Multi-mode System on Chip Resource Reconfiguration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing systems with System on a Chip (SoC) are typically limited to specific functionalities based on their configuration, requiring additional silicon and resources to provide additional features, which restricts clients to pre-determined services without flexibility.
Innovation Solution
The SoC is configured to operate in multiple modes, allowing it to provide either networking services, compute services, or both concurrently, by repurposing and reconfiguring resources between a network compute subsystem and a server compute subsystem, managed by a management compute subsystem, using techniques such as PCIe devices and reconfigurable resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If additional functionalities are provided in a traditional SoC, then more silicon and resources are required, but this increases device complexity and costs
Solution Approach 1:
The patent implements a universal resource pool that can be dynamically allocated to different compute subsystems. The same physical resources (cores, memory, I/O) can serve multiple functions by being reassigned between network compute subsystem and server compute subsystem based on operational mode, eliminating the need for dedicated hardware for each function.
Solution Approach 2:
The patent introduces dynamic reconfiguration capability where the SoC can switch between different operational modes (network-only, server-only, or hybrid) by dynamically allocating resources from the resource pool to different compute subsystems. This dynamic allocation allows the system to adapt its functionality without physical reconfiguration.
2Productivity
If a SoC is configured for specific functionality, then performance is optimized, but adaptability to different services is reduced
Solution Approach 1:
The patent divides the SoC into dedicated compute subsystems (network compute subsystem, server compute subsystem) and a shared resource pool. Each compute subsystem can be optimized for its specific function while sharing underlying resources, allowing both specialized performance and flexible allocation.
Solution Approach 2:
The patent enables parameter changes in resource allocation by modifying the operational mode of the SoC. The management compute subsystem can adjust resource distribution parameters dynamically, changing how much of the resource pool is allocated to network versus server functions based on current needs.
3Adaptability or versatility
If resources are reallocated between compute subsystems, then flexibility is improved, but system complexity increases
Solution Approach 1:
The patent implements self-service mechanisms where the management compute subsystem automatically handles resource allocation and mode switching based on predefined criteria or external inputs. The system can autonomously reconfigure resources without requiring complex external control logic for each allocation decision.
Solution Approach 2:
The patent introduces a management compute subsystem as an intermediary between the resource pool and compute subsystems. This intermediary layer simplifies resource management by providing a centralized control point that handles allocation decisions, shielding the underlying complexity from the compute subsystems themselves.
4Reliability
If dedicated hardware is provided for each function, then reliability is improved, but silicon area and cost increase
Solution Approach 1:
The patent merges previously separate dedicated hardware resources into a unified resource pool. By combining memory, cores, and I/O resources into shared pools that serve multiple compute subsystems, the total silicon area is reduced while maintaining the reliability needed for each function through proper resource isolation and management.
Data Source
AI summary
A system on a chip (SoC) can be configured to operate in one of a plurality of modes. In a first mode, the SoC can be operated as a network compute subsystem to provide networking services only. In a second mode, the SoC can be operated as a server compute subsystem to provide compute services only. In a third mode, the SoC can be operated as a network compute subsystem and the server compute subsystem to provide both networking and compute services concurrently.