Serial Differential Interconnect for In-Band Link Power Transitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current interconnect architectures in high-performance computing systems face challenges in meeting the increasing demand for bandwidth and power efficiency, particularly in servers and mobile devices, where existing solutions often compromise between performance and power consumption.
Innovation Solution
The development of a High Performance Interconnect (HPI) architecture that employs a layered protocol stack, including a transaction layer, link layer, and physical layer, with features such as credit-based flow control, virtual channels, and a coherence protocol to support cache-coherent systems, enabling efficient data transfer and power management across multiple processor sockets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional multi-drop buses are used for interconnect, then device complexity is reduced, but bandwidth and communication performance are insufficient for modern high-performance computing
Solution Approach 1:
The interconnect architecture is segmented into multiple independent point-to-point links instead of a single shared bus. Each link connects specific devices (e.g., processor to memory, I/O devices) with dedicated pathways, enabling parallel data transfer and eliminating bus contention, thereby increasing bandwidth without requiring complex arbitration protocols
Solution Approach 2:
The architecture transitions from a one-dimensional shared bus to a multi-dimensional hierarchical interconnect structure with multiple layers (e.g., L1 cache interconnect, L2 interconnect, memory controller interconnect). This dimensional expansion allows simultaneous communication across different levels without interference, significantly improving overall system bandwidth
2Productivity
If multiple processor sockets are added to increase computing power, then processing capability improves, but communication between sockets becomes more critical and complex
Solution Approach 1:
The interconnect architecture provides universal communication pathways that can handle multiple types of traffic (data, commands, acknowledgments) simultaneously. Virtual channels and QoS mechanisms enable the same physical infrastructure to serve diverse communication needs between multiple sockets, reducing the need for separate dedicated communication paths for each socket pair
Solution Approach 2:
Intermediate communication structures such as switches or routers are introduced between processor sockets to manage communication traffic. These intermediaries route messages efficiently between sockets, load balance communication demands, and isolate complexity at the communication infrastructure level from the processing cores
3Productivity
If interconnect bandwidth is increased to meet demand, then data transfer performance improves, but power consumption increases
Solution Approach 1:
The interconnect system dynamically adjusts operational parameters based on actual communication needs. Link width, data rate, and active participant identification are modified in real-time to match traffic patterns, enabling high bandwidth when needed while reducing power consumption during low-activity periods through adaptive rather than static high-performance configuration
Solution Approach 2:
Different segments of the interconnect are optimized locally based on their specific requirements. Critical high-bandwidth paths (e.g., between memory controllers and RAM) maintain high performance with dedicated resources, while less critical paths use reduced bandwidth configurations, balancing overall power consumption against performance requirements
4Productivity
If existing interconnect architectures are used, then device complexity is maintained at acceptable levels, but they cannot meet the increasing demand for higher bandwidth and performance
Solution Approach 1:
The interconnect architecture implements nested hierarchical levels where lower-level interconnects (e.g., on-chip connections) are contained within higher-level structures (e.g., off-chip interposers). This nesting allows standardized modules to be reused across different hierarchy levels, managing complexity through modular repetition rather than custom design at each level
Data Source
AI summary
A physical layer (PHY) is coupled to a serial, differential link that is to include a number of lanes. The PHY includes a transmitter and a receiver to be coupled to each lane of the number of lanes. The transmitter coupled to each lane is configured to embed a clock with data to be transmitted over the lane, and the PHY periodically issues a blocking link state (BLS) request to cause an agent to enter a BLS to hold off link layer flit transmission for a duration. The PHY utilizes the serial, differential link during the duration for a PHY associated task selected from a group including an in-band reset, an entry into low power state, and an entry into partial width state.