Conditional Modular Subtraction Instruction for Faster HE Arithmetic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Homomorphic encryption (HE) operations, particularly polynomial multiplication in finite fields, are computationally expensive and pose a bottleneck in privacy-preserving machine learning (PPML) due to inefficient polynomial multiplication and modular operations like NTT, which are not adequately optimized.
Innovation Solution
Introduce a set of instructions, including the vpcondsubuq instruction, to optimize forward and inverse NTT operations and element-wise modular multiplication by performing conditional subtraction, reducing the number of instructions required for HE operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional instructions are used for HE operations, then computational correctness is maintained, but computational expense and execution time increase significantly
Solution Approach 1:
The patent combines multiple sequential instructions into a single conditional modular subtraction instruction. Specifically, it merges comparison operations, conditional branching, and subtraction operations into one atomic instruction that performs modular subtraction only when needed, thereby reducing the total instruction count and improving execution speed while maintaining computational correctness
Solution Approach 2:
The patent introduces a new instruction parameter type (the conditional modular subtraction instruction) that changes the operational parameters of the CPU instruction set. This new instruction type enables direct modular arithmetic operations without requiring multiple separate instructions, thereby optimizing the execution parameters for HE workloads
2Loss of time
If multiple sequential instructions are used for conditional subtraction, then operational flexibility is maintained, but execution time increases
Solution Approach 1:
The patent merges comparison, branching, and subtraction operations into a single atomic conditional modular subtraction instruction. This eliminates the need for multiple sequential instructions while maintaining operational flexibility, thereby reducing execution time without sacrificing ease of operation
Solution Approach 2:
The conditional modular subtraction instruction is designed to automatically determine whether subtraction is needed based on the operands' values. The instruction internally performs the comparison and conditional logic, making the system self-determining about when modular subtraction is required, thereby reducing execution time without requiring external control complexity
Data Source
AI summary
One embodiment provides a processor comprising first circuitry to decode an instruction into a decoded instruction, the instruction to indicate a first source operand and a second source operand and second circuitry including a processing resource to execute the decoded instruction, wherein responsive to the decoded instruction, the processing resource is to output a result of first source operand data minus second source operand data in response to a determination by the processing resource that the first source operand data is greater than or equal to the second source operand data, otherwise the processing resource is to output the first source operand data.