Release of the PriDec optimization solver

The salient features of v0.6.0 are

• the release of the primal decomposition (PriDec) solver for structured two-stage problems
• improved support for (NVIDIA) GPUs for solving sparse optimization problems via NVIDIA's cuSOLVER API and newly developed condensed optimization kernels.

Other notable capabilities include

• improved accuracy in the computations of the search directions via Krylov-based iterative refinement
• design of a matrix interface for sparse matrices in compressed sparse row format and (capable) CPU reference implementation

Elastic mode, Krylov solvers, and misc bug fixes

The release added new algorithmic features to the NLP solver(s) and associated linear algebra KKT systems

• soft feasibility restoration
• Relaxer of equality constraints at the NLP formulation level
• Krylov interfaces and implementation for CG and BiCGStab
• protype of the condensed linear system and initial Krylov-based iterative refinement
• update of the Magma solver class for the latest Magma API
• elastic mode

This release also includes several bug fixes.

xSDK compliance

Updated xSDK compliance document for HiOp.

MDS device computations, porting of sparse kernels, and HiOp-PriDec

The salient features of the major release are

• update of the interface to MAGMA and capability for running mixed dense-sparse (MDS) problems solely in the device memory space
• added interface PARDISO linear solver
• porting of the sparse linear algebra kernels to device via RAJA performance portability layer
• various optimizations and bug fixes for the RAJA-based dense linear algebra kernels
• Primal decomposition solver HiOp-PriDec available as a release candidate

MAGMA GPU and PARDISO interface

• solver class for Magma was updated to work on the GPU

• interface for PARDISO (CPU)

• bug fixes for PriDec solver

Dev workflow update

Updated development workflow on supported CI and fixed a couple of other issues.

small fixes

Fixes for a couple of bugs in RAJA linear algebra implementation and in NLP scaling

Bug fixing for sparse and MDS problems

Fixes a couple of small bugs (corner cases) in the linear algebra objects

Sparse optimization solver and enhanced support for device computations

Minor fixes missed in v0.4 release.

The salient features of v0.4 release are: [list updated in release v0.4.1]

• Development of a sparse NLP solver and associated sparse NLP interface
• Update of the mixed dense-sparse NLP solver to support full GPU compute mode
• Unit testing and documentation were expanded and consolidated
• Added partial support for device/GPU computations for the linear algebra of sparse NLPs
• Second-order corrections to the quasi- and full-Newton search directions
• Support for the relaxing and adjustment of variables and constraints bounds
• Implemented gradient-based scaling of the problem
• Least-squares initialization and computation of the duals
• Support for inertia computation and regularization for KKT linearizations systems

Sparse optimization solver and enhanced support for device computations

The salient features of this release are

• Development of a sparse NLP solver and associated sparse NLP interface
• Update of the mixed dense-sparse NLP solver to support full GPU compute mode
• Unit testing and documentation were expanded and consolidated
• Added partial support for device/GPU computations for the linear algebra of sparse NLPs
• Second-order corrections to the full-Newton search directions
• Support for the relaxing and adjustment of variables and constraints bounds
• Implemented gradient-based scaling of the problem
• Least-squares initialization and computation of the duals
• Support for inertia computation and regularization for KKT linearizations systems