Tiny-HD: Ultra-Efficient Hyperdimensional Computing Engine for IoT Applications
Behnam Khaleghia, Hanyang Xub, Justin Morrisc and Tajana Šimunić Rosingd
CSE Department, UC San Diego, La Jolla, CA 92093, USA
abkhaleghi@ucsd.edu
bhax032@ucsd.edu
cj1morris@ucsd.edu
dtajana@ucsd.edu
ABSTRACT
Hyperdimensional computing (HD) is a new braininspired algorithm that mimics the human brain for cognitive tasks. Despite its inherent potential, the practical efficiency of HD is tied to the underlying hardware, which throttles the efficiency of HD in conventional microprocessors. In this paper, we propose tiny-HD, a light-weight dedicated HD platform that targets low power, high energy efficiency, and low latency, while being configurable to support various applications. We leverage an enhanced HD encoding that alleviates the memory requirements and also simplifies the dataflow to make tiny-HD flexible with an efficient architecture. We further augment tiny-HD by pipelining the stages and resource sharing, as well as a data layout that enables opportunistic power reduction. We compared tiny-HD in terms of area, performance, power, and energy consumption with the stateof- the-art HD platforms. tiny-HD occupies ∼0.5mm2, consumes 1.6mW standby and 9.6mW runtime power (at 400 MHz), with a 0.016 ms latency on a set of IoT benchmarks. tiny-HD consumes average per-query energy of 160 nJ, which outperforms the stateof- the-art FPGA and ASIC implementations by 95.5× and 11.2×, respectively.
