Directions:

1) Create a database.  You might need to increase the size of tempdb.
In order to avoid filling the log, I checkpointed the database
after every set of queries (see closedb() in wisc_util.c).  Ideally
logging should be turned off for result tables from queries
but I never figured out how to do it.

2) Edit the file wiscbench.h. Things to edit include:

	/* verbose flag. 1 enables additional debugging output */
	#define VERBOSE 0

	/* name of the database created in step 1 */
	#define DATABASE "wiscbench"

	/* names and sizes of the two base relations */
	/* these will be created automatically by wisc_dbgen.c*/
	#define BASERELATION1 "TENKTUP1"
	#define BASERELATION2 "TENKTUP2"
	#define BASERELSIZE 10000

	/* names and size of two prime relations.  should */
	/* be 10% of size of base relations. Probably you can */
	/* leave the names APRIME and BPRIME alone */
	/* these will be created automatically by wisc_dbgen.c*/

	#define PRIME1 "APRIME"
	#define PRIME2 "BPRIME"
	#define PRIMESIZE 1000
    
For any reasonable configuration you should be running at least with
100,000 tuples in the base relations.  1,000,000 is minimal
for even small configurations.

3) Type "make wiscbench".  This should compile the benchmark programs
   (the Makefile is not quite right.  For some reason, everything is remade
   even if only one file has been changed.  It used to work but
   stopped for reason.) 

4) Type "wisc_run.csh" to run the benchmark.  The queries scale automatically
as the size of the database is changed.  First, queries without indices
are run.  Then the indices are created and the indexed queries are run.
Output is kind of messy as I never figured out how to turn off all
the garbage sybase likes to spit out.

5) Good luck. Let me know if you have problems.  
