref:`NUMA Node Issues `, for more information. -For more information see the *Intel Performance Scaled Messaging 2 -(PSM2) Programmer's Guide* and the *Intel Omni-Path Performance Tuning -Guide*, which can be found on the `Intel Omni-Path web site -`_. +For more information see the *Cornelis Networks Performance Scaled Messaging 2 +(PSM2) Programmer's Guide* and the *Cornelis Networks Omni-Path Performance +Tuning Guide*, which can be found in the `Cornelis Networks Customer Center +`_. .. error:: TODO Are Intel/OPA references still correct? diff --git a/docs/tuning-apps/networking/ofi.rst b/docs/tuning-apps/networking/ofi.rst index fb38d29047d..3f7563985d1 100644 --- a/docs/tuning-apps/networking/ofi.rst +++ b/docs/tuning-apps/networking/ofi.rst @@ -131,6 +131,5 @@ Users can restrict access to a single HFI using the environment variable: More details can be found on the PSM2 Programmer's Guide and the Omni-Path Fabric Performance Tuning Guide. -Please also see `the full Omni-Path documentation -`_ +Please see the `Cornelis Networks Customer Center `_ for more details.