Memory Optimization of a Quantum-Inspired Processor for High-Speed Computation

The Quantum-like Accelerator for Tangled (QAT) Processor is a high-performance computational unit built to simulate quantum superposition states. It does this by processing multiple bits in parallel, similar to the working of SIMD (Single Instruction, Multiple Data) architectures, while avoiding wavefunction collapse. Designed to work alongside the SHAKTI C-class processor, the QAT improves both data transfer efficiency and overall computational throughput through custom Instruction Set Architecture extensions. One of its more pressing challenges is its large memory footprint which is largely a consequence of its extensive Array of Bits (AoB) structure. To address this challenge, bit compression techniques such as quantization have been implemented, effectively reducing memory usage from 65,536 AoB to 32,768 AoB while maintaining computational fidelity. This enables the QAT to store twice the amount of data, significantly improving efficiency in quantum-inspired computing environments. The proposed architecture paves the way for enhanced quantum-classical hybrid processing, offering a scalable and efficient solution for high-performance computing applications.