Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper analyzes five photonic microring tensor core designs, focusing on how circuit ordering, unary encoding, and homodyne accumulation impact scalability. They find that unary encoding and homodyne accumulation offer the most promising route to increased parallelism in photonic tensor cores. The analysis uses a common optical power model to compare the different designs.
Unary encoding and homodyne accumulation could unlock significantly higher parallelism in photonic tensor cores, offering a potential leap in compute density for AI hardware.
We analyze five photonic microring tensor core designs with a common optical power model. The results show that circuit ordering, unary encoding, and homodyne accumulation shape scalability, with the last two offering the strongest path to higher parallelism.