Compiling and running MPI programs - Revision history

Bremer at 10:41, 11 November 2019

2019-11-11T10:41:53Z

Thw: /* Infiniband */

2010-08-09T12:47:21Z

‎Infiniband

Thw at 08:56, 2 April 2009

2009-04-02T08:56:00Z

Thw at 08:44, 15 January 2009

2009-01-15T08:44:01Z

Thw: New page: We only support openMPI, currently the best MPI implementation. The widely used MPICH2 implementation, is difficult to maintain and Debian/Ubuntu packages are not available. The lamMPI imp...

2009-01-05T13:35:42Z

New page: We only support openMPI, currently the best MPI implementation. The widely used MPICH2 implementation, is difficult to maintain and Debian/Ubuntu packages are not available. The lamMPI imp...

New page

We only support openMPI, currently the best MPI implementation. The widely used MPICH2 implementation, is difficult to maintain and Debian/Ubuntu packages are not available. The lamMPI implementation has stopped in favour to openMPI.

== Compiling ==

To compile MPI programms your are recommended to use the commands mpicc, mpif77, mpif90. By default this wrapper refer to the GNU Compilers gcc and gfortran. To use the wrapper with the Intel Compilers you have to define some environment variables.

@@ Line 12: / Line 12: @@
 </pre>
-Especially if you use icc, it will be better if use the recent [[Intel_Compiler_Temp|Intel Compiler 11.0]].
+Especially if you use icc, it will be better if you use the recent [[Intel_Compiler_Temp|Intel Compiler 11.0]].
 == Infiniband ==
-The latest nodes of DFG and 'quantum' Cluster have Infiniband Network. It is accessible through the queue 'dfg-ib' or 'quantum' and is used by default when using our openMPI installation. Infiniband provides high bandwith with low latency. It can transport 20 GB/s with latency of 4 µs. In comperison, normal Gigabit Ethernet provides 1 GB/s with at least 30 µs latency. Latency is the time a data packet needs to travel from the source to its target. This is the main drawback when using Gigibit Ethernet for MPI communication.
+The 'itp'-, 'itp-big'-, 'dfg-xeon'-, 'iboga'-, 'dreama'- and 'barcelona'-nodes have Infiniband Network. It is used by default when using our openMPI installation. Infiniband provides high bandwith with low latency. It can transport 20 GB/s with latency of 4 µs. In comparison, normal Gigabit Ethernet provides 1 GB/s with at least 30 µs latency. Latency is the time a data packet needs to travel from the source to its target. This is the main drawback when using Gigibit Ethernet for MPI communication.

← Older revision		Revision as of 12:47, 9 August 2010
Line 16:		Line 16:
	== Infiniband ==		== Infiniband ==

−	The latest nodes of DFG Cluster have Infiniband Network. It is accessible through the queue 'dfg-ib' and is used by default when using our openMPI installation. Infiniband provides high bandwith with low latency. It can transport 20 GB/s with latency of 4 µs. In comperison, normal Gigabit Ethernet provides 1 GB/s with at least 30 µs latency. Latency is the time a data packet needs to travel from the source to its target. This is the main drawback when using Gigibit Ethernet for MPI communication.	+	The latest nodes of DFG and 'quantum' Cluster have Infiniband Network. It is accessible through the queue 'dfg-ib' or 'quantum' and is used by default when using our openMPI installation. Infiniband provides high bandwith with low latency. It can transport 20 GB/s with latency of 4 µs. In comperison, normal Gigabit Ethernet provides 1 GB/s with at least 30 µs latency. Latency is the time a data packet needs to travel from the source to its target. This is the main drawback when using Gigibit Ethernet for MPI communication.

← Older revision		Revision as of 08:56, 2 April 2009
Line 11:		Line 11:

	</pre>		</pre>
−	~~<br>~~	+
		+	Especially if you use icc, it will be better if use the recent [[Intel_Compiler_Temp\|Intel Compiler 11.0]].
		+
		+	== Infiniband ==
		+
		+	The latest nodes of DFG Cluster have Infiniband Network. It is accessible through the queue 'dfg-ib' and is used by default when using our openMPI installation. Infiniband provides high bandwith with low latency. It can transport 20 GB/s with latency of 4 µs. In comperison, normal Gigabit Ethernet provides 1 GB/s with at least 30 µs latency. Latency is the time a data packet needs to travel from the source to its target. This is the main drawback when using Gigibit Ethernet for MPI communication.

@@ Line 1: / Line 1: @@
-We only support openMPI, currently the best MPI implementation. The widely used MPICH2 implementation, is difficult to maintain and Debian/Ubuntu packages are not available. The lamMPI implementation has stopped in favour to openMPI.
+We only support openMPI, currently the best MPI implementation. The widely used MPICH2 implementation, is difficult to maintain and Debian/Ubuntu packages are not available. The lamMPI implementation has stopped development in favour to openMPI.
-== Compiling ==
+== Compiling  ==
 To compile MPI programms your are recommended to use the commands mpicc, mpif77, mpif90. By default this wrapper refer to the GNU Compilers gcc and gfortran. To use the wrapper with the Intel Compilers you have to define some environment variables.