|
For the past 20 years, the
GDSII format governed the efficiency and performance of
our data files, databases and the industry’s
computational algorithms. It governs bit-efficiency for
both database and data files. It governs as well the
shape and thus the run time of the algorithms that drive
the performance of our databases and computational
geometry operations.
Squeezing out more
efficiency and performance from existing data structures
constrained by GDSII, the format whose framers pegged
its life expectancy at five years, has become
increasingly difficult, risky and expensive. As data
structures failed to accommodate the need for greater
computing efficiency and performance, brute force
compute power overcame its shortcomings.
Data structure does drive
computational efficiency and performance. Today, the
data structure within databases cause the creation of a
large amount of redundant structural overhead. There are
many, many repeated instances of data. Databases and
computational geometry operations conduct computations
on 10-100X more data records than necessary. There is a
direct relation between the number of required
computations and run time performance.
OASIS low-level primitives
in the data structure of databases squeezes out
repetitive instances of data as successfully as they did
at the file level. The next step with OASIS as an
implementation is to modify database low-level
primitives and shape the algorithms so that they can
detect and operate off the new data structure.
Computational algorithms that can detect and compute off
the new data structure run 10-100X faster.
|
