Lapack Exception

General Usage
1

Hello,

After a search on Google, I can’t find the meaning of this error code.

I understand that 0 means that everything is right, a negative number refers to the position of a wrong argument, but I can’t find the meaning of code 9.

Thanks,

Maxime

ERROR: LoadError: LAPACKException(9)
Stacktrace:
 [1] chklapackerror at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/lapack.jl:38 [inlined]
 [2] gesdd!(::Char, ::Array{Float64,2}) at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/lapack.jl:1576
 [3] #svdfact!#53(::Bool, ::Nothing, ::Function, ::Array{Float64,2}) at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/svd.jl:57
 [4] #svdfact! at ./<missing>:0 [inlined]
 [5] #svdfact#54(::Bool, ::Nothing, ::Function, ::Array{Float64,2}) at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/svd.jl:105
 [6] #svdfact at ./<missing>:0 [inlined]
 [7] #svd#57(::Bool, ::Nothing, ::Function, ::Array{Float64,2}) at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/svd.jl:171
 [8] svd at /Users/osx/buildbot/slave/package_osx64/build/usr/share/julia/stdlib/v0.7/LinearAlgebra/src/svd.jl:165 [inlined]
 [9] truncate_svd(::Array{Float64,2}, ::Int64) at /Users/maxime/Git/TennetOpti.jl/Sources/tennet_opti.jl:13
 [10] update(::MPS{2,10}, ::Array{Float64,4}, ::Int64) at /Users/maxime/Git/TennetOpti.jl/Sources/mps_learning.jl:90
 [11] update(::MPS{2,10}, ::Array{Features{2,10},1}, ::Array{Float64,1}, ::Float64, ::Int64) at /Users/maxime/Git/TennetOpti.jl/Sources/mps_learning.jl:82
 [12] train(::MPS{2,10}, ::Array{Features{2,10},1}, ::Array{Float64,1}, ::Float64, ::Int64) at /Users/maxime/Git/TennetOpti.jl/Sources/mps_learning.jl:103
 [13] top-level scope
 [14] include at ./boot.jl:314 [inlined]
 [15] include_relative(::Module, ::String) at ./loading.jl:1067
 [16] include(::Module, ::String) at ./sysimg.jl:29
 [17] exec_options(::Base.JLOptions) at ./client.jl:327
 [18] _start() at ./client.jl:457
1 Like
2

Apparently it stands for

DBDSDC did not converge, updating process failed.

A minimal working example (or a dump of the matrix in eg JLD2 format) would be interesting.

3

FYeah, it was indeed a convergence problem. The element of the matrix were too big due to a previous normalization problem.

However, for future work, it would be convenient to have more explicit error message when doing linear algebra.

3 Likes