How can a sparse direct solver fail? That seems strange to me. With 1000 x 1000 even a full direct solver should take no time:
julia> A = rand(1000,1000); b = rand(1000); @time A \ b; 0.023178 seconds (15 allocations: 7.645 MB)
How can a sparse direct solver fail? That seems strange to me. With 1000 x 1000 even a full direct solver should take no time:
julia> A = rand(1000,1000); b = rand(1000); @time A \ b; 0.023178 seconds (15 allocations: 7.645 MB)