SHA-2 Hashing Pipeline With Three-Block State Update
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SHA-2 hashing algorithms struggle with throughput limitations, particularly in high-frequency processor environments, requiring multiple cycles per round and excessive hardware resources, which is inefficient for single-message processing.
Innovation Solution
Partitioning the SHA-2 state update into three pipelined blocks with concurrent 1-cycle loops and feedback loops that send data 'back-in-time', allowing for efficient computation with minimal hardware overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing SHA-2 hashing algorithms are implemented in high-frequency processor environments, then cryptographic security is maintained, but throughput is limited due to requiring multiple cycles per round and excessive hardware resources
Solution Approach 1:
The SHA-2 state update function is partitioned into three separate blocks (first block for variables A-D, second block for variables E-H, third block for temporary word computation). Each block can be executed independently in a pipelined manner, allowing concurrent execution across multiple cycles and improving throughput without requiring excessive hardware resources.
Solution Approach 2:
The patent implements dynamic pipelining where the three blocks are executed in overlapping time periods with feedback loops that feed results back to earlier stages. The first block of cycle N feeds into the second block of cycle N+1, which feeds into the third block of cycle N+2, creating a dynamic pipeline that maintains continuous operation at 1-cycle per round performance.
2Speed
If multiple cycles per round are used in existing SHA-2 implementations, then computation accuracy is ensured, but processing speed decreases
Solution Approach 1:
The patent performs preliminary partitioning of the state update function into three blocks before execution, preparing the computational structure in advance. This allows the processor to execute each block in a single cycle with proper pipelining, eliminating the need for multiple cycles per round while maintaining computation accuracy through the structured feedback loops.
Solution Approach 2:
The pipelined execution model ensures continuous useful action by overlapping the execution of three blocks across consecutive cycles. While the first block executes in cycle N, the second block executes in cycle N+1, and the third block executes in cycle N+2, creating uninterrupted computational flow that achieves 1-cycle per round effective throughput.
3Use of energy by moving object
If existing SHA-2 implementations are deployed, then cryptographic hashing functionality is provided, but power consumption increases due to excessive hardware resources
Solution Approach 1:
By segmenting the state update into three distinct blocks that can be executed sequentially with pipelining, the patent reduces the instantaneous hardware resource requirements compared to traditional parallel implementations. This segmentation allows the same functionality to be achieved with lower power consumption while maintaining throughput through continuous pipelined execution.
Data Source
AI summary
Provided are a method, a system, a computer program product, an Application Specific Integrated Circuit, and a cryptographic module, where a partitioning is performed of a SHA-2 state update into three blocks. A set of operations associated with the three blocks are executed in a pipelined manner, where each block feeds back for computation into itself.