Microsector Debug Trace Architecture for Faster FPGA Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing programmable logic devices face inefficiencies in resource allocation due to sector-based allocations that are imprecise and wasteful, leading to overallocation and slower configuration times, particularly in partial reconfigurations and single-event upset detection.
Innovation Solution
Implementing a microsector architecture with reduced column counts and 1-bit data registers, combined with a micro-network-on-chip (MicroNOC) data transmission method, which facilitates high-bandwidth data transactions and direct addressing methods, enables faster configuration times and reduces power consumption by optimizing data handling processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If sector-based allocation is used in programmable logic fabric, then resource allocation is simplified, but allocation precision deteriorates leading to overallocation and waste
Solution Approach 1:
The patent divides the logic fabric into smaller microsectors instead of using large sectors. Each microsector can be independently allocated and configured, enabling fine-grained resource management. This segmentation allows precise matching of allocated resources to actual circuit design requirements, eliminating the overallocation problem inherent in sector-based approaches while maintaining allocation simplicity through standardized microsector units.
2Productivity
If sector-based allocation is used, then configuration process is faster, but resource utilization efficiency deteriorates due to overallocation
Solution Approach 1:
By segmenting the fabric into microsectors, the system achieves both fast configuration and high resource utilization. The standardized microsector units can be quickly allocated and configured, maintaining configuration speed. Simultaneously, the fine-grained control enabled by microsector segmentation allows precise resource matching, eliminating waste from overallocation and improving overall resource utilization efficiency.
Solution Approach 2:
The patent changes the fundamental allocation unit from large sectors to smaller microsectors, effectively changing the granularity parameter of resource allocation. This parameter change enables the system to achieve both fast configuration (through standardized units) and high resource utilization (through fine-grained control), resolving the contradiction between configuration speed and resource efficiency.
3Device complexity
If traditional sector architecture is used, then device complexity is lower, but configuration time increases significantly
Solution Approach 1:
The microsector architecture segments the logic fabric into smaller, standardized units that can be independently configured. This segmentation enables parallel configuration operations and reduces the overall configuration time by a factor of five or six compared to traditional sector architecture. The standardized microsector design maintains relatively low device complexity while achieving dramatic configuration time improvements.
4Adaptability or versatility
If soft logic routing is used for data transmission, then routing flexibility is maintained, but transaction speed deteriorates and power consumption increases
Solution Approach 1:
The patent introduces a micro-network-on-chip (microNOC) as an intermediary for data transmission between microsectors. The microNOC provides high-bandwidth, fast data transactions while the soft logic routing within microsectors maintains routing flexibility. This intermediary approach resolves the contradiction by offloading high-speed transmission to the microNOC while preserving adaptive routing capabilities in the logic fabric itself.
Data Source
Figure 1
Figure 2~3
Figure 4A~4B
AI summary
Systems and methods described herein may relate to data transactions involving a microsector architecture. Control circuitry may organize transactions to and from the microsector architecture to, for example, enable direct addressing transactions as well as batch transactions across multiple microsectors. A data path disposed between programmable logic circuitry of a column of microsectors and a column of row controllers may form a micro-network-on-chip used by a network-on-chip to interface with the programmable logic circuitry.