Menu

Case Studies

Home /
Case Studies
Processor Development

Project Overview

Project Summary

AIONSI executed a full Netlist-to-GDSII physical design engagement for a networking processor built on SMC 28nm technology. The project involved implementing a flat chip architecture with multiple high-speed analog IPs and stringent performance targets, requiring deep expertise in floor planning, congestion optimization, timing closure, and signoff methodologies.

Key Achievements

Key Specifications

  • Technology Node: 28nm (TSMC)

  • Chip Dimensions: 4.5 mm × 5.535 mm

  • Team Size: 5 Physical Design Engineers

  • Primary Tool: Cadence Innovus Implementation System

  • Scope: Complete Turnkey PD — Netlist to GDSII

  • Target Application: High-performance Networking Chip

  • Operating Frequency: 700 MHz

  • Integrated Analog IPs: PLL, DDR PHY, SERDES blocks

  • Design Approach: Flat chip architecture


Major Features & Technical Highlights

1. Clocking Complexity

The design contained 10 independent clocks, including scan clocks, requiring:

  • Careful CTS planning

  • Clock skew and insertion delay optimization

  • Multi-domain timing analysis and cleanup

2. Congestion Management & Area Optimization

Reducing routing congestion and meeting area targets was a significant challenge due to:

  • Analog IP placement constraints

  • High standard-cell density

  • Flat implementation approach

The team optimized placement, routing layers, blockages, and pin assignment to achieve a clean layout within the required die size.

3. Timing Closure Across All TSMC Signoff Corners

Complete timing closure was achieved across:

  • All PVT corners

  • Functional, test, and scan modes

  • Setup, hold, and clock-gating checks

This ensured robust performance for silicon across all operating conditions.


Final Outcome

  • Successful delivery of GDSII with full signoff compliance

  • Achieved 700 MHz performance target

  • Completed routing and timing closure without violating congestion thresholds

  • Managed complex analog IP integration in a flat physical design flow

  • Demonstrated strong teamwork and efficient execution with a compact 5-engineer team