Differences

This shows you the differences between two versions of the page.

--- doku:time [2015/03/26 16:19] – ir
+++ doku:time [2015/04/17 11:45] (current) – [Hybrid OMP and MPI] ir
@@ Line 4: / Line 4: @@
 or do not provide adequate access to high resolution timers (See also Section 8.6 Timers and Synchronization in
 [[http://www.mpi-forum.org/docs/mpi-3.0/mpi30-report.pdf|mpi-3.0 Standard]].).
+==== MPI internal functions ====
 === MPI_WTIME() ===
@@ Line 13: / Line 15: @@
 MPI_WTIME returns a floating-point number of seconds, representing elapsed wallclock
-time since some time in the past and may be used as follows:
+time since some time in the past.
-<code>{
+Generally, **times can ONLY be compared within the same MPI-process.** The times returned are **local to the node that called them**. Different nodes need not to be synchronized.
-double starttime, endtime;
+(MPI_WTIME_IS_GLOBAL is a Boolean variable that is 1 if synchronized and 0 if not. In Open MPI
-starttime = MPI_Wtime();
+MPI_WTIME_IS_GLOBAL does not have a valid value since the clocks are not guaranteed to be synchronized.)
-.... stuff to be timed ...
-endtime = MPI_Wtime();
+MPI_WTIME has to be called inside the MPI region
-printf("That took %f seconds\n",endtime-starttime);
+<code>
-}</code>
+MPI_Init(&argc, &argv);
-The times returned are **local to the node that called them**. Different nodes need not to be synchronized.
+...
-(MPI_WTIME_IS_GLOBAL is a Boolean variable that is 1 if synchronized and 0 if not.)
+MPI_Finalize( );
+</code>
+otherwise its behavior is undefined.
 === MPI_WTICK() ===
 <code>double MPI_Wtick(void)
@@ Line 28: / Line 33: @@
 DOUBLE PRECISION MPI_WTICK()</code>
 MPI_WTICK returns the resolution of MPI_WTIME in seconds, i.e. the time between successive clock ticks given in seconds.
+=== Examples ===
+In this  {{:doku:mpitime.c|example file}}, t1=MPI_WTIME() and t2=MPI_WTIME() enclose the region wich should be timed.
+There are two possibilities. Either a MPI_Barrier can be set before each time call. Then the time measured is the time needed that all MPI processes are finished. Or t1 and t2 are defined without a barrier, yielding the execution times for the individual processes.
+==== omp internal calls ====
+=== omp_get_wtime ===
+c example:
+<code>
+    double start = omp_get_wtime( );
+    Sleep(1000);
+    double end = omp_get_wtime( );
+    double wtick = omp_get_wtick( );
+</code>
+==== gettimeofday() ====
+gives the time of your wrist watch.
+==== clock ====
+gives the amount of time that the CPU is busy for a particular task from one call to the next.
+===== Hybrid OMP and MPI ======
+{{ :doku:mpi_omp:hybrid.png?400 |}}
+{{ :doku:mpi_omp:hybridbarrier.png?400 |}}
+{{:doku:mpi_omp:hybridtime.c|}}
+====== Serial codes ======
+==== fortran ====
+<code>
+SYSTEM_CLOCK
+DATE_AND_TIME
+CPU_TIME
+</code>