Two-level Benchmarks 1997
We ask you to post any comments and suggestions about the contents of this
directory using our
PosterNotes.
We have generated PLA equivalence classes with number of inputs
ranging from 3 to 16 and syndrome of 0.125, 0.25, 0.5. There are 100 PLAs
in each class. The first sets of experiments on some of these descriptions
are available in
1996-Thesis-PhD-Zemva.
A syndrome for an n-input PLA is defined as
s = (no. of 1's in the PLA output)/(2^n)
It is known that PLAs with s=0.5 involve highest amount of computations and
hence may result in a circuit of highest complexity for a given input size.
For example, see:
K.T. Cheng and V.D. Agrawal,
"An Entropy Measure for the Complexity of Multi-Output Boolean Functions",
Design Automation Conference, pp. 302-305, June 1991.
As shown in the list,
the current directory stores PLAs
with s=0.125 and s=0.5 and input sizes of n=4 to n=8. Efficient
construction of PLAs of other input sizes and syndrome are under
construction. If you need PLAs of other input size/syndrome, please send a
request to ghosh@cbl.ncsu.edu.
To use these PLAs:
For more details about data formats, see the User Guide
1991-IWLS-UG-Saeyang