PACE® 2
Optoelectronic Accelerated Computing Card
-
8bit
Optical Computing Output Precision
-
128x128
Configurable Optical Matrix
-
3D TSV Advanced Packaging Technology
-
Over 40,000 Photonic Devices
Product Overview
PACE®️2 is an optoelectronic accelerated computing card released by Lightelligence in 2025, showing key features of high performance, programmability, versatility, and advanced packaging technology. With the essence of a 128×128 optical matrix multiplier, it allows dynamic configuration of computational matrix coefficients via APIs, offering exceptional programmability.
Through a PCIe Gen4 x16 high-speed interface, Tianshu integrates seamlessly into existing computer systems, making it suitable for a wide range of high-performance computing scenarios. It is designed to meet the computing demands of AI, complex data analytics, and other data-intensive applications.
-
Optoelectronic ComputingPioneering integration of optical computing into electronic computing to deliver general-purpose computing functions
-
New Architecture DesignLargest optical computing matrix of 128x128
-
3D TSV Advanced Packaging TechnologyHigh-efficiency integration and packaging with over 40,000 photonic devices
-
Commercial Algorithm ApplicationFirst implementation of optoelectronic hybrid computing in complex commercial models
Data Parameters
-
Heterogeneous ComputingArchitecture
-
Optical Tensor Processing UnitPIC Core
-
Electronic Tensor Acceleration Unit and Optical Computing Control UnitEIC core
-
1GHzChip Clock Speed
-
200psPIC Latency
-
INT/UINT8 INT/UINT4EIC Precision Support
-
INT4/UINT4PIC precision supports input
-
INT4/INT8/INT16PIC precision supports output
-
32 TOPsEIC Peak Performance
-
32 TOPsPIC Peak Performance
Application Scenarios
Optical computing is accelerating commercial deployment in AI for Science, biomedicine, financial engineering, materials discovery, and beyond.
-
Financial EngineeringEmpowering high- frequency trading and risk analysis
-
Electronic Design Automation (EDA)Assisting complex IC design and verification
-
AI-related TasksImage analysis, image recognition, and industrial quality inspection
