Twan Koolen
Twan Koolen
For this issue, it's not really about performance; the current timing checks are with trivial problems where I'm just making sure the `solve` call doesn't return *too quickly* when I...
I've been using `@constraint(m, v >= zeros(n))` everywhere so I hadn't run into this yet. That will always remain a valid way to do this, so it's certainly an option...
This probably also becomes easier if we switch to a Cassette-based implementation.
Could you post your `Project.toml`?
Yeah I'm a little behind in terms of MOI support. I'll try to get to this soonish.
It appears the first order of business is getting OSQP.jl to work on Julia 1.3. It doesn't seem to be passing its own tests.
> I'm not sure what exactly we can do about it, other than to keep making noise over at JuliaLang/julia#28669 Thanks for doing so and reevaluating the current state of...
Now we should remember that `FunctionWrappers.FunctionWrapper` is not actually functional on Julia 1, so the very thing that makes it fast could also be what makes it not work. Also,...
This is because `1.0^2` becomes `Base.literal_pow(^, 1.0, Val{2})`, and `Val{2}` is a `DataType` (as opposed to `Val{2}()`), so dynamic dispatch happens. The issue is fixed on Julia master, so I'm...
So the same thing also happens for Gurobi, and it looks like both OSQP.jl and LinQuadOptInterface.jl have essentially the same bug. The problem is that a difference between the current...