v3.13.0

DiffEqGPU v3.13.0

Diff since v3.12.0

Merged pull requests:

• Add DAE support for GPU kernels with mass matrices and initialization (#361) (@ChrisRackauckas)
• Fix invalid compat bound (#415) (@maleadt)
• Update test/Project.toml (#416) (@ChrisRackauckas)
• Fix all CI failures: typos, Runic, test compat, callback test (#419) (@ChrisRackauckas)
• Update ModelingToolkit requirement to 11.17.0 in /test in the all-julia-packages group across 1 directory (#420) (@dependabot[bot])
• Fix format-check CI: add setup-julia step for Runic (#421) (@ChrisRackauckas)
• Migrate CUDA tests from Buildkite to GitHub Actions self-hosted runners (#422) (@ChrisRackauckas-Claude)
• Fix documentation linkcheck, example blocks, and MTK DAE GPU test (#423) (@ChrisRackauckas-Claude)
• Use OrdinaryDiffEqCore in second readme example (#424) (@asinghvi17)
• Bump codecov/codecov-action from 5 to 6 (#426) (@dependabot[bot])
• Update RecursiveArrayTools compat to support v4 (#428) (@ChrisRackauckas-Claude)

Closed issues:

• EnsembleGPUKernel fails with MTK generated problems (#375)
• EnsembleGPUArray fails on problems from MTK (#425)

v3.12.0

DiffEqGPU v3.12.0

Diff since v3.11.0

Merged pull requests:

• Skip FiniteDiff tests on OpenCL due to LLVM SPIR-V backend bug (#407) (@ChrisRackauckas-Claude)
• Fix oneAPI CI failure on Julia 1.10 due to stale deps Manifest (#408) (@ChrisRackauckas-Claude)
• Fix GPU kernel crash from SciMLLogging ENV access (#410) (@ChrisRackauckas-Claude)
• Replace BracketingNonlinearSolve with hand-written GPU-compatible ITP (#411) (@ChrisRackauckas-Claude)
• Fix GPUVern7 adaptive continuous callback interpolation issue (#413) (@ChrisRackauckas-Claude)

Closed issues:

• GPUVern7 adaptive continuous callbacks give incorrect results (error ≈ 60) (#412)

v3.11.0

DiffEqGPU v3.11.0

Diff since v3.10.0

Merged pull requests:

• Fix LinSolveGPUSplitFactorize init_cacheval for LinearSolve v3 (#404) (@ChrisRackauckas-Claude)
• Add KernelAbstractions to test dependencies (#406) (@ChrisRackauckas-Claude)

v3.10.0

DiffEqGPU v3.10.0

Diff since v3.9.0

Merged pull requests:

• Bump actions/checkout from 4 to 6 (#384) (@dependabot[bot])
• Add Julia ecosystem to Dependabot configuration (#385) (@ChrisRackauckas-Claude)
• fix interpolation for GPUVern7 and GPUVern9 (#387) (@jfeist)
• Add JET.jl static analysis tests (#388) (@ChrisRackauckas-Claude)
• Fix minimum version bounds for SciMLBase, DiffEqBase (#390) (@ChrisRackauckas-Claude)
• Switch from JuliaFormatter to Runic.jl for code formatting (#394) (@ChrisRackauckas-Claude)
• Bump actions/checkout from 4 to 6 (#395) (@dependabot[bot])
• Performance improvements: fix Julia 1.12 compat and remove duplicate k1 computation (#396) (@ChrisRackauckas-Claude)
• Add missing maxthreads method for JLBackend (#397) (@ChrisRackauckas-Claude)
• Expand JET.jl static analysis test coverage (#398) (@ChrisRackauckas-Claude)
• Fix minimum version bounds for RecursiveArrayTools, SimpleDiffEq, and StaticArrays (#399) (@ChrisRackauckas-Claude)
• Update event handling for DiffEqBase.jl #1274 API changes (#402) (@ChrisRackauckas-Claude)

Closed issues:

• Removal of determine_event_occurence in DiffEqBase v6.205.0 broke DiffEqGPU precompilation (#400)

v3.9.0

DiffEqGPU v3.9.0

Diff since v3.7.0

Merged pull requests:

• CompatHelper: bump compat for AMDGPU in [weakdeps] to 2, (keep existing compat) (#353) (@github-actions[bot])
• Remove Julia v1.9 get_extension compatibility code (#354) (@ChrisRackauckas)
• Apply JuliaFormatter to fix code formatting (#355) (@ChrisRackauckas)
• Fix 1 typos + Add spell checking configuration (#357) (@ChrisRackauckas)
• [ci skip] Fix spelling errors (#358) (@ChrisRackauckas)
• Add downgrade CI workflow (#359) (@ChrisRackauckas)
• Fix master (#362) (@ChrisRackauckas)
• Fix Julia 1.11 AbstractVectorOfArray indexing deprecation warnings (#365) (@ChrisRackauckas)
• Add CPU backend testing support via KernelAbstractions (#366) (@ChrisRackauckas)
• Fix remaining CI warnings and test failures (#367) (@ChrisRackauckas)
• Fix spelling error in determine_event_occurrence function (#368) (@ChrisRackauckas)
• Drop old versions from Project.toml compat constraints (#370) (@ChrisRackauckas)
• Apply JuliaFormatter to fix code formatting (#371) (@ChrisRackauckas)
• Add JLArrays.jl support for CPU-based array computations (#374) (@ChrisRackauckas-Claude)
• Add OpenCL.jl backend for CPU-based GPU debugging (#377) (@ChrisRackauckas-Claude)
• Resolve Metal.jl type instability for saveat and literals; Metal and AMDGPU build fix (#383) (@Ambar-13)

Closed issues:

• Tutorial fails using Metal.jl (#315)
• Add stepping to GPU integrators (#325)
• JLArrays.jl support? (#373)
• saveat option does not work with Metal (#379)

v3.7.0

DiffEqGPU v3.7.0

Diff since v3.6.0

Merged pull requests:

• Use the ImmutableODEProblem from SciMLBase (#351) (@ChrisRackauckas)

v3.6.0

DiffEqGPU v3.6.0

Diff since v3.5.0

Merged pull requests:

• Add a tutorial which showcases ModelingToolkit and SII (#349) (@ChrisRackauckas)
• Properly throw an error for non-diagonal noise in EnsembleGPUArray (#350) (@ChrisRackauckas)

Closed issues:

• Tutorial utilising MTK to highlight the utility of DiffEqGPU.jl (#330)
• Proposal: GPU-Accelerated Ensemble Simulation with Randomized Decay Rates (#340)

v3.5.0

DiffEqGPU v3.5.0

Diff since v3.4.1

Merged pull requests:

• reapply formatter (#324) (@ArnoStrouwen)
• ci: standardise workflows using SciML's reusable workflows (#327) (@thazhemadam)
• CompatHelper: bump compat for Metal in [weakdeps] to 1, (keep existing compat) (#328) (@github-actions[bot])
• CompatHelper: bump compat for AMDGPU in [weakdeps] to 0.9, (keep existing compat) (#329) (@github-actions[bot])
• CompatHelper: bump compat for Flux to 0.15 for package docs, (keep existing compat) (#333) (@github-actions[bot])
• CompatHelper: bump compat for Flux to 0.16 for package docs, (keep existing compat) (#334) (@github-actions[bot])
• CompatHelper: bump compat for oneAPI in [weakdeps] to 2, (keep existing compat) (#335) (@github-actions[bot])
• Add kernel to EnsembleGPUKernel documentaion (#337) (@murenkov)
• CompatHelper: bump compat for LinearSolve to 3, (keep existing compat) (#338) (@github-actions[bot])
• CompatHelper: bump compat for ForwardDiff to 1, (keep existing compat) (#341) (@github-actions[bot])
• CompatHelper: bump compat for ForwardDiff to 1 for package docs, (keep existing compat) (#342) (@github-actions[bot])
• GPU-Accelerated Ensemble Simulation with Randomized Decay Rates (#345) (@ParyaRoustaee)
• updated the random_decay tutorial with StaticArrays and improve benchmark and plots (#346) (@ParyaRoustaee)
• Bump all tests to LTS and latest (#347) (@ChrisRackauckas)
• Fix master tests (v1.10 and v1.11) (#348) (@ChrisRackauckas)

Closed issues:

• Docstring nitpick (#339)

v3.4.1

DiffEqGPU v3.4.1

Diff since v3.4.0

Merged pull requests:

• Updates for HIP-based AMDGPU (#322) (@jpsamaroo)

v3.4.0

DiffEqGPU v3.4.0

Diff since v3.3.0

Merged pull requests:

• Update Citation (#318) (@utkarsh530)
• CompatHelper: bump compat for RecursiveArrayTools to 3, (keep existing compat) (#319) (@github-actions[bot])
• CompatHelper: bump compat for Adapt to 4, (keep existing compat) (#320) (@github-actions[bot])
• CompatHelper: bump compat for Adapt to 4 for package docs, (keep existing compat) (#321) (@github-actions[bot])

Closed issues:

• Callbacks don't work with EnsembleCPUArray() (#316)