张量核单元的计算模型
A Computational Model for Tensor Core Units
摘要：
为了满足深度神经网络的高效训练和推理需求，已经引入了大量特定于领域的硬件架构，例如Google Tensor Processing Units和NVIDIA Tensor Cores。这些架构的共同特征是用于有效地计算给定小尺寸的密集矩阵乘法的硬件电路。为了拓宽利用这些系统的算法的类，我们提出了一个计算模型，命名为TCU模型，它捕获的能力，本机乘小矩阵。然后，我们使用TCU模型设计快速算法的几个问题，包括矩阵运算（密集和稀疏乘法，高斯消元），图形算法（传递闭包，所有对最短距离），离散傅立叶变换，模板计算，整数乘法，多项式评估。最后，我们强调TCU模型和外部存储器模型之间的关系。
英文摘要：
To respond to the need of efficient training and inference of deep neural networks, a plethora of domain-specific hardware architectures have been introduced, such as Google Tensor Processing Units and NVIDIA Tensor Cores. A common feature of these architectures is a hardware circuit for efficiently computing a dense matrix multiplication of a given small size. In order to broaden the class of algorithms that exploit these systems, we propose a computational model, named the TCU model, that captures the ability to natively multiply small matrices. We then use the TCU model for designing fast algorithms for several problems, including matrix operations (dense and sparse multiplication, Gaussian Elimination), graph algorithms (transitive closure, all pairs shortest distances), Discrete Fourier Transform, stencil computations, integer multiplication, and polynomial evaluation. We finally highlight a relation between the TCU model and the external memory model.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
