Intel has made a significant breakthrough in integrated photonics technology, demonstrating the industry’s first fully integrated optical compute interconnect chiplet co-packaged with an Intel CPU and running live data. This achievement is expected to revolutionize high-speed data processing for AI infrastructure.
According to Thomas Liljeberg, senior director of Product Management and Strategy at Intel’s Integrated Photonics Solutions Group, this technology will enable customers to seamlessly integrate co-packaged silicon photonics interconnect solutions into next-generation compute systems, boosting bandwidth, reducing power consumption, and increasing reach.
The OCI chiplet supports 64 channels of 32 gigabits per second data transmission in each direction on up to 100 meters of fiber optics, addressing AI infrastructure’s growing demands for higher bandwidth, lower power consumption, and longer reach. This technology has the potential to revolutionize high-performance AI infrastructure, enabling future scalability of CPU/GPU cluster connectivity and novel compute architectures.
Revolutionizing High-Speed Data Processing with Optical Compute Interconnect Chiplet
The ever-increasing demand for high-speed data processing in artificial intelligence (AI) infrastructure has led to a significant breakthrough in integrated photonics technology. Intel Corporation has demonstrated the industry’s first fully integrated optical compute interconnect (OCI) chiplet, co-packaged with an Intel CPU and running live data. This achievement represents a leap forward in high-bandwidth interconnect by enabling co-packaged optical input/output (I/O) in emerging AI infrastructure for data centers and high-performance computing (HPC) applications.
The OCI chiplet is designed to support up to 4 terabits per second (Tbps) bidirectional data transfer, compatible with peripheral component interconnect express (PCIe) Gen5. The live optical link demonstration showcases a transmitter (Tx) and receiver (Rx) connection between two CPU platforms over a single-mode fiber (SMF) patch cord. The CPUs generated and measured the optical Bit Error Rate (BER), and the demo showcases the Tx optical spectrum with 8 wavelengths at 200 gigahertz (GHz) spacing on a single fiber, along with a 32 Gbps Tx eye diagram illustrating strong signal quality.
The Need for Optical I/O in AI Infrastructure
The current electrical I/O solutions used in CPUs and GPUs are limited in capacity and range, much like horse-drawn carriages distributing goods. Replacing these electrical I/O solutions with optical I/O solutions is akin to using cars and trucks that can deliver much larger quantities of goods over much longer distances. This level of improved performance and energy cost is what optical I/O solutions like Intel’s OCI chiplet bring to AI scaling.
The unsustainable power requirements of AI infrastructure make it critical to adopt energy-efficient solutions. The co-packaged OCI solution consumes only 5 pico-Joules (pJ) per bit compared to pluggable optical transceiver modules at about 15 pJ/bit. This level of hyper-efficiency is crucial for data centers and high-performance computing environments.
Intel’s Leadership in Silicon Photonics
As a market leader in silicon photonics, Intel leverages more than 25 years of internal research from Intel Labs, which pioneered integrated photonics. Intel was the first company to develop and ship silicon photonics-based connectivity products with industry-leading reliability at high volume to major cloud service providers.
Intel’s main differentiator is unparalleled integration using hybrid laser-on-wafer technology and direct integration, which yield higher reliability and lower costs. This unique approach enables Intel to deliver superior performance while maintaining efficiency. Intel’s robust, high-volume platform boasts shipping over 8 million PICs with over 32 million integrated on-chip lasers, showing a laser failures-in-time (FIT) rate of less than 0.1.
The Future of Connectivity
Intel’s current OCI chiplet is a prototype, and the company is working with select customers to co-package OCI with their SoCs as an optical I/O solution. As the AI infrastructure landscape evolves, Intel remains at the forefront, driving innovation and shaping the future of connectivity.
The development of next-generation 200G/lane PICs to support emerging 800 Gbps and 1.6 Tbps applications is underway. Additionally, Intel is implementing a new silicon photonics fab process node with state-of-the-art device performance, higher density, better coupling, and vastly improved economics. Advancements in on-chip laser and semiconductor optical amplifier (SOA) performance, cost, and power are also being made.
The future of high-speed data processing looks bright with the advent of optical compute interconnect chipsets like Intel’s OCI. As AI infrastructure continues to evolve, innovations like these will play a critical role in shaping its future.
External Link: Click Here For More
