Word-Level Operator-to-Cell Mapping for Arithmetic Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current design optimization methods face challenges in efficiently mapping arithmetic operators to cells at a word-level, often requiring bit-level mapping that increases runtime and memory usage while compromising accuracy.
Innovation Solution
A system and method for word-level operator-to-cell mapping that utilizes datapath synthesis and technology mapping to associate arithmetic operators with cells, minimizing area and power consumption while meeting timing constraints, and optionally decomposing operators into smaller cells to reduce node count and error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bit-level mapping is used to map arithmetic operators to cells, then mapping precision is improved, but runtime and memory usage increase
Solution Approach 1:
The patent segments the mapping process into two distinct levels: word-level mapping that operates on entire words as units, and bit-level mapping that operates on individual bits. This segmentation allows the system to perform coarse mapping at the word level to reduce runtime and memory usage, then apply fine-grained bit-level mapping only where necessary to maintain precision.
Solution Approach 2:
The patent introduces a new dimension of abstraction by operating at the word level rather than strictly at the bit level. This dimensional change allows the mapping system to work with higher-level abstractions (words) that encompass multiple bits, thereby reducing the complexity and resource requirements of the mapping process while still achieving accurate results through selective refinement.
2Measurement precision
If bit-level mapping is used to map arithmetic operators to cells, then mapping precision is improved, but memory usage increases
Solution Approach 1:
The patent segments the mapping process into two distinct levels: word-level mapping that operates on entire words as units, and bit-level mapping that operates on individual bits. This segmentation allows the system to perform coarse mapping at the word level to reduce runtime and memory usage, then apply fine-grained bit-level mapping only where necessary to maintain precision.
Solution Approach 2:
The patent introduces a new dimension of abstraction by operating at the word level rather than strictly at the bit level. This dimensional change allows the mapping system to work with higher-level abstractions (words) that encompass multiple bits, thereby reducing the complexity and resource requirements of the mapping process while still achieving accurate results through selective refinement.
3Measurement precision
If operators are decomposed into smaller cells, then mapping accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the mapping process into two distinct levels: word-level mapping that operates on entire words as units, and bit-level mapping that operates on individual bits. This segmentation allows the system to perform coarse mapping at the word level to reduce runtime and memory usage, then apply fine-grained bit-level mapping only where necessary to maintain precision.
Data Source
AI summary
A mapping system, method and computer program product are provided. In use, at least one arithmetic operator is received. Further, the at least one arithmetic operator is mapped to at least one cell, at a word-level.