1: \begin{abstract}

2: 	Asynchronous optimization algorithms have seen great success in

3: 	solving large-scale real-world problems. However, most of their convergence results under total asynchronism are either asymptotic or absent. In this paper, we derive explicit convergence rates for two asynchronous optimization algorithms, proximal incremental gradient (PIAG) and asynchronous block-coordinate descent (Async-BCD) methods, under a specific total asynchrony model. We show that the derived rates are optimal in terms of order for both methods. The optimal rates provide an insightful understanding of how growing of unbounded delays deteriorates the convergence rate of asynchronous optimization algorithms. Our theoretical findings are demonstrated by a numerical example.

4: 	\end{abstract}

5: