Maxence Gollier
Maxence Gollier
I have an issue when using `PDENLPModels.jl`. When I call ``` (rows,cols) = jac_structure(nlp) ``` on some `PDENLPModel`, the tuple `(rows,cols)` can have duplicate entries which creates bugs on some...
#106
Hello, I am trying to implement an efficient optimization solver. In order to test performance I use `ADNLPModels` models from `OptimizationProblems.jl`, I am currently trying to catch type instabilities/allocations using...
It would be useful to have an implementation of the operator $$v\mapsto Hv + \sigma v$$, where $$H$$ is some quasi-Newton operator and $$\sigma > 0$$. We should be able...
We have an `InverseLBFGSOperator` but not an `InverseLSR1` one, we should add it.
#347 @amontoison @tmigot This is where I am so far. Attached are my benchmark results... [grad_benchmarks.zip](https://github.com/user-attachments/files/20606813/grad_benchmarks.zip)
@amontoison i investigated a bit, it seems that BenchmarkCI is just broken with Julia 1.12. Why don't we use `PkgBenchmark.jl`. Waiting for your opinion and then i will try to...
https://github.com/JuliaSmoothOptimizers/LinearOperators.jl/blob/4575ca56d78d1cd21a080280af54540c384b3add/src/lbfgs.jl#L195-L197 This is highly inefficient. There are too many memory accesses being performed. We should replace with ``` @. q += bx .* data.b[k] - ax .* data.a[k] ``` In...