In the hardware design space exploration process, it is critical to optimize both hardware parameters and algorithm-to-hardware mappings. Previous work has largely approached this simultaneous optimization problem by separately exploring the hardware design space and the mapspace - both individually large and highly nonconvex spaces - independently. The resulting combinatorial explosion has created significant difficulties for optimizers. In this paper, we introduce DOSA, which consists of differentiable performance models and a gradient descent-based optimization technique to simultaneously explore both spaces and identify high-performing design points. Experimental results demonstrate that DOSA outperforms random search and Bayesian optimization by 2.80x and 12.59x, respectively, in improving DNN model energy-delay product, given a similar number of samples. We also demonstrate the modularity and flexibility of DOSA by augmenting our analytical model with a learned model, allowing us to optimize buffer sizes and mappings of a real DNN accelerator and attain a 1.82x improvement in energy-delay product.
@article{arxiv.2509.10702,
title = {DOSA: Differentiable Model-Based One-Loop Search for DNN Accelerators},
author = {Charles Hong and Qijing Huang and Grace Dinh and Mahesh Subedar and Yakun Sophia Shao},
journal= {arXiv preprint arXiv:2509.10702},
year = {2025}
}