-
Notifications
You must be signed in to change notification settings - Fork 1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix use of TAO solution vector #28923
base: next
Are you sure you want to change the base?
Conversation
We were changing the internal tao parameters vector and we should not have been doing that because it changed the way the tao optimization algorithms worked. We now have two parameter vectors, _parameters for TAO to take ownership of and another called _local_parameters that is still owned by OptimizeSolve. We can now update the _local_parameters with the data from TAO and compute our objective and gradient from this. This causes some poorly designed tests to diff and time out. These tests will be cleaned up in next set of commits.
… mead algorithm This test was diff'ing because it relied on the incorrect use of nelder mead to converge
…s set extremely loose to make this solve solve in reasonable amount of time. This loose tolerance will result diffs in the solution. Nelder mead is not meant to be used with so many parameters. This problem is still being tested with gradient based methods
… PETSC errors after fixing the parameters vector. This test is optimizing a quadratic equation that converges in a single iteration. And with the fix to the parameters vector, the linesearch algorithms now work more robustly.
… the per iteration exodus output diff. The test and input files were also cleaned up
…nce with slightly different optimized values than before. closes idaholab#28922
Job Documentation, step Docs: sync website on e62b0fc wanted to post the following: View the site here This comment will be updated on new commits. |
Job Coverage, step Generate coverage on e62b0fc wanted to post the following: Framework coverage
Modules coverageOptimization
Full coverage reportsReports
This comment will be updated on new commits. |
@lindsayad Will you review this? I documented the changes to the tests in the commit messages. |
Job Apptainer moose-openmpi on e62b0fc : invalidated by @lindsayad |
Job Conda (Rocky) on e62b0fc : invalidated by @lindsayad |
Job Conda (Ubuntu) on e62b0fc : invalidated by @lindsayad |
Job Conda MOOSE (Linux) on e62b0fc : invalidated by @lindsayad |
Job Conda build (ARM Mac) on e62b0fc : invalidated by @lindsayad |
Job Conda build (Intel Mac) on e62b0fc : invalidated by @lindsayad |
Job Python 3.11 on e62b0fc : invalidated by @lindsayad |
Job Python 3.9 on e62b0fc : invalidated by @lindsayad |
Job Modules parallel on e62b0fc : invalidated by @lindsayad |
closes #28922