Detailed Project Report · Custom CMOS · Arithmetic · 2025–2026

16-bit Multiply–Accumulate Unit

Qinyan Zeng
Department of Electrical and Computer Engineering
University of Southern California · qyzeng.work@gmail.com

Abstract

The MAC project extends the custom arithmetic work to a hierarchical multiply–accumulate datapath. Partial products are formed with Radix-4 Booth encoding, reduced with 6-to-2 compressors, and resolved with a carry-lookahead adder before physical verification and extraction.

Keywords—Cadence Virtuoso, Radix-4 Booth, 6-to-2 compressor, CLA, DRC, LVS, PEX.

I. Project summary

Datapath16-bit MAC
EncodingRadix-4 Booth
Reduction6-to-2 compressors
Final adderCarry-lookahead

II. Scope and contribution

  • Constructed the transistor-level MAC hierarchy.
  • Integrated Booth encoding, compressor reduction, and final addition.
  • Completed layout, DRC, LVS, and PEX.

III. Methodology

  1. Radix-4 partial-product encoding
  2. Compressor-tree reduction
  3. Carry-lookahead final addition
  4. Hierarchical custom layout

IV. Verification

  • Block-level and integrated functional simulation.
  • DRC and LVS.
  • PEX-based post-layout verification.

V. Reported results

  • Completed the 16-bit MAC from transistor-level schematic through extracted layout.
  • Closed the stated physical verification flow.

VI. Artifacts and data availability

The résumé establishes the following artifact classes. Public-safe figures and report excerpts will be attached after review:

  • Datapath hierarchy
  • Booth encoder
  • Compressor stages
  • Integrated layout
  • DRC/LVS/PEX summary