[Swift-commit] r8217 - SwiftTutorials/ATPESC_2014-08-14/swift-k/doc

[yadunandb at ci.uchicago.edu]

Author: yadunandb
Date: 2014-08-14 15:00:44 -0500 (Thu, 14 Aug 2014)
New Revision: 8217

Modified:
   SwiftTutorials/ATPESC_2014-08-14/swift-k/doc/README
Log:

Committing changes from midway to tukey


Modified: SwiftTutorials/ATPESC_2014-08-14/swift-k/doc/README
===================================================================
--- SwiftTutorials/ATPESC_2014-08-14/swift-k/doc/README	2014-08-14 19:59:18 UTC (rev 8216)
+++ SwiftTutorials/ATPESC_2014-08-14/swift-k/doc/README	2014-08-14 20:00:44 UTC (rev 8217)
@@ -9,7 +9,7 @@
 ////
 
 This tutorial is viewable at:
-http://swiftlang.org/tutorials/midway/tutorial.html
+http://swift-lang.org/ATPESC_2014/swift-k/README.html
 
 ////
 
@@ -23,7 +23,7 @@
 
 Compute-node exercises, running apps via qsub and aprun:
 
-  p4 - Running apps on RCC Midway cluster
+  p4 - Running apps on Tukey
   p5 - Running on multiple pools of compute nodes
   p6 - Running a more complex workflow pattern
 
@@ -48,26 +48,22 @@
 clusters, clouds, grids, and supercomputers.
 
 In this tutorial, you'll be able to first try a few Swift examples
-(parts 1-3) on a Midway login host, to get a sense of the
-language. Then in parts 4-6 you'll run similar workflows on Midway
+(parts 1-3) on a Tukey login host, to get a sense of the
+language. Then in parts 4-6 you'll run similar workflows on Tukey
 compute nodes, and see how more complex workflows can be expressed
 with Swift scripts.
 
 Swift tutorial setup
 --------------------
-If you are using a temporary/guest account to access Midway, follow
-the instructions at http://docs.rcc.uchicago.edu/tutorials/intro-to-rcc-workshop.html
-for more information on using a yubikey to log in.
 
-Once you are logged into the Midway login host, run the following commands
+Once you are logged into the Tukey login host, run the following commands
 to set up your environment.
 
 -----
 $ cd $HOME
-$ wget http://swiftlang.org/tutorials/midway/swift-midway-tutorial.tar.gz
-$ tar xvfz swift-midway-tutorial.tar.gz
-$ cd swift-midway-tutorial
-$ source setup.sh
+$ svn co https://svn.ci.uchicago.edu/svn/vdl2/SwiftTutorials/ATPESC_2014-08-14/swift-k
+$ cd swift-k
+$ source setup.tukey.sh
 -----
 
 Verify your environment
@@ -79,20 +75,13 @@
 $ swift -version  # verify that you have Swift 0.95 RC1
 -----
 
-NOTE: If you re-login or open new ssh sessions, you must re-run `source setup.sh` in each ssh shell/window.
+Ensure that you have the PROJECT env variable set
 
-To check out the tutorial scripts from SVN
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-If you later want to get the most recent version of this tutorial from
-the Swift Subversion repository, do:
-
 -----
-$ svn co https://svn.ci.uchicago.edu/svn/vdl2/SwiftTutorials/swift-midway-tutorial
+$ export PROJECT=ATPESC2014
 -----
 
-This will create a directory called "swift-midway-tutorial" which contains all of the
-files used in this tutorial.
+NOTE: If you re-login or open new ssh sessions, you must re-run `source setup.sh` in each ssh shell/window.
 
 
 Simple "science applications" for the workflow tutorial
@@ -458,13 +447,13 @@
 on the site's compute nodes.
 
 
-Running applications on Midway compute nodes with Swift
+Running applications on Tukey compute nodes with Swift
 -------------------------------------------------------
 
-Part 4: Running a parallel ensemble on Midway compute nodes
+Part 4: Running a parallel ensemble on Tukey compute nodes
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-`p4.swift` will run our mock "simulation" applications on Midway
+`p4.swift` will run our mock "simulation" applications on Tukey
 compute nodes. The script is similar to as `p3.swift`, but specifies
 that each simulation app invocation should additionally return the
 log file which the application writes to `stderr`.
@@ -507,33 +496,39 @@
 Swift's `swift.properties` configuration file allows many parameters to
 specify how jobs should be run on a given cluster.
 
-Consider for example that Midway has several Slurm partitions. The sandyb
-partition has 16 cores, and the westmere partition has 12 cores. Depending
-on the application and which partitions are busy, you may want to modify
-where you run.
 
 Here is an example of the swift.properties in the part04 directory:
 
 -----
-site=westmere
-site.westmere.slurm.reservation=workshop
-site.westmere.slurm.exclusive=false
+site=tukey
+
+site.tukey {
+   jobmanager=localhost:cobalt
+   taskWalltime=00:15:00
+   workdir=/home/$USER/swiftwork
+   filesystem=local
+   maxJobs=4
+   tasksPerWorker=8
+   taskThrottle=300
+   initialscore=10000
+}
+
+
+app.tukey.simulate=$PWD/../app/simulate
+app.tukey.stats=$PWD/../app/stats
+app.tukey.md=$PWD/../bin/md
+app.tukey.mdstats=$PWD/../app/mdstats
+app.tukey.mdmaxk=$PWD/../app/mdmaxk
+app.tukey.renderframe=$PWD/../app/renderframe
+app.tukey.convert=/usr/bin/convert
 -----
 
-The first line, site=westmere, allows you to define which partition to run on. Swift includes
-templates for each partition that provides some reasonable default values. Other valid partitions
-are amd, bigmem, gpu, local, and sandyb.
+The first line, site=tukey, allows you to define which site to run on. Swift also includes
+templates for several sites that provides some reasonable default values.
 
 The second and third line override some of the default values, by specifying a reservation that will
 be used for this session, and enabling node sharing.
 
-Try changing the queue to run on the sandyb queue. The new swift.properties should look like this:
------
-site=sandyb
-site.sandyb.slurm.reservation=workshop
-site.sandyb.slurm.exclusive=false
-----
-
 Performing larger Swift runs
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~