summaryrefslogtreecommitdiffstats
path: root/testsuites/psxtmtests/README
diff options
context:
space:
mode:
authorJoel Sherrill <joel.sherrill@OARcorp.com>2010-12-06 16:35:31 +0000
committerJoel Sherrill <joel.sherrill@OARcorp.com>2010-12-06 16:35:31 +0000
commit519928c33386f1d6d5ef6cb9d34d93d2ef89c8a0 (patch)
tree74be705b7729a84fd13ad6d997bf2202909ab837 /testsuites/psxtmtests/README
parent2010-12-03 Joel Sherrill <joel.sherrill@oarcorp.com> (diff)
downloadrtems-519928c33386f1d6d5ef6cb9d34d93d2ef89c8a0.tar.bz2
2010-12-06 Joel Sherrill <joel.sherrill@oarcorp.com>
* .cvsignore, ChangeLog, Makefile.am, README, configure.ac, psxtmtests_plan.csv, psxtmmutex02/.cvsignore, psxtmmutex02/Makefile.am, psxtmmutex02/init.c, psxtmmutex02/psxtmmutex02.doc, psxtmthread01/.cvsignore, psxtmthread01/Makefile.am, psxtmthread01/init.c, psxtmthread01/psxtmthread01.doc, psxtmthread03/.cvsignore, psxtmthread03/Makefile.am, psxtmthread03/init.c, psxtmthread03/psxtmthread03.doc: New files.
Diffstat (limited to 'testsuites/psxtmtests/README')
-rw-r--r--testsuites/psxtmtests/README25
1 files changed, 25 insertions, 0 deletions
diff --git a/testsuites/psxtmtests/README b/testsuites/psxtmtests/README
new file mode 100644
index 0000000000..428c41eaac
--- /dev/null
+++ b/testsuites/psxtmtests/README
@@ -0,0 +1,25 @@
+#
+# $Id$
+#
+
+This directory contains the RTEMS POSIX Timing Test Suite.
+The tests in this directory are used to measure the execution
+time of RTEMS directive and some critical internal functions.
+The results of these test are useful for comparison of RTEMS
+performance
+
++ between versions of RTEMS
++ between different boards
++ across different architectures
+
+These tests attempt to benchmark RTEMS as the user would. They measure
+the execution time for a directive (under various circumstances)
+from the time RTEMS is entered until it is executed. Where possible,
+the time is the average of at least 100 invocations of the directive.
+
+The accuracy of the times reported by these benchmarks is very dependent
+on the resolution of the timer being used. It is also important to
+insure that all sources of hardware interrupts are disabled during
+execution of the tests. This insures that the directive time reported
+does not include any interrupt time.
+