Mathieu Besançon
Mathieu Besançon
Exactly the same issue as https://github.com/JuliaSparse/SparseArrays.jl/pull/93 since the other one was closed :/ ```julia julia> using SparseArrays julia> direction = spzeros(3); direction[1] = 2 2 julia> argmin(direction) === nothing true...
```julia using LinearAlgebra, SparseArrays using BenchmarkTools julia> M = rand(500, 300); julia> S = spzeros(size(M)...); julia> @btime dot($M, $S); 521.241 μs (0 allocations: 0 bytes) ``` In comparison, even materializing...
Otherwise the problem is completely fixed
Should this be implemented in a non-type-pirating version? If packages depend on GenericSVD, they are aware of it and can call a non-pirate method for `svd!` or `svd`
Instead of keeping the list of authors as a unique string, it would make sense to at least have the option to get it as a vector of individual authors?
Similar to https://github.com/coin-or/python-mip/issues/255, this is to have an issue to track the possible integration of SCIP. All features required by python-mip should be there, and the package could directly depend...
In some cases, one would be interested in the cuts generated by the OA algorithm in addition to the solution etc, for instance if we run it in a limited...
This builds SCIP-SDP, a SCIP extension for mixed-integer semidefinite optimization