Computational Storage Device Fixed Point Data Bitwise Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computational storage devices (CSDs) face challenges in efficiently managing data storage based on varying computational precision requirements for different forms of media, such as audio, video, and image processing, which demand different levels of processing power, performance, and latency.
Innovation Solution
A method and apparatus for a CSD that includes a processor configured to obtain fixed point data, determine computational precision requirements, separate the data into two groups of bits (most significant bits (MSB) and least significant bits (LSB), and store them separately in a non-volatile memory array, allowing for flexible processing based on precision needs, including low precision processing when sufficient and full precision when required.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full precision fixed point data is always stored and processed, then processing accuracy is maintained, but bandwidth and resource usage increase
Solution Approach 1:
The patent dynamically changes the precision parameter of fixed point data based on computational requirements. The processor determines whether full precision or reduced precision is sufficient for a given computation, and adjusts the data representation accordingly. This resolves the contradiction by allowing the system to maintain full precision when needed for accuracy while using reduced precision to conserve bandwidth and resources when full accuracy is not required.
2Adaptability or versatility
If data is separated into MSB and LSB groups and stored separately, then flexible processing is enabled, but device complexity increases
Solution Approach 1:
The patent segments fixed point data into two separate groups: most significant bits (MSB) and least significant bits (LSB). These segmented data groups are stored separately in the non-volatile memory array, enabling flexible processing where only the necessary precision level is retrieved and processed. This segmentation approach resolves the contradiction by providing processing flexibility through selective data retrieval while managing complexity through an organized storage structure.
Data Source
AI summary
A computational storage device (CSD) is provided with a processor that obtains fixed point data having an initial precision (e.g., 32 bits per word) and determines a computation precision requirement for the fixed point data (such as a requirement for regular precision processing as opposed to low precision processing). The processor separates the fixed point data, based on the computational precision requirement, into a first group of bits, e.g., the most significant bits, and a second group of bits, e.g., the least significant bits, then separately stores the first the second groups of bits in the NVM array so that the different groups of bits can be fetched and managed separately. In this manner, bitwise grouping of fixed point data may be exploited to facilitate low precision processing when it is sufficient, while also accommodating full or regular precision processing when needed. Various methods are also described.