Skip to content

Conversation

@tso-martin
Copy link

  1. Update value and dual calls to Plasmo 0.5.4
  2. Add example with static partition (useful if KaHyPar is not availble on the system)

@codecov-commenter
Copy link

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (2a48bef) 89.13% compared to head (fbd99bb) 89.13%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files
@@           Coverage Diff           @@
##           master       #5   +/-   ##
=======================================
  Coverage   89.13%   89.13%           
=======================================
  Files           4        4           
  Lines         313      313           
=======================================
  Hits          279      279           
  Misses         34       34           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@tso-martin: Are the value and dual functions not working as expected when given the graph as an argument in Plasmo v0.5.4? Removing the graph argument should cause issues when running on multiple threads.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I get errors for each thread, when value and dual are called with the graph argument. Example:

 nested task error: KeyError: key MOI.VariableIndex(37) not found
    Stacktrace:
      [1] getindex
        @ xxx\.julia\packages\MathOptInterface\yczX1\src\Utilities\CleverDicts.jl:164 [inlined]
      [2] getindex
        @ xxx\.julia\packages\MathOptInterface\yczX1\src\Utilities\copy\index_map.jl:64 [inlined]
      [3] get(node_pointer::Plasmo.NodePointer, attr::MathOptInterface.VariablePrimal, idx::MathOptInterface.VariableIndex)
        @ Plasmo xxx\SchwarzTryout\dev\Plasmo\src\moi_backend_node.jl:287
      [4] value(graph::OptiGraph, var::VariableRef)
        @ Plasmo xxx\SchwarzTryout\dev\Plasmo\src\optigraph.jl:514
      [5] (::SchwarzOpt.var"#25#27"{OptiGraph})(::Pair{Int64, VariableRef})
        @ SchwarzOpt .\none:0
      [6] iterate(g::Base.Generator, s::Vararg{Any})
        @ Base .\generator.jl:47 [inlined]
      [7] _all(f::Base.var"#384#386", itr::Base.Generator{Dict{Int64, VariableRef}, SchwarzOpt.var"#25#27"{OptiGraph}}, ::Colon)
        @ Base .\reduce.jl:1287
      [8] all
        @ .\reduce.jl:1283 [inlined]
      [9] Dict(kv::Base.Generator{Dict{Int64, VariableRef}, SchwarzOpt.var"#25#27"{OptiGraph}})
        @ Base .\dict.jl:111
     [10] _do_iteration(subproblem_graph::OptiGraph)
        @ SchwarzOpt xxx\SchwarzTryout\dev\SchwarzOpt\src\optimizer.jl:305
     [11] macro expansion
        @ xxx\SchwarzTryout\dev\SchwarzOpt\src\optimizer.jl:414 [inlined]
     [12] (::SchwarzOpt.var"#45#threadsfor_fun#32"{SchwarzOpt.var"#45#threadsfor_fun#31#33"{…}})(tid::Int64; onethread::Bool)
        @ SchwarzOpt .\threadingconstructs.jl:214
     [13] #45#threadsfor_fun
        @ SchwarzOpt .\threadingconstructs.jl:181 [inlined]
     [14] (::Base.Threads.var"#1#2"{SchwarzOpt.var"#45#threadsfor_fun#32"{SchwarzOpt.var"#45#threadsfor_fun#31#33"{…}}, Int64})()
        @ Base.Threads .\threadingconstructs.jl:153

Removing the graph argument worked fine in multiple threads

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Interesting, that is definitely a bug. Without the graph argument, value grabs the last solution the node was used in. So it's possible it takes the wrong one if there are multiple subgraphs that use it. I will take a look at this and get a fix out. Thanks for checking into this!

@jalving
Copy link
Member

jalving commented Sep 1, 2024

Hey @tso-martin; would you be able to split your example to a different pull request?
I think the fixes you're proposing certainly eliminate the threading issue, but it may lead to weird convergence with the algorithm.

In the meantime, I am fixing the current main branch of SchwarzOpt.jl to v0.5.4 and working on updating this package to work with Plasmo.jl v0.6+ which will provide a cleaner interface to work with.

Edit: Actually, this seems to be an issue inside the serial implementation with value and dual here (it fails on a single thread). i will take a deeper look soon.

@tso-martin
Copy link
Author

Hi @jalving, I am a bit busy at the moment. Would 10th September be okay for you?

@jalving
Copy link
Member

jalving commented Sep 8, 2024

@tso-martin no rush at all; i will also try to get to this soon.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants