Doctoral Thesis Proposal - Mingkuan Xu

Location:
In Person and Virtual - ET - Gates Hillman 8102 and Zoom

Speaker:
MINGKUAN XU , Ph.D. Student, Computer Science Department , Carnegie Mellon University
https://mingkuan.taichi.graphics/

Optimization and Simulation of Quantum Circuits

Optimizing quantum circuits and simulating them at scale remain critical bottlenecks: manual design of quantum circuit optimizations is labor-intensive and device-specific, while simulators struggle with exponential resource costs. This thesis delivers tools to tackle these challenges. 

First, I introduce Quartz, a superoptimizer that automates the generation and verification of circuit transformations for arbitrary quantum gate sets. By systematically exploring small circuits and employing an automated theorem prover (Z3), Quartz discovers both expert-designed and novel optimizations, outperforming hand-tuned optimizers across various gate sets. 

Next, I present Atlas, a distributed GPU-based simulator that hierarchically partitions circuits to exploit available data parallelism while minimizing communication costs, running over 2× faster than state-of-the-art GPU simulators. Atlas minimizes communication overhead via integer linear programming to allocate "nearby" gates to "nearby" GPUs and maximizes throughput through dynamic programming for kernel scheduling. 

Finally, I propose an initial formal verification framework to certify each application of transformation-based optimizers like Quartz, paving the way for full correctness guarantees. Together, these contributions advance automated, scalable, and reliable quantum computing workflows for emerging devices. 

Thesis Committee

Zhihao Jia (Co-chair)
Umut A. Acar (Co-chair)
Ryan O'Donnell
Yongshan Ding (Yale University)
 

Additional Information

In Person and Zoom Participation.  See announcement.

Add event to Google
Add event to iCal