Releases · exanauts/ExaPF.jl

Update to KA 0.6 (#151) (@michel2323)
Allow to run GPU kernels in batch (#155) (@frapac)
Refactoring of tests directory (#156) (@frapac)
Remove IterativeSolvers from ExaPF (#161) (@frapac)
Add @const macro in KA's kernels (#162) (@frapac)
Weekly CI on default branch (master) (#163) (@michel2323)
Fix matpower parser on two breaking cases (#164) (@frapac)
Evaluator: fix slowdown in SlackEvaluator's Hessian (#165) (@frapac)
Change names of attributes in PolarNetworkCache (#167) (@frapac)
[CI] test (again) power flow with matpower (#170) (@frapac)
Bumping CUDA.jl to 3.x (#171) (@michel2323)
Use norm2 instead of normInf in power flow (#172) (@frapac)
Allow CUDA 2.x (#173) (@michel2323)
Use latest KA 0.6 (#174) (@michel2323)
Switch to Krylov.jl 0.7.3 (#176) (@frapac)
Restrict CUDA.jl to 3.2 due to @allowscalar change (#178) (@michel2323)
Disallow scalar indexing on GPU - part I (#181) (@frapac)
Fix type inferences in ExaPF.PowerSystem (#182) (@frapac)
hotfix: phase shift was set implicitly to 0 in the MATPOWER parser (#183) (@frapac)
[FINAL] Integrate batch Hessian in ExaPF (#185) (@frapac)
Disallow scalar indexing on GPU - part II (#186) (@frapac)
Update to KA 0.7 and CUDA.jl 3.3 (#187) (@michel2323)
Fix ProxALEvaluator (#188) (@frapac)
Fix for broken CUDA on moonshot (#189) (@michel2323)
Add script to benchmark power flow solver (#190) (@frapac)
Remove all Evaluators from ExaPF (#191) (@frapac)
Hotfix: Fix weekly GA schedule (#192) (@michel2323)
Add codecov to github actions (#193) (@frapac)
Remove deprecated functions (#195) (@frapac)
Refactor AD (#196) (@michel2323)
Move HessianLagrangian structure to ExaOpt (#197) (@frapac)
Various version bumps (#198) (@michel2323)
Another look at the line flow gradient (#201) (@michel2323)
Fix breaking [email protected] interface change (#202) (@michel2323)
Add support to CUDA.jl 3.5 (#204) (@frapac)
Moving data to ExaData artifact (#205) (@michel2323)
Artifact update with added ProxAL data (#207) (@michel2323)
Release 0.6.0 (#208) (@frapac)
[ci-skip] Setup badges in README (#209) (@frapac)

Contributors

const, michel2323, and frapac

Assets 2

19 Apr 17:29

github-actions

v0.5.0

34f3a71

v0.5.0

ExaPF v0.5.0

Diff since v0.4.0

Closed issues:

Improve the naming of the API (#35)
Non-deterministic out-of-bounds error in AD backend (#54)
Implement a new structure NewtonRaphson to store options of the algorithm (#57)
Implement line power constraints (#58)
Implement getters/setters (#59)
Trigger Julia Registrator (#70)
Implement ramping constraints for active power generation (#73)
Race condition in views with same indices on GPUs (#89)
Change indexing of control variable (#92)
Efficient transpose Jacobian vector product for reactive power generation (#95)
Julia 1.6: unexpected error during compilation of overdub (#105)
Implement reduced Hessian on GPU (#108)
[Evaluator] Add reduced Hessian for ProxALEvaluator (#120)
[Evaluator] Allow to evaluate reduced Jacobian on the GPU (#123)
Store factorization in LinearSolver.DirectSolver (#142)

Merged pull requests:

Merge back v0.4.0 (#71) (@michel2323)
Integration with ProxAL (#72) (@frapac)
Refactoring AutoDiff (#75) (@michel2323)
Add ProxALEvaluator to call ExaPF from ProxAL (#76) (@frapac)
Add TagBot Github Actions (#77) (@frapac)
Update Krylov.jl version (#78) (@frapac)
Cache physical values only in buffer in ReducedSpaceEvaluator (#82) (@frapac)
Zygote for flow_constraints (#83) (@michel2323)
Remove deprecated scripts from repo (#87) (@frapac)
Implement lineflow constraints (#91) (@frapac)
Implement reduced Hessian on CPU (#93) (@frapac)
Clean importation of external modules (#96) (@frapac)
Implement reduced Hessian for ReducedSpaceEvaluator (#97) (@frapac)
[Evaluators] Improve compile time by fixing type inference issue (#98) (@frapac)
Hessian adjoint tangent projection using ForwardDiff over adjoint residual function (#99) (@michel2323)
Implement Hessian-vector product in MOI wrapper (#100) (@frapac)
Update documentation (#101) (@frapac)
Fix the performance of the seeding by moving it onto the host (#102) (@michel2323)
Parallel CPU and improved GPU preconditioner (#103) (@michel2323)
Update benchmark script and KA everywhere (#104) (@michel2323)
Generic AutoDiff for constraints (#106) (@frapac)
Handwritten adjoints and removal of Zygote (#107) (@michel2323)
Implement adjoint! function (#109) (@frapac)
Implement Hessian of lineflow (#111) (@frapac)
polar: add generic jtprod evaluation (#112) (@frapac)
Update NEWS.md (#113) (@frapac)
Adjoint of residual, reactive power, and line flows on the GPU using KA (#114) (@michel2323)
Cleaning / Part I (#115) (@frapac)
Cleaning / Part II: move Parsers and Indexes into PowerSystem (#116) (@frapac)
Create system image (#117) (@michel2323)
Evaluator: compute reduced Hessian on GPU (#118) (@frapac)
MOI: load Hessian only if available (#121) (@frapac)
Partial seeding and extraction for Hv (#125) (@michel2323)
CI update (#127) (@michel2323)
Compute reduced Jacobian on GPU (#128) (@frapac)
Evaluator: fix scaling of gradient in ProxAlEvaluator (#129) (@frapac)
Non-allocating kernels for adjoints (#131) (@frapac)
Fix powerflow with nonzero slack's voltage angle (#132) (@frapac)
[MOI] clean Hessian in MathOptInterface wrapper (#133) (@frapac)
Update documentation (#134) (@frapac)
[ci skip] update github action file (#135) (@frapac)
Update quickstart guide (#140) (@frapac)
[ci skip] add CONTRIBUTING.md (#141) (@frapac)
Store factorization directly inside DirectSolver (#144) (@frapac)
Heisenbug (#145) (@michel2323)
Update API of ExaPF before release 0.5.0 (#146) (@frapac)
Improve performance of reduced Hessian on GPU (#147) (@frapac)
[ci-skip] add quickstart in README file (#148) (@frapac)
Minor documentation updates (#149) (@michel2323)
evaluator: fix MOI wrapper for ProxALEvaluator (#150) (@frapac)
Add benchmark documentation (#152) (@michel2323)
Release 0.5 (#153) (@michel2323)