Datapath Placement Using Tiered Assignment for Hybrid ICs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional automated placement tools underperform in hybrid integrated circuit designs containing both datapath and random logic, leading to poor datapath alignment and increased congestion, as they fail to account for the distinct formulation of datapath logic and random logic.
Innovation Solution
A method is introduced that identifies datapath clusters using machine-learning based datapath extraction, determines datapath width, and assigns cells to tiers based on connectivity, optimizing cell placement within each tier to improve alignment and mitigate congestion through a tiered placement approach.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional automated placement tools are used for hybrid integrated circuit designs, then the placement process is automated and efficient, but datapath alignment deteriorates and congestion increases
Solution Approach 1:
The patent segments the placement process into two distinct phases: datapath placement and random logic placement. The datapath placement tool specifically handles datapath cells with constrained positioning to ensure precise alignment, while the random logic placement tool handles remaining cells. This segmentation resolves the contradiction by allowing automated efficiency overall while achieving precision where needed.
Solution Approach 2:
The patent applies local quality by implementing different placement strategies for different regions of the circuit. Datapath regions receive specialized placement treatment with alignment constraints and dedicated tooling, while random logic regions use standard automated placement. This localized approach maintains high productivity overall while achieving superior datapath alignment in critical areas.
2Productivity
If conventional automated placement tools are used for hybrid integrated circuit designs, then the placement process is automated and efficient, but congestion increases
Solution Approach 1:
The patent segments the placement process to handle congestion separately in different regions. By placing datapath cells first with dedicated constraints and then placing random logic cells in remaining spaces, the tool prevents congestion in datapath regions while maintaining overall automation efficiency.
Solution Approach 2:
The patent performs preliminary action by placing datapath cells before random logic cells. This sequencing allows the datapath placement tool to establish critical pathways first, preventing congestion before random logic placement occurs. The preliminary placement of datapath cells creates a framework that guides subsequent random logic placement to avoid congested areas.
3Adaptability or versatility
If datapath cells are allowed to move freely during placement, then placement flexibility is maximized, but datapath alignment deteriorates
Solution Approach 1:
The patent applies local quality by imposing different mobility constraints on different cell types. Datapath cells are constrained to specific regions or positions to ensure alignment, while random logic cells maintain full mobility for optimal placement. This localized constraint strategy preserves flexibility where needed while achieving precision where required.
Solution Approach 2:
The patent segments the cell population into datapath cells and random logic cells, applying different placement rules to each segment. Datapath cells are subject to alignment constraints that limit their movement freedom, while random logic cells enjoy unrestricted movement. This segmentation resolves the contradiction by applying flexibility and precision selectively to different cell types.
Data Source
AI summary
Datapath placement defines tiers for placement sets of a cell cluster, assigns cells to the tiers constrained by the datapath width, and then orders cells within each tier. Clusters are identified using machine-learning based datapath extraction. Datapath width is determined by computing a size of a bounding box for cells in the cluster. Placement sets are identified using a breadth-first search beginning with input cells for the cluster. Tiers are initially defined using logic depth assignment. A cell may be assigned to a tier by pulling the cell from the next higher tier to fill an empty location or by pushing an excess cell into the next higher tier. Cells are ordered within each tier using greedy cell assignment according to a wirelength cost function. The datapath placement can be part of an iterative process which applies spreading constraints to the cluster based on computed congestion information.