Shanshan Xie
Research Experience
Exploring, creating, and debugging
eDRAM-CIM
Low Power Compute-in-Memory Design for ML Accelerator
- Successfully taped out this project from circuit design, simulation, custom-layout, APR and post-silicon verification(on-going and got accepted to ISSCC 2021.
- Designed memory blocks to be used for on-chip machine learning computation using 65nm CMOS process.
- Own all digital and analog blocks in the design while optimizing power consumption and throughput.
- Worked on behavior model of analog and digital circuits using Verilog AMS and completed mixed signal simulation.
- Developed energy model to evaluate the design and used Python to model the circuit performance.
Advanced Encryption Standard (AES)
- Successfully taped out this project from circuit design, simulation, custom-layout, APR and post-silicon verification.
- Utilized Cadence Virtuoso to design and simulate power management unit as a supporting circuit for the AES core using 40nm CMOS process.
- Performed digital development activities such as RTL design, synthesis and validation.
- Validated post silicon circuit performance in the lab in a team of two.
- Designed 4-layer PCB for testing purpose using Altium Designer.
Analog IC Design Competition
- Designed a low power differential operational transconductance amplifier (OTA) with common mode feedback using the 0.18um CMOS process.
- Achieved 120uW by using current-reuse telescopic topology with rail-to-rail class-AB output stage.
- Won second prize in the final analog IC design competition.
Independent Study: Ultra-Low Power Pressure Ulcer Prevention System
WPI, August 2017 – May 2018
Advisor: Professor John McNeill, Dean of Engineering
- Developed a time-based resistance measurement algorithm with error analysis, simulation and verification to save more power and increase the accuracy for the resistive-force sensor system
- Modeled non-idealities due to the creep behavior of piezoresistance for the conductor-filled polymer composites
- Calibrated resistive-force sensor using MATLAB to bring down the measurement error from 10% to 2%
Major Qualifying Project (Senior Thesis)
WPI, September 2017 – May 2018
Advisor: Professor Reinhold Ludwig, Professor and Department Head
- Developed a small, accurate, and efficient integrated circuits capable of measuring a variety of vital parameters to create an innovative wearable health monitor on a team of three.
- Proposed features to monitor the user’s heartbeat, blood pressure, internal temperature, respiration rate, and galvanic skin response, as well as providing fall detection.
- Worked on SPI Communication between microcontroller CC2650 and accelerometer ADXL362, using Code Composer Studio and Analog-to-Digital Converter inside the microcontroller to process ECG heart rate monitor AD8233 analog data.
