Back to Search
Start Over
HTVM: Efficient Neural Network Deployment On Heterogeneous TinyML Platforms
- Source :
- 2023 60th ACM/IEEE Design Automation Conference (DAC), San Francisco, CA, USA, 2023, pp. 1-6
- Publication Year :
- 2024
-
Abstract
- Optimal deployment of deep neural networks (DNNs) on state-of-the-art Systems-on-Chips (SoCs) is crucial for tiny machine learning (TinyML) at the edge. The complexity of these SoCs makes deployment non-trivial, as they typically contain multiple heterogeneous compute cores with limited, programmer-managed memory to optimize latency and energy efficiency. We propose HTVM - a compiler that merges TVM with DORY to maximize the utilization of heterogeneous accelerators and minimize data movements. HTVM allows deploying the MLPerf(TM) Tiny suite on DIANA, an SoC with a RISC-V CPU, and digital and analog compute-in-memory AI accelerators, at 120x improved performance over plain TVM deployment.<br />Comment: Presented at DAC2023. Open-source code is available at https://github.com/KULeuven-MICAS/htvm
Details
- Database :
- arXiv
- Journal :
- 2023 60th ACM/IEEE Design Automation Conference (DAC), San Francisco, CA, USA, 2023, pp. 1-6
- Publication Type :
- Report
- Accession number :
- edsarx.2406.07453
- Document Type :
- Working Paper
- Full Text :
- https://doi.org/10.1109/DAC56929.2023.10247664