Two-Phase Arbitration for Shared Optical Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As semiconductor-based microprocessors approach physical limits, increased heat dissipation and power consumption prompt the need for alternative designs, such as parallel processing systems with multiple microprocessors, which require higher I/O bandwidth and efficient data arbitration mechanisms to manage data sharing and memory access among nodes.
Innovation Solution
A method and system for arbitration in an arbitration domain, where nodes receive arbitration requests, assign consecutive time slots based on priorities, and send data through a shared data channel using a two-phase arbitration mechanism, enabling efficient data transmission among nodes in a macro-chip architecture with optical data communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If multiple microprocessors work in parallel to surpass physical limits, then processing power is improved, but I/O bandwidth requirements increase and heat dissipation worsens
Solution Approach 1:
The patent segments the I/O communication into multiple arbitration domains, where each domain handles specific traffic patterns. This segmentation allows parallel processing systems to manage I/O bandwidth requirements efficiently by distributing traffic across multiple domains rather than overwhelming a single communication channel.
Solution Approach 2:
The patent implements periodic time-slot-based arbitration where nodes are assigned specific time slots for data transmission. This periodic action structure allows the system to manage high I/O bandwidth requirements through organized, time-divided multiplexing, reducing contention and improving overall bandwidth utilization in parallel processing systems.
2Power
If multiple microprocessors work in parallel to surpass physical limits, then processing power is improved, but heat dissipation worsens
Solution Approach 1:
The patent divides the parallel processing system into multiple arbitration domains with dedicated arbitration mechanisms. This segmentation distributes the computational and communication workload across multiple independent domains, preventing heat concentration in a single processor or communication channel, thereby managing heat dissipation in high-power parallel systems.
3Productivity
If time slots are assigned based on priorities to reduce contention, then data transmission efficiency is improved, but arbitration complexity increases
Solution Approach 1:
The patent segments the arbitration process into multiple independent arbitration domains, each with its own arbitration logic. This segmentation reduces arbitration complexity by localizing decision-making to smaller domains rather than requiring a single complex global arbitration mechanism, while still achieving efficient data transmission through priority-based time slot assignment within each domain.
Solution Approach 2:
The patent implements preliminary priority assignment and time slot allocation before actual data transmission occurs. Nodes declare their priorities and receive pre-assigned time slots in advance, which eliminates the need for complex real-time arbitration during data transmission, thereby improving data transmission efficiency while keeping arbitration complexity manageable through pre-computed schedules.
4Productivity
If consecutive time slots are assigned to sending nodes based on priorities, then bandwidth utilization is improved, but arbitration domain coordination complexity increases
Solution Approach 1:
The patent divides the system into multiple arbitration domains that operate semi-independently with their own priority-based time slot assignment mechanisms. This segmentation improves bandwidth utilization within each domain while reducing coordination complexity between domains, as each domain manages its own arbitration locally rather than requiring centralized coordination across the entire parallel processing system.
Data Source
AI summary
A method for arbitration in an arbitration domain. The method includes: receiving, by each node of a plurality of nodes in the arbitration domain, an arbitration request from each sending node of the plurality of nodes in the arbitration domain, where the plurality of nodes in the arbitration domain each use a shared data channel to send data to a set of receiving nodes; assigning, by each node in the arbitration domain, consecutive time slots to each sending node based on a plurality of priorities assigned to the plurality of nodes in the arbitration domain; for each time slot: sending, from the arbitration domain, a switch request to a receiving node designated by the sending node, where the receiving node is in the set of receiving nodes; and sending, by the sending node, data to the receiving node via the shared data channel during the time slot.