ExaModels v0.9.2

Diff since v0.9.1

Merged pull requests:

• [documentation] Add a paragraph about other GPU backends (#183) (@amontoison)
• CompatHelper: bump compat for NLPModelsIpopt in [weakdeps] to 0.11, (keep existing compat) (#186) (@github-actions[bot])
• Update compat entries in Project.toml (#187) (@amontoison)

Closed issues:

• [documentation] Explain how to use ExaModels.jl on AMD and Intel GPUs (#182)
• Precompiling a model (#184)

ExaModels v0.9.1

Diff since v0.9.0

Merged pull requests:

• Register docstring fixed (#175) (@sshin23)
• Fix nesting structs in generators. (#176) (@hfytr)
• Slight performance deterioration fixed (#177) (@sshin23)
• iterate defined for parindex as well (#178) (@sshin23)
• Bump ExaModels version to 0.9.1 (#180) (@sshin23)
• Potential fix for intel CI break (#181) (@sshin23)

Closed issues:

• Minor doc fix (bivariate function registration) (#174)

ExaModels v0.9.0

Diff since v0.8.3

There are no breaking changes

New Features

• Support for struct data type (#160)
• OpenCL support with PoCL tests (#163)
• Format .jl files (#165)

Maintenance

• Update MOI extension (#155)

CI Improvement

• registry update added to ci (#162)

Bug Fixes

• Fix parameter-only gradients (#158)

Documentation

• lvar/uvar and lcon/ucon explained (#166)
• README simd explanation edit (#164)

Merged pull requests:

• Update MOI extension (#155) (@klamike)
• GPU dependencies reduced (#157) (@sshin23)
• Fix parameter-only gradients (#158) (@klamike)
• CompatHelper: bump compat for AMDGPU in [weakdeps] to 2, (keep existing compat) (#159) (@github-actions[bot])
• Support for struct data type (#160) (@sshin23)
• Registry update added to ci (#162) (@sshin23)
• OpenCL support with PoCL tests (#163) (@klamike)
• README simd explanation edit (#164) (@sshin23)
• [Documentation] lvar/uvar and lcon/ucon explained (#166) (@sshin23)
• CompatHelper: add new compat entry for OpenCL in [weakdeps] at version 0.10, (keep existing compat) (#168) (@github-actions[bot])
• Bump ExaModels.jl version to v0.9 (#171) (@sshin23)

Closed issues:

• document constraint bounds (#99)
• Unsafe use of MOI (#100)
• Documentation nits (#101)
• Extracting solution of decision variables (#128)
• ExaModels optimizer API for JuMP does not solve the correct model (#130)
• OpenCL support (#161)

ExaModels v0.8.3

Diff since v0.8.2

Merged pull requests:

• Add parameters (#148) (@klamike)
• CUDA group added to ci runners (#149) (@sshin23)
• Self-hosted runners added (#150) (@sshin23)
• Drop buildkite CI (#152) (@sshin23)

Closed issues:

• Add support for model parameters in ExaModels (#147)

ExaModels v0.8.2

Diff since v0.8.1

Merged pull requests:

• minor fix (?) in quadrotor example (#137) (@jbcaillau)
• GPU Implementation of hprod! for unconstrained (#140) (@sshin23)
• Empty sum handling for CPU (#142) (@sshin23)
• Bump ExaModels version to v0.8.2 (#144) (@sshin23)

Closed issues:

• Unconstrained hprod! fails with CuArray (#139)
• MadNLP fails to handle constraint over an empty set when running on CPU (#141)

ExaModels v0.8.1

Diff since v0.8.0

Merged pull requests:

• Use Type instead of DataType (#124) (@blegat)
• CompatHelper: bump compat for oneAPI in [weakdeps] to 2, (keep existing compat) (#125) (@github-actions[bot])
• Update CITATION.bib (#129) (@amontoison)
• Add a method hess_coord! for the Hessian of the objective (#135) (@amontoison)

Closed issues:

• The CI build for the documentation is broken (#123)

ExaModels v0.8.0

Diff since v0.7.1

Merged pull requests:

• CompatHelper: add new compat entry for JuMP in [weakdeps] at version 1, (keep existing compat) (#94) (@github-actions[bot])
• Pass option prod to ExaModel when instantiating from JuMP (#97) (@frapac)
• Add constraint docs (#98) (@jerryfung110)
• register asinh and acosh in default lib (#108) (@FHoltorf)
• Fix interface with NLPModels (#112) (@frapac)
• [buildkite] Add CI builds on GPUs (#113) (@amontoison)
• Bump GPU pacakge versions (#114) (@sshin23)
• Add constraint docs (#98) (#115) (@sshin23)
• Resolves ambiguity in ExaCore(::Type{Float64}, ::CPU) (#116) (@sshin23)
• Remove need for specialized GPU functions (#117) (@sshin23)
• Add supports for fixed indexing (#119) (@sshin23)
• Bump ExaModels version to v0.8.0 (#120) (@sshin23)

Closed issues:

• Running into a mysterious zero div error (#75)
• oneAPI reduce workaround (#83)
• ExaCore(::Type{Float64}, ::CPU) is ambiguous (#103)
• Doc examples: assertion error on nnzj? (#104)
• constraint! cannot accept fixed index (#110)
• Ambiguous method for jtprod! (#111)

ExaModels v0.7.1

Diff since v0.7.0

Merged pull requests:

• MadNLP-related documentation update (#76) (@sshin23)
• Better handling of empty case (#77) (@sshin23)
• Docs update with reexporting MadNLP (#78) (@sshin23)
• oneAPI 2-d generator bug fix (#80) (@sshin23)
• Bump AMDGPU compat versions (#81) (@sshin23)
• oneAPI sum workaround (#82) (@sshin23)
• GPU support for CompressedNLPModels (#88) (@sshin23)
• Remove need for resize! in CompressedNLPModel (#90) (@sshin23)
• bump ExaModels version to 0.7.1 (#92) (@sshin23)

Closed issues:

• convert_data Float32 support and missing documantation (#86)

ExaModels v0.7.0

Diff since v0.6.0

Merged pull requests:

• Bump ExaModels version to v0.6.0 (#64) (@sshin23)
• Benchmark and Examples moved (#65) (@sshin23)
• Add support for maximization (#66) (@sshin23)
• Better explain objective syntax (#67) (@sshin23)
• Add support for multi-indexed constraints (#69) (@sshin23)
• Add support for single variable, constraint, objective (#70) (@sshin23)
• Variable bound check (#72) (@sshin23)
• Bump ExaModels version to v0.7 (#74) (@sshin23)

Closed issues:

• Indexing Variable Out of Bounds (#32)
• Better single objective/constraint API (#34)
• Better explain what objective does (#35)
• Double/tripple indexing for constraints (#48)
• [ExaModelsJuMP] Wrapper is broken if JuMP's objective sense is set to Max (#56)

ExaModels v0.6.0

Diff since v0.5.0

Merged pull requests:

• Update docs/make.jl (#55) (@sshin23)
• Full Optimizer approach for MOI interface (#57) (@sshin23)
• MadNLP extension hot fix (#58) (@sshin23)
• MOI interface hot fix (#60) (@sshin23)
• JuMP Interface documentation update (#62) (@sshin23)
• Repo URL change (#63) (@sshin23)