Crate srsadmm_core 

srsadmm-core

This library provides the core functionality for the srsadmm project, including matrix operations, problem formulation, and resource management. It is meant for usage with the tokio runtime and the srsadmm-lambda-mm AWS Lambda function. For more information, please see the GitHub repository: https://github.com/buk0vec/srsadmm.

Functionality

• Local and remote state management
• Matrix operations
• LASSO-specific operations
• ADMM problem formulation and solving
• Timing and logging

Features

• accelerate - Use the accelerate backend for matrix operations
• netlib - Use the netlib backend for matrix operations
• openblas - Use the openblas backend for matrix operations
• linfa - Adds a utility function to compute the optimal objective value for Lasso regression using the linfa and linfa-elasticnet libraries. Useful for testing and validation.
• rayon - Adds support for parallelization using the rayon library. Really not necessary for the ADMM algorithm, but speeds up the problem instance generation.

Modules

ops

Matrix operations on distributed variables

problem

Problem formulation and solving

resource

Resource management for distributed variables

subproblem

Utility functions for splitting and combining distributed variables

timing

Timing and logging utilities

utils

Utility functions for LASSO problems

variable

Distributed variable management