Scaling Up Silicon Photonic-based Accelerators: Challenges and Opportunities
arXiv:2109.08025 · doi:10.1063/5.0070992
Abstract
Digital accelerators in the latest generation of CMOS processes support multiply and accumulate (MAC) operations at energy efficiencies spanning 10-to-100~fJ/Op. But the operating speed for such MAC operations are often limited to a few hundreds of MHz. Optical or optoelectronic MAC operations on today's SOI-based silicon photonic integrated circuit platforms can be realized at a speed of tens of GHz, leading to much lower latency and higher throughput. In this paper, we study the energy efficiency of integrated silicon photonic MAC circuits based on Mach-Zehnder modulators and microring resonators. We describe the bounds on energy efficiency and scaling limits for NxN optical networks with today's technology, based on the optical and electrical link budget. We also describe research directions that can overcome the current limitations.
This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in "M. A. Al-Qadasi, L. Chrostowski, B. J. Shastri, and S. Shekhar , "Scaling up silicon photonic-based accelerators: Challenges and opportunities", APL Photonics 7, 020902 (2022)", and may be found at https://doi.org/10.1063/5.0070992
References in corpus (5)
- Photonics for artificial intelligence and neuromorphic computing
- Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
- Efficient, Compact and Low Loss Thermo-Optic Phase Shifter in Silicon
- Prospects and applications of photonic neural networks
- Silicon Photonic Architecture for Training Deep Neural Networks with Direct Feedback Alignment
Cited by in corpus (9)
- Silicon Photonic Architecture for Training Deep Neural Networks with Direct Feedback Alignment
- An Electro-Photonic System for Accelerating Deep Neural Networks
- Integrated Photonic Tensor Processing Unit for a Matrix Multiply: a Review
- Photonic neural networks based on integrated silicon microresonators
- Addressing the programming challenges of practical interferometric mesh based optical processors
- AnalogVNN: A fully modular framework for modeling and optimizing photonic neural networks
- Bayesian Photonic Accelerators for Energy Efficient and Noise Robust Neural Processing
- Optimized architectures for universal quantum state transformations using photonic path and polarization
- Scaling Analog Photonic Accelerators for Byte-Size, Integer General Matrix Multiply (GEMM) Kernels