1 files changed, 41 insertions, 0 deletions
diff --git a/ACE/TAO/orbsvcs/tests/Notify/performance-tests/scripts/3_Path_Period_10ms_Lanes/README b/ACE/TAO/orbsvcs/tests/Notify/performance-tests/scripts/3_Path_Period_10ms_Lanes/README
new file mode 100644
index 00000000000..55e2003fc75
--- /dev/null
+++ b/ACE/TAO/orbsvcs/tests/Notify/performance-tests/scripts/3_Path_Period_10ms_Lanes/README
@@ -0,0 +1,41 @@
+## $Id$
+
+3 Paths : Load vs Throughput experiment with period=10ms
+========================================================
+
+In this test we setup 3 Paths (In a Path, a single supplier sends
+events to a consumer and only that consumer) at High, Medium and Low Priority.
+
+Each Supplier tries to send an event every 10ms (100Hz).
+Each consumer performs CPU work according to the load supplied to it.
+
+We increase the load and measure the throughput recorded.
+
+
+Running the experiment:
+======================
+
+Run the "run_all.pl" script in this directory to run the test using RT-Notification.
+as "run_all.pl --rt"
+
+If you want to generate results using the CosNotification, run as
+"run_all.pl --cos"
+
+To generate the data for both cases run as "run_all.pl --all"
+
+The RT-Notification results are generated in the Load_X/rt_output directory.
+The CosNotification results are generated in the Load_X/cos_output directory.
+
+e.g., Load_0/rt_output/c_high.dat would contain the statistics
+recorded for the consumer when the Load is 0.
+
+
+Expected Results:
+================
+
+RT-Notification : As the load increases the throughput for the consumers should fall
+however the high priority path should be least affected followed by
+the medium and then the low priority.
+
+CosNotification : As the CosNotification is unable to communicate
+priorities, the throughput of all 3 paths will fall in tandem.