Accelerating HDC-CNN Hybrid Models Using Custom Instructions on RISC-V GPUs

TL;DR

This paper presents custom RISC-V GPU instructions tailored for hybrid HDC-CNN models, significantly boosting performance and energy efficiency in neural network processing.

Contribution

It introduces novel custom instructions for RISC-V GPUs to accelerate HDC-CNN hybrid models, enhancing their efficiency and programmability.

Findings

Performance improved by up to 56.2 times in microbenchmarks

Custom instructions enable efficient hybrid HDC-CNN processing

RISC-V GPUs are suitable for energy-efficient neural network acceleration

Abstract

Machine learning based on neural networks has advanced rapidly, but the high energy consumption required for training and inference remains a major challenge. Hyperdimensional Computing (HDC) offers a lightweight, brain-inspired alternative that enables high parallelism but often suffers from lower accuracy on complex visual tasks. To overcome this, hybrid accelerators combining HDC and Convolutional Neural Networks (CNNs) have been proposed, though their adoption is limited by poor generalizability and programmability. The rise of open-source RISC-V architectures has created new opportunities for domain-specific GPU design. Unlike traditional proprietary GPUs, emerging RISC-V-based GPUs provide flexible, programmable platforms suitable for custom computation models such as HDC. In this study, we design and implement custom GPU instructions optimized for HDC operations, enabling efficient processing for hybrid HDC-CNN workloads. Experimental results using four types of custom HDC instructions show a performance improvement of up to 56.2 times in microbenchmark tests, demonstrating the potential of RISC-V GPUs for energy-efficient, high-performance computing.