Changes between Initial Version and Version 1 of SimtestNew

Timestamp:

Feb 24, 2009, 4:23:58 PM (17 years ago)

Author:

trac

Comment:

--

SimtestNew

@@ +1 @@
+=== Introduction ===
+
+The IPP Simulation Test ("Simtest") has been revised to be more automatic than its previous incarnations.  The desire has been for it to run through with the least possible amount of user intervention and interactivity.  At the present time (release 2.6), it requires a few pantasks commands to set up, and will run all the way through the creation, registration, detrend, chip, camera, fake, warp, stack, and diff stages; the magic stage will also be populated, though not run.
+
+=== WARNING ===
+
+This page refers to the ''revised'' Simtest, which will not be released until IPP 2.6.  If you are using an earlier version of the IPP (e.g., 2.5.1), then these instructions are not relevant to you.
+
+=== Instructions ===
+
+Having installed the IPP, go to a fresh directory.
+
+<pre>
+user@computer:/path/to/simtest>pantasks
+ 
+Welcome to pantasks - parallel task scheduler
+ 
+ 
+pantasks: module simtest.pro
+pantasks: simtest
+USAGE: simtest (testroot) (dbname) (hostname) (init)
+ if (init) == 0, just run the analysis tasks
+ if (init) == 1, (re)create the database and (re)inject the images
+ if (init) == 2, restart completely from scratch
+pantasks: simtest simtest.basic dbname computer 2
+</pre>
+
+Here:
+ * <code>simtest.basic</code> specifies the setup (see <code>ippTasks/simtest.basic.config</code>).
+ * <code>dbname</code> is your database name: it must have already been set up (e.g., using <code>dbadmin</code>), though it needn't have been created (using <code>pxadmin</code>) since the <code>simtest</code> will do that, depending on the <code>init</code> value you give it.
+ * <code>computer</code> is the name of a processing node: you must be able to <code>ssh</code> to this node without it prompting for a password.
+ * The <code>init</code> value (in the example here, <code>2</code>) specifies what needs to be done.
+
+After this, the database will be (re)created, simulated images will be generated, and the usual processing should automatically take place all the way through to the diff stage.  As usual, you can track how things go using the <code>status</code> command in pantasks, or using the IPP Monitor (if you've got it set up).