Simply speaking, an OASIS database is one which
uses OASIS low-level primitives within its native data
structure, or as a thin non-intrusive layer to
represent and store geometric data in layout
databases. The primitives allow the representation of
repeated data by a single canonical representation
along with a repetition function. Singular
representations of data, if implemented properly, can
achieve significant speed-up for Stream IN/Out
operations. CAD operations able to detect and digest
repetition data from an OASIS database can achieve
significant speed-up as well.
Oasis Tooling's technology, called dbPrune prunes
away equivalent geometrical shapes, instances and text
records, replacing them by a single copy, annotated
with repetition information. dbPrune comprises a
thin non-intrusive layer on top of an existing
database, and several computational geometry
algorithmic techniques which give an order of
magnitude performance and disk-usage advantage when
compared to state-of-art databases. This repetition
information can be attached as a property on the
element, or can be stored outside the database in an
auxiliary file. We have implemented this on the
OpenAccess database reference implementation. In a
pruned database, GDSII Stream-In/Out remains possible.
dbPrune will not interfere with other facets of
a design database such as HDL, logic, circuit and
physical layout simulation results.
Included with dbPrune are pattern recognition,
repetition detection and equivalence class detection
along with several data optimization versus
performance optimization methods.
dbPrune is part of the functionality of our C++/C
API or can be purchased separately.
