[Assorted-commits] SF.net SVN: assorted: [379] hash-join/trunk/README

[<yan...@us...>]

Revision: 379
          http://assorted.svn.sourceforge.net/assorted/?rev=379&view=rev
Author:   yangzhang
Date:     2008-02-11 15:19:42 -0800 (Mon, 11 Feb 2008)

Log Message:
-----------
more informative readme

Modified Paths:
--------------
    hash-join/trunk/README

Modified: hash-join/trunk/README
===================================================================
--- hash-join/trunk/README	2008-02-11 23:11:38 UTC (rev 378)
+++ hash-join/trunk/README	2008-02-11 23:19:42 UTC (rev 379)
@@ -1,3 +1,9 @@
+% Parallel Hash Join
+% Yang Zhang
+
+Overview
+--------
+
 This is a simple implementation of parallel hash joins. I'm using this as a
 first step in studying the performance problems in multicore systems
 programming. This implementation is tailored for a particular dataset, the IMDB
@@ -3,11 +9,31 @@
 `movies.list` and `actresses.list` files, which may be found [here].
 
-The `tools/` directory contains `DbPrep.scala`, which is a filter for the
-`.list` files to prepare them to be more easily parsed by the hash join
-application.
+Requirements
+------------
 
-The `tools/` directory also contains `LogProc.scala`, which processes stdout
-concatenated from multiple runs of the program. This will produce the time and
-speedup plots illustrating the scalability of the system.
+- [C++ Commons] svn r370+
+- [libstdc++] v4.1
 
+Supporting Tools
+----------------
+
+`DbPrep` filters the `.list` files to prepare them to be parsed by the hash
+join.
+
+`LogProc` processes stdout concatenated from multiple runs of the program. This
+will produce the time and speedup plots illustrating the scalability of the
+system. This has actually been made into a generic tool and will be moved to
+its own project directory later.
+
+`Titles` extracts the titles from the output of `DbPrep` on `movies.list`.
+
+Related
+-------
+
+I used [HashDist] to experiment with the chaining of various hash functions on
+this dataset and observe the distribution.
+
 [here]: http://us.imdb.com/interfaces#plain
+[libstdc++]: http://gcc.gnu.org/libstdc++/
+[C++ Commons]: http://assorted.sf.net/
+[HashDist]: http://assorted.sf.net/


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.