linearized.t	-- Shows our accuracy on linearized subscripts
by3.t		-- the example that broke KAP, Paraphase2 and Parascope 

example1.t      -- An example of killing
example1a.t     -- continues example 1
example1b.t     -- continues example 1a
example3.t	-- An example with a temporary work array that 
		   can be expanded/privatized
example4.t	-- A more subtle version of example 3
example5.t      -- A more subtle version of example 3 in which the
		   dependence carried by the outer loop cannot be killed
example6.t	-- An example showing the refinement of a coupled dependence
example7.t	-- Shows the ability to zap a dependence
olda.t		-- An example of code from the Perfect club benchmarks
		   showing that we can perform the analysis required to
		   privatize arrays in real codes

ivr.t		-- Displays our induction variable recognition capabilities

mxm.t		-- Apply unified transformation to see effects of different
		   loop orderings on parallelism and cache performance
lu.t		-- Apply unified transformation to obtain all 6 permutations
		   of the loops (something that can't be obtained using
		   loop interchange)
petersen.t	-- Paul Petersen described this example as a code fragment
		   who's parallelism could not be found by traditional 
		   analysis methods. Using expansion and a unified 
		   transformation allows us find the parallelism.
syr2k.t		-- Using any combination of loop interchanges, this code
		   will still have 4 references that are not stride 0 or
		   stride 1 in the innermost loop. By applying a unified
		   transformation with normalization turned on, we find
		   skewed transformations that make all array references
		   stride 1.
