Semiconductor Memory Random Number Generator Circuit Scale
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional random number sequences generated in a two-dimensional arrangement often suffer from correlation and bias towards 0 or 1 in the column direction, leading to poor randomness, and increasing randomness in this arrangement results in a larger circuit scale.
Innovation Solution
A semiconductor memory device with a random number generator that includes M-sequence shift registers and a coefficient selecting unit, which varies the feedback loop coefficients based on address values to reduce correlation among initial values, and a bit selecting unit to output N bits from the M-bit sequence, ensuring high randomness in both row and column directions with a reduced circuit scale.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the Hamming distance between initial values in adjacent rows is small, then the circuit scale is reduced, but the random number sequences become correlated and biased towards 0 or 1 in the column direction
Solution Approach 1:
The patent applies dynamics by making the feedback loop coefficients variable rather than fixed. The coefficient selecting unit dynamically changes coefficients based on address values, allowing the system to adapt to different positions in the two-dimensional arrangement. This dynamic adjustment ensures that initial values have sufficient Hamming distance without requiring a larger fixed circuit structure, thus maintaining randomness while controlling circuit scale.
Solution Approach 2:
The patent changes the parameter of feedback loop coefficients based on address values. By selecting different coefficients from multiple sets depending on the current address, the system achieves high randomness in the column direction while maintaining a compact circuit scale. This parameter change strategy resolves the contradiction by allowing small Hamming distance in some contexts while achieving high randomness through coefficient variation.
2Reliability
If the Hamming distance between initial values in adjacent rows is increased, then randomness in the column direction is improved, but the circuit scale increases
Solution Approach 1:
The system uses dynamic coefficient selection to achieve high randomness without proportionally increasing circuit scale. The coefficient selecting unit adapts coefficients based on address values, allowing the same hardware structure to achieve high Hamming distance effects through software-controlled parameter variation rather than hardware expansion.
Solution Approach 2:
The patent makes the feedback loop coefficients universal by providing multiple coefficient sets that can be selected based on address values. This multi-functionality allows a single circuit structure to serve multiple purposes: achieving high randomness in different directions and positions without requiring separate dedicated circuits for each function, thus avoiding circuit scale increase.
3Reliability
If multiple random number sequences are arranged in a two-dimensional manner, then randomness in the row direction is improved, but correlation and bias occur in the column direction
Solution Approach 1:
The patent applies dynamics by making the feedback loop coefficients variable rather than fixed. The coefficient selecting unit dynamically changes coefficients based on address values, allowing the system to adapt to different positions in the two-dimensional arrangement. This dynamic adjustment ensures that initial values have sufficient Hamming distance without requiring a larger fixed circuit structure, thus maintaining randomness while controlling circuit scale.
Solution Approach 2:
The patent applies local quality by making the feedback loop coefficients variable rather than fixed. The coefficient selecting unit dynamically changes coefficients based on address values, allowing the system to adapt to different positions in the two-dimensional arrangement. This dynamic adjustment ensures that initial values have sufficient Hamming distance without requiring a larger fixed circuit structure, thus maintaining randomness while controlling circuit scale.
Data Source
AI summary
According to one embodiment, semiconductor memory device and a random number generator includes A semiconductor memory device includes: a semiconductor memory 30, a random number generator 10 generating a random number sequence, and a data writing unit 20 storing data in the semiconductor memory 30 using the random number sequence. The random number generator 10 includes: a random number generating unit generating an M-bit random number sequence; a coefficient selecting unit outputs a first coefficient or a second coefficient to the random number generating unit; and a bit selecting unit which outputs the random number sequence obtained by selecting N bits from M-bit random number sequence output from the random number generating unit.