BwUniCluster 2.0 Maintenance/2020-10/Software Issues - Revision history

M Janczyk: Blanked the page

2023-08-15T08:17:14Z

Blanked the page

← Older revision		Revision as of 10:17, 15 August 2023
Line 1:		Line 1:
	After the last regular [[BwUniCluster_2.0_Maintenance/2020-10\|maintenance]] interval (from 06.10.2020 to 13.10.2020) the following issues with Intel MPI exist:

	* Intel MPI 2018 is incompatible with Red Hat 8.2. Any invocation, even a simple "Hello World" MPI program, will result in a crash. The ''mpi/impi/2018'' module has therefore been removed.

	* There is a bug in Intel MPI 2019.x which leads to crashes when multiple MPI applications which are linked against Intel MPI 2019.x are run on the same node (e.g. in the "single" partition). The first application will run normally, but all others will crash. This can be fixed by setting the environment variable '''I_MPI_HYDRA_TOPOLIB="ipl"'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.

	* There is a bug in Intel MPI 2019.x which leads to incorrect CPU binding/affinity in conjunction with the Slurm batch system used on the clusters. All MPI ranks will run on the same CPU core instead of being bound to all available CPU cores. This can be fixed by setting the environment variable '''I_MPI_HYDRA_BOOTSTRAP_EXEC_EXTRA_ARGS="--cpu-bind=none"'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.

	There is a number of Third-Party software modules installed on the cluster system which come with their own copies of various Intel MPI library versions. These software modules fall into the following categories:

	== Corrected software modules ==

	The following software modules have been corrected by the HPC software maintainers. They should currently work as expected.

	* ''StarCCM+'': The included Intel MPI 2018 library was replaced with a more recent version.

	* ''LS-DYNA'': The included Intel MPI library was replaced with a more recent version.

	* ''CST'': The license does not allow multi-node jobs, so the problematic code paths cannot be used.

	== Software modules with known fixes ==

	The following software modules require additional user interaction to work:

	* ''ANSYS Mechanical'' and ''Fluent'': The software has to be switched to OpenMPI using the '''-mpi=openmpi''' command line argument.

	* ''ANSYS CFX'': The software has to be switched to OpenMPI using the '''-start-method 'Open MPI Distributed Parallel' ''' command line argument.

	== Software modules without known fixes ==

	For the following software modules there is currently no known fix:

	* ''Abaqus'': Comes with Intel MPI 2017, for which there is currently no known fix. We are working on a solution. The ''cae/abaqus/2019'' software modules will not be removed because it can still be used for pre-/post-processing and single-node parallelisation using e.g. OpenMP.

S Raffeiner: /* Software modules without known fixes */

2020-10-27T10:18:27Z

Software modules without known fixes

← Older revision		Revision as of 12:18, 27 October 2020
Line 31:		Line 31:
	For the following software modules there is currently no known fix:		For the following software modules there is currently no known fix:

		⚫	* ''Abaqus'': Comes with Intel MPI 2017, for which there is currently no known fix. We are working on a solution. The ''cae/abaqus/2019'' software modules will not be removed because it can still be used for pre-/post-processing and single-node parallelisation using e.g. OpenMP.
	* ''cae/abaqus/2019'' (comes with Intel MPI 2017).We are working on a solution.

⚫	~~Non-~~working software modules will not be removed because ~~they~~ can still be used for pre-/post-processing and single-node parallelisation using e.g. OpenMP.

S Raffeiner: /* Corrected software modules */

2020-10-27T10:17:21Z

Corrected software modules

← Older revision		Revision as of 12:17, 27 October 2020
Line 13:		Line 13:
	The following software modules have been corrected by the HPC software maintainers. They should currently work as expected.		The following software modules have been corrected by the HPC software maintainers. They should currently work as expected.

	* StarCCM+: The included Intel MPI 2018 library was replaced with a more recent version.		* ''StarCCM+'': The included Intel MPI 2018 library was replaced with a more recent version.

	* LS-DYNA: The included Intel MPI library was replaced with a more recent version.		* ''LS-DYNA'': The included Intel MPI library was replaced with a more recent version.

	* CST: The license does not allow multi-node jobs, so the problematic code paths cannot be used.		* ''CST'': The license does not allow multi-node jobs, so the problematic code paths cannot be used.

	== Software modules with known fixes ==		== Software modules with known fixes ==

S Raffeiner at 10:13, 27 October 2020

2020-10-27T10:13:00Z

← Older revision		Revision as of 12:13, 27 October 2020
Line 5:		Line 5:
	* There is a bug in Intel MPI 2019.x which leads to crashes when multiple MPI applications which are linked against Intel MPI 2019.x are run on the same node (e.g. in the "single" partition). The first application will run normally, but all others will crash. This can be fixed by setting the environment variable '''I_MPI_HYDRA_TOPOLIB="ipl"'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.		* There is a bug in Intel MPI 2019.x which leads to crashes when multiple MPI applications which are linked against Intel MPI 2019.x are run on the same node (e.g. in the "single" partition). The first application will run normally, but all others will crash. This can be fixed by setting the environment variable '''I_MPI_HYDRA_TOPOLIB="ipl"'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.

	* There is a bug in Intel MPI 2019.x which leads to incorrect CPU binding/affinity in conjunction with the Slurm batch system used on the clusters. All MPI ranks will run on the same CPU core instead of being bound to all available CPU cores. This can be fixed by setting the environment variable '''I_MPI_HYDRA_BOOTSTRAP_EXEC_EXTRA_ARGS="--cpu-bind=none""'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.		* There is a bug in Intel MPI 2019.x which leads to incorrect CPU binding/affinity in conjunction with the Slurm batch system used on the clusters. All MPI ranks will run on the same CPU core instead of being bound to all available CPU cores. This can be fixed by setting the environment variable '''I_MPI_HYDRA_BOOTSTRAP_EXEC_EXTRA_ARGS="--cpu-bind=none"'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.

	There is a number of Third-Party software modules installed on the cluster system which come with their own copies of various Intel MPI library versions. These software modules fall into the following categories:		There is a number of Third-Party software modules installed on the cluster system which come with their own copies of various Intel MPI library versions. These software modules fall into the following categories:

S Raffeiner at 10:11, 27 October 2020

2020-10-27T10:11:02Z

← Older revision		Revision as of 12:11, 27 October 2020
Line 7:		Line 7:
	* There is a bug in Intel MPI 2019.x which leads to incorrect CPU binding/affinity in conjunction with the Slurm batch system used on the clusters. All MPI ranks will run on the same CPU core instead of being bound to all available CPU cores. This can be fixed by setting the environment variable '''I_MPI_HYDRA_BOOTSTRAP_EXEC_EXTRA_ARGS="--cpu-bind=none""'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.		* There is a bug in Intel MPI 2019.x which leads to incorrect CPU binding/affinity in conjunction with the Slurm batch system used on the clusters. All MPI ranks will run on the same CPU core instead of being bound to all available CPU cores. This can be fixed by setting the environment variable '''I_MPI_HYDRA_BOOTSTRAP_EXEC_EXTRA_ARGS="--cpu-bind=none""'''. The ''mpi/impi/2019'' and ''mpi/impi/2020'' modules provided on the cluster already set this variable.

	There is a number of Third-Party software modules installed on the cluster system which come with their own copies of various Intel MPI library versions. These software modules fall into ~~onthe~~ following categories.		There is a number of Third-Party software modules installed on the cluster system which come with their own copies of various Intel MPI library versions. These software modules fall into the following categories:

	== Corrected software modules ==		== Corrected software modules ==

S Raffeiner at 10:10, 27 October 2020

2020-10-27T10:10:37Z

@@ Line 1: / Line 1: @@
-After the [[BwUniCluster_2.0_Maintenance/2020-10|last maintenance]] the default settings of the following software modules offered on the system are currently know to be be incompatible with the new system:
+After the last regular [[BwUniCluster_2.0_Maintenance/2020-10|maintenance]] interval (from 06.10.2020 to 13.10.2020) the following issues with Intel MPI exist:
 * ''ANSYS Mechanical'' and ''Fluent'': The software has to be switched to OpenMPI using the '''-mpi=openmpi''' command line argument.
@@ Line 5: / Line 27: @@
 * ''ANSYS CFX'': The software has to be switched to OpenMPI using the '''-start-method 'Open MPI Distributed Parallel' ''' command line argument.

S Raffeiner: Created page with "After the last maintenance the default settings of the following software modules offered on the system are currently know to be be in..."

2020-10-27T09:00:36Z

Created page with "After the last maintenance the default settings of the following software modules offered on the system are currently know to be be in..."

New page

After the [[BwUniCluster_2.0_Maintenance/2020-10|last maintenance]] the default settings of the following software modules offered on the system are currently know to be be incompatible with the new system:

* ''ANSYS Mechanical'' and ''Fluent'': The software has to be switched to OpenMPI using the '''-mpi=openmpi''' command line argument.

* ''ANSYS CFX'': The software has to be switched to OpenMPI using the '''-start-method 'Open MPI Distributed Parallel' ''' command line argument.

* For the ''cae/abaqus/2019'' module there is currently no known fix. We are working on a solution.