1: \begin{abstract}

2:     We introduce the GEneralized Newton Inexact Operator Splitting solver (\method{}) for large-scale convex optimization.

3:     \method{} speeds up ADMM by approximately solving approximate subproblems: 

4:     it uses a second-order approximation to the most challenging ADMM subproblem and solves it inexactly with a fast randomized solver.

5:     Despite these approximations, \method{} retains the convergence rate of classic ADMM and can detect primal and dual infeasibility from the algorithm iterates.

6:     At each iteration, the algorithm solves a positive-definite linear system that arises from a second-order approximation of the first subproblem and computes an approximate proximal operator.

7:     \method{} solves the linear system using an indirect solver with a randomized preconditioner, 

8:     making it particularly useful for large-scale problems with dense data.

9:     Our high-performance open-source implementation in Julia allows users to specify convex optimization problems directly 

10:     (with or without conic reformulation) and allows extensive customization.

11:     We illustrate \method{}'s performance on a variety of problem types.

12:     Notably, \method{} is more than ten times faster than existing solvers on large-scale, dense problems.

13:     \ifsubmit

14:         \keywords{Convex optimization \and Operator splitting \and Inexact ADMM \and Large-scale optimization \and Julia}

15:         \subclass{90C25 \and 90C06 \and 90C90 \and 65K05 \and 65K10}

16:     \fi

17: \end{abstract}

18: