v1.0.0

ParallelParticleSwarms v1.0.0

Diff since v0.1.0

Breaking changes

• No major API change, but major changes to underlying dependencies

Merged pull requests:

• Apply JuliaFormatter to fix code formatting (#53) (@ChrisRackauckas)
• Add downgrade CI workflow (#55) (@ChrisRackauckas)
• Apply JuliaFormatter to fix code formatting (#56) (@ChrisRackauckas)
• Migrate from CompatHelper to Dependabot (#57) (@ChrisRackauckas-Claude)
• Bump julia-actions/setup-julia from 1 to 2 (#59) (@dependabot[bot])
• Bump codecov/codecov-action from 3 to 5 (#60) (@dependabot[bot])
• Bump julia-actions/cache from 1 to 2 (#61) (@dependabot[bot])
• Bump actions/checkout from 3 to 6 (#62) (@dependabot[bot])
• Fix dependency compat bounds for Julia 1.12 (#65) (@ChrisRackauckas-Claude)
• Bump the all-julia-packages group across 2 directories with 7 updates (#68) (@dependabot[bot])
• Documentation improvements: fix badge URL, docstrings, and example (#70) (@ChrisRackauckas-Claude)
• Fix precompilation and improve startup time (#71) (@ChrisRackauckas-Claude)
• Improve static analysis and fix bugs found by JET.jl (#74) (@ChrisRackauckas-Claude)
• Switch from JuliaFormatter to Runic.jl for code formatting (#75) (@ChrisRackauckas-Claude)
• NeuralODE benchmarks: add global optimization comparison (#76) (@AdityaPandeyCN)
• Bump actions/checkout from 4 to 6 (#77) (@dependabot[bot])
• Fix package version compatibility by removing test/Project.toml (#78) (@ChrisRackauckas-Claude)
• Fix minimum version bounds for DiffEqGPU (#80) (@ChrisRackauckas-Claude)
• Add test/Project.toml with dependencies and compatibility (#82) (@utkarsh530)
• Fix spinlock release mechanism (#83) (@AdityaPandeyCN)

Closed issues:

• Version bumps (#66)
• Version bump to 0.2.0 needed for release (#72)

v0.1.0

ParallelParticleSwarms v0.1.0

Merged pull requests:

• Implement basic GPU Particle Swarm (#1) (@utkarsh530)
• Add asynchronous particle swarm algorithm (#8) (@utkarsh530)
• [WIP]Some restructuring and API changes (#9) (@Vaibhavdixit02)
• [WIP] Basic non-bounded case initialization (#12) (@Vaibhavdixit02)
• Start implementation of sync PSO on GPU (#17) (@utkarsh530)
• Add parallel ODE parameter estimation with PSO (#18) (@utkarsh530)
• Add examples for ODE parameter estimation (#19) (@utkarsh530)
• Add basic iris NN example (#23) (@Vaibhavdixit02)
• Fix GPU parallelization (#24) (@utkarsh530)
• Initialize unify and use LHS (#26) (@Vaibhavdixit02)
• Switch to KernelAbstractions (#28) (@jpsamaroo)
• Use penalty method for constrained optimization (#33) (@Vaibhavdixit02)
• Towards cleaner and modular internals in the code (#34) (@utkarsh530)
• lbfgs (#35) (@Vaibhavdixit02)
• Append solve time in solution interface (#36) (@utkarsh530)
• ParallelPSOKernel: Do proper locking (#37) (@jpsamaroo)
• Update Hybrid API and misc cleanups (#38) (@utkarsh530)
• Update ParallelSyncPSO to use Shared Memory (#39) (@utkarsh530)
• [WIP] Add benchmark scripts (#40) (@utkarsh530)
• Fix queue lock algorithm for ParallelPSOKernel (#42) (@utkarsh530)
• Update solvers to run CPUs as well (#43) (@utkarsh530)
• [WIP] Add more benchmarks (#44) (@utkarsh530)
• Start towards caching and perf optimizations (#45) (@utkarsh530)
• Prepare the package for release (#50) (@utkarsh530)
• Rename PSOGPU->ParallelParticleSwarms (#51) (@utkarsh530)

Closed issues:

• Asynchronous Particle Swarm (#2)
• Resolve race condition in synchronous Particle Swarm (#3)
• Parallel-reduce Particle Swarm (#5)
• Connections of PSO with Multi-start optimization algorithms (#7)
• Use extensions for different backends (#10)
• Implement initialization of particles when upper and lower bounds are not given (#11)
• Explore quasi-monte carlo initialization (#13)