MPI: Difference between revisions
No edit summary |
No edit summary |
||
Line 135: | Line 135: | ||
==MPI Libraries for parallel jobs on the teaching cluster== | ==MPI Libraries for parallel jobs on the teaching cluster== | ||
All compute nodes on | All compute nodes on Sapelo2 have Infiniband (IB) interconnect via '''EDR''' Infiniband network (100Gbps). Various IB-enabled MPI libraries are available and users can set the environment variables for the MPI library of choice by loading the corresponding module file. | ||
For more information on Environment Modules, please see the [[Lmod]] page. | |||
The following MPI libraries are available: | |||
<!-- | <!-- | ||
====MVAPICH2==== | ====MVAPICH2==== |
Revision as of 11:17, 6 September 2023
MPI Libraries for parallel jobs on Sapelo2
All compute nodes on Sapelo2 have Infiniband (IB) interconnect via EDR Infiniband network (100Gbps). Various IB-enabled MPI libraries are available and users can set the environment variables for the MPI library of choice by loading the corresponding module file.
For more information on Environment Modules, please see the Lmod page.
The following MPI libraries are available:
OpenMPI
You can find all OpenMPI modules available on Sapelo2 by running the following command on a Sapelo2 node:
module spider OpenMPI
The module names have the format OpenMPI/Version-CompilerToolchain-ToolchainVersion.
For example, these are some of the modules available:
[shtsai@ss-sub1 ~]$ module spider OpenMPI ----------------------------------------------------------------------------------------------------------------------------------------------- OpenMPI: ----------------------------------------------------------------------------------------------------------------------------------------------- Description: The Open MPI Project is an open source MPI-3 implementation. Versions: OpenMPI/3.1.4-GCC-8.3.0 OpenMPI/4.1.1-GCC-11.2.0 OpenMPI/4.1.2-GCC-11.2.0 OpenMPI/4.1.4-GCC-11.3.0 OpenMPI/4.1.4-GCC-12.2.0 ------------------------------------------------------------------------------------------------------------------------------------------------ For detailed information about a specific "OpenMPI" package (including how to load the modules) use the module's full name. Note that names that have a trailing (E) are extensions provided by other modules. For example: $ module spider OpenMPI/4.1.4-GCC-12.2.0 -----------------------------------------------------------------------------------------------------------------------------------------------
Once the appropriate module is loaded, you can compile code with mpicc, mpiCC, mpicxx, mpic++, mpifort, mpif90, mpif77, etc. and you can run applications that were linked to the MPI libraries loaded by the module.
Intel MPI
You can find all Intel MPI modules available on Sapelo2 by running the following command on a Sapelo2 node:
module spider impi
The module names have the format impi/Version-CompilerToolchain-ToolchainVersion. For example, these are some of the modules available:
[shtsai@ss-sub1 ~]$ module spider impi
------------------------------------------------------------------------------------------------------------------------------------
impi:------------------------------------------------------------------------------------------------------------------------------------
Description: Intel MPI Library, compatible with MPICH ABIVersions: impi/2021.4.0-intel-compilers-2021.4.0 impi/2021.6.0-intel-compilers-2022.1.0 impi/2021.9.0-intel-compilers-2023.1.0 Other possible modules matches: iimpi------------------------------------------------------------------------------------------------------------------------------------
To find other possible module matches execute:$ module -r spider '.*impi.*'------------------------------------------------------------------------------------------------------------------------------------
For detailed information about a specific "impi" package (including how to load the modules) use the module's full name. Note that names that have a trailing (E) are extensions provided by other modules. For example:$ module spider impi/2021.9.0-intel-compilers-2023.1.0------------------------------------------------------------------------------------------------------------------------------------
Once the appropriate module is loaded, you can compile code with mpicc, mpiicc, mpicxx, mpiicpc, mpiifort, mpif90, mpif77, etc. and you can run applications that were linked to the MPI libraries loaded by the module.
MPI Libraries for parallel jobs on the teaching cluster
All compute nodes on Sapelo2 have Infiniband (IB) interconnect via EDR Infiniband network (100Gbps). Various IB-enabled MPI libraries are available and users can set the environment variables for the MPI library of choice by loading the corresponding module file.
For more information on Environment Modules, please see the Lmod page.
The following MPI libraries are available:
OpenMPI
You can find all OpenMPI modules available on the teaching cluster by running the following command:
module spider OpenMPI
The module names have the format OpenMPI/Version-CompilerToolchain-ToolchainVersion.
For example, these are some of the modules available:
zhuofei@teach-sub1 ~$ ml spider OpenMPI ------------------------------------------------------------------------------------------------------------------------------------ OpenMPI: ------------------------------------------------------------------------------------------------------------------------------------ Description: The Open MPI Project is an open source MPI-3 implementation. Versions: OpenMPI/3.1.4-GCC-8.3.0 OpenMPI/4.1.1-GCC-11.2.0 OpenMPI/4.1.2-GCC-11.2.0 OpenMPI/4.1.4-GCC-11.3.0 OpenMPI/4.1.4-GCC-12.2.0 -------------------------------------------------------------------------------------------------------------------------------------- For detailed information about a specific "OpenMPI" package (including how to load the modules) use the module's full name. Note that names that have a trailing (E) are extensions provided by other modules. For example: $ module spider OpenMPI/4.1.4-GCC-12.2.0 ------------------------------------------------------------------------------------------------------------------------------------
Once the appropriate module is loaded, you can compile code with mpicc, mpiCC, mpicxx, mpic++, mpifort, mpif90, mpif77, etc. and you can run applications that were linked to the MPI libraries loaded by the module.
Intel MPI
You can find all Intel MPI modules available on the teaching cluster by running the following command:
module spider impi
The module names have the format impi/Version-CompilerToolchain-ToolchainVersion. For example, these are some of the modules available:
zhuofei@teach-sub1 ~$ ml spider impi ------------------------------------------------------------------------------------------------------------------------------------ impi: impi/2018.5.288-iccifort-2019.5.281 ------------------------------------------------------------------------------------------------------------------------------------ Description: Intel MPI Library, compatible with MPICH ABI Versions: impi/2021.4.0-intel-compilers-2021.4.0 impi/2021.6.0-intel-compilers-2022.1.0 impi/2021.9.0-intel-compilers-2023.1.0 Other possible modules matches: iimpi ------------------------------------------------------------------------------------------------------------------------------------ To find other possible module matches execute: $ module -r spider '.*impi.*' ------------------------------------------------------------------------------------------------------------------------------------ For detailed information about a specific "impi" package (including how to load the modules) use the module's full name. Note that names that have a trailing (E) are extensions provided by other modules. For example: $ module spider impi/2021.9.0-intel-compilers-2023.1.0 ------------------------------------------------------------------------------------------------------------------------------------
Once the appropriate module is loaded, you can compile code with mpicc, mpiicc, mpicxx, mpiicpc, mpiifort, mpif90, mpif77, etc. and you can run applications that were linked to the MPI libraries loaded by the module.