Performance analysis of self-electro-optic-effect-device-based (SEED-based) smart-pixel arrays used in data sorting

Area of Science:

• Optoelectronics
• Semiconductor device physics
• Computer architecture

Background:

• Self-electro-optic-effect-device (SEED) based smart-pixel arrays are crucial for parallel optical processing.
• Performance is influenced by semiconductor technology and pixel complexity.
• Sorting tasks serve as a benchmark for evaluating smart-pixel array capabilities.

Purpose of the Study:

• To analyze performance factors of SEED-based smart-pixel arrays.
• To identify optimal configurations for processing power and communication rates.
• To highlight areas for technological advancement in smart-pixel design.

Main Methods:

• Analysis of performance factors considering semiconductor technology and pixel complexity.
• Evaluation using a sorting task as a practical application.
• Comparison of different smart-pixel configurations, including CMOS-SEED nodes.

Main Results:

• Complementary metal-oxide semiconductor (CMOS)-SEED 2 × 2 self-routing nodes operated with quasi-continuous-wave (quasi-cw) mode lasers demonstrate superior processing power.
• These configurations achieve the highest on- and off-chip communication rates.
• The study underscores the limitations of current designs for demanding applications.

Conclusions:

• CMOS-SEED 2 × 2 nodes with quasi-cw lasers represent a high-performance solution for smart-pixel arrays.
• Advancements in front-end amplifiers are essential to fully realize the potential of smart-pixel technology.
• Further research into integrated optoelectronic systems is warranted.