Completion Optimization
How can a PDBF be optimized before conventional logic synthesis?
The Official Home of GT Synthesis
A unified research program built on Completion Optimization, Contextual PDBFs, and Native Sequential Device PDBFs.

The trilogy
How can a PDBF be optimized before conventional logic synthesis?
How can ordinary RTL reveal hidden partially defined Boolean functions?
Why is sequential logic inherently partially defined?
The scientific progression
Completion Optimization selects a favorable function from its Legal Completions.
Unreachable local terms inside RTL blocks become contextual don’t-cares.
ONLY reachable behavior defines correctness; unreachable behavior defines optimization freedom.
Signature examples
One don’t-care, two Legal Completions, and a decisive implementation difference.
A familiar decoder demonstrating multiple area–depth trade-offs across Legal Completions.
Two unreachable state/input combinations reduce the complete implementation from 7 gates to 4.
Semantics-driven design
Characters, tokens, dates, sensor measurements, protocol symbols, and controller states often occupy only a fraction of their available encoding spaces.
Those restrictions are not artificial don’t-cares. They come directly from application semantics and naturally define PDBFs that GT can optimize.
Explore semantics-driven hardware designSequential logic is inherently partially defined.
ONLY reachable behavior defines correctness.
Unreachable behavior defines optimization freedom.