Closed abhx7 closed 2 months ago
Hi abhx7,
I'm afraid the adjoint solver from Eilmer 4 has been depreciated: It was built as an experiment and even though it worked, it was too restrictive for general use.
A new shape optimisation code is currently being tested for Eilmer v5 that will be available in the near future, but at present it is not quite ready.
Nick,
I should mention that, if you don't have too many design variables, it is still perfectly possible to compute sensitivites using finite differences if you have to!
Alright, thank you. I was planning to use finite difference if i couldn't use the adjoint solver.
For a small number of design variables, this is likely a tractable approach.
On Fri, 28 June 2024, 7:00 pm abhx7, @.***> wrote:
Alright, thank you. I was planning to use finite difference if i couldn't use the adjoint solver.
— Reply to this email directly, view it on GitHub https://github.com/gdtk-uq/gdtk/issues/66#issuecomment-2196450969, or unsubscribe https://github.com/notifications/unsubscribe-auth/AWONKB36OJZ3EOLSDGHE353ZJUQZNAVCNFSM6AAAAABJ7UULFGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOJWGQ2TAOJWHE . You are receiving this because you are subscribed to this thread.Message ID: @.***>
How do I use install the executable for for adjoint solver to calculate shape sensitivities? It has been commented out in the Makefile.