1: \begin{abstract}

2: We describe an asynchronous parallel stochastic proximal coordinate

3: descent algorithm for minimizing a composite objective function, which

4: consists of a smooth convex function added to a separable convex function.

5: In contrast to previous analyses, our model of asynchronous

6: computation accounts for the fact that components of the unknown

7: vector may be written by some cores simultaneously with being read by

8: others.  Despite the complications arising from this possibility, the

9: method achieves a linear convergence rate on functions that satisfy an

10: optimal strong convexity property and a sublinear rate ($1/k$) on

11: general convex functions. Near-linear speedup on a multicore system

12: can be expected if the number of processors is $O(n^{1/4})$. We

13: describe results from implementation on ten cores of a multicore

14: processor.

15: \end{abstract}

16: