Melanson Denis, Abu Khater Mohammad, Aifer Maxwell, Donatella Kaelan, Hunter Gordon Max, Ahle Thomas, Crooks Gavin, Martinez Antonio J, Sbahi Faris, Coles Patrick J
Normal Computing Corporation, New York, NY, USA.
Nat Commun. 2025 Apr 22;16(1):3757. doi: 10.1038/s41467-025-59011-x.
Recent breakthroughs in artificial intelligence (AI) algorithms have highlighted the need for alternative computing hardware in order to truly unlock the potential for AI. Physics-based hardware, such as thermodynamic computing, has the potential to provide a fast, low-power means to accelerate AI primitives, especially generative AI and probabilistic AI. In this work, we present a small-scale thermodynamic computer, which we call the stochastic processing unit. This device is composed of RLC circuits, as unit cells, on a printed circuit board, with 8 unit cells that are all-to-all coupled via switched capacitances. It can be used for either sampling or linear algebra primitives, and we demonstrate Gaussian sampling and matrix inversion on our hardware. The latter represents a thermodynamic linear algebra experiment. We envision that this hardware, when scaled up in size, will have significant impact on accelerating various probabilistic AI applications.
人工智能（AI）算法最近的突破凸显了对替代计算硬件的需求，以便真正释放AI的潜力。基于物理的硬件，如热力学计算，有潜力提供一种快速、低功耗的方式来加速AI原语，特别是生成式AI和概率AI。在这项工作中，我们展示了一种小型热力学计算机，我们称之为随机处理单元。该设备由印刷电路板上的RLC电路作为单元组成，有8个单元通过开关电容进行全对全耦合。它可用于采样或线性代数原语，我们在硬件上演示了高斯采样和矩阵求逆。后者代表了一个热力学线性代数实验。我们设想，这种硬件在规模扩大后，将对加速各种概率AI应用产生重大影响。
