Fixed Point to Floating Point Conversion Rounding Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for converting fixed point numbers to floating point numbers are inefficient due to the need for additional rounding logic after normalization, which adversely affects performance by increasing pipeline depth and requiring additional stages in data processing operations.
Innovation Solution
A data processing apparatus and method that determines the bit location of the most significant bit in a fixed point number and generates a rounding bit sequence before passing it through adder logic, allowing for efficient incorporation of rounding increments during the conversion process, thereby eliminating the need for a separate rounding stage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional conversion method with separate rounding stage is used, then correct rounding is achieved, but pipeline depth increases and performance deteriorates
Solution Approach 1:
The patent determines the bit location of the most significant bit and pre-calculates the rounding requirement before the normalization stage. This preliminary determination of rounding needs allows the system to prepare rounding information in advance, eliminating the need for a separate post-normalization rounding stage and thus maintaining high conversion speed while ensuring accurate rounding.
Solution Approach 2:
The patent divides the conversion process into distinct functional segments: determination logic for identifying the most significant bit location, low order bit analysis logic for determining rounding requirements, and generation logic for creating the rounding bit sequence. This segmentation allows each component to operate independently and in parallel where possible, improving overall efficiency while maintaining rounding accuracy.
2Manufacturing precision
If additional rounding logic stage is added after normalization, then rounding accuracy is improved, but device complexity increases
Solution Approach 1:
The patent merges the rounding determination function into the existing normalization pipeline by integrating determination logic and low order bit analysis logic that operate concurrently with normalization. The rounding bit sequence is generated and made available during the normalization process itself, combining what would traditionally be separate stages into a unified operation, thereby reducing device complexity while maintaining rounding accuracy.
Solution Approach 2:
By determining the rounding requirement in advance through preliminary analysis of the most significant bit location and least significant bits, the system prepares all necessary rounding information before normalization completes. This preliminary action eliminates the need for additional post-normalization rounding logic stages, reducing device complexity while ensuring accurate rounding.
3Ease of manufacture
If serial processing steps are used for conversion, then implementation simplicity is maintained, but processing time increases
Solution Approach 1:
The patent segments the conversion process into independent logical units: determination logic for most significant bit location, low order bit analysis logic for rounding determination, and generation logic for rounding bit sequence creation. These segmented components can be implemented using standard logic blocks and can operate in parallel, reducing conversion time while maintaining implementation simplicity through modular design.
Solution Approach 2:
The determination logic and low order bit analysis logic perform preliminary actions to identify rounding requirements before the full conversion process completes. This allows rounding information to be prepared in advance, enabling parallel processing paths and reducing the overall critical path time without complicating the implementation, as the preliminary logic uses straightforward bit analysis.
Data Source
AI summary
A data processing apparatus and method are provided for converting an m-bit fixed point number to a rounded floating point number having an n-bit significand, where n is less than m. The data processing apparatus comprises determination logic for determining the bit location of the most significant bit of the value expressed within the m-bit fixed point number, and low order bit analysis logic for determining from a selected number of least significant bits of the m-bit fixed point number a rounding signal indicating whether a rounding increment is required in order to generate the n-bit significand. Generation logic is then arranged in response to the rounding signal to generate a rounding bit sequence appropriate having regard to the bit location determined by the determination logic. Adder logic then adds the rounding bit sequence to the m-bit fixed point number to generate an intermediate result, whereafter normalisation logic shifts the intermediate result to generate the n-bit significand. At this point, due to the incorporation of the rounding information prior to the addition, the generated n-bit significand is correctly rounded.