More specialized routines for Triangular matrices #5255
Conversation
|
Ready to merge pending review |
| end | ||
|
|
||
| #Generic solver using naive substitution | ||
| function naivesub!(A::Triangular, b::AbstractVector, x::AbstractVector=b) |
There was a problem hiding this comment.
Why not A_ldiv_B!? I would also prefer overwriting the right hand side. It is more in line with BLAS.
There was a problem hiding this comment.
I like having the name of the algorithm in the function (which isn't exported anyway). But I can definitely add A_ldiv_B! as a synonym.
I believe this function already overwrites the r.h.s. by default, since x is passed by reference and defaults to b.
There was a problem hiding this comment.
I like having the name of the algorithm in the function
Ok, but I'd like a A_ldiv_B! synonym (and the spelling naïvesub!;-)).
I believe this function already overwrites the r.h.s.
My mistake. I missed the =b part.
There was a problem hiding this comment.
he spelling naïvesub!;-)
I once had a long argument with an American about the correct spelling of "naïve". He refused to budge, so I called him a naïf. Apparently calling people "naïf" in the US is very amusing to them.
My mistake. I missed the =b part.
=b
- Generic inverse replaced with generic solve on identity matrix - Specialized transpose, ctranspose, istril, istriu
- wraps LAPACK wrapper trcon!, reciprocal of condition number estimate for triangular matrices - type stability bugfixes to generic cond() - Triangular-specific tests for linear solver and cond()
- Wraps LAPACK.trevc!
- New eigvecs{T<:BlasFloat}(A::Triangular{T}) method
Addresses #3688 - Provides diag, generic multiplication and other simple routines - Eigensystem solvers: eigvals, eigvecs, eigfact - Provides simple wrappers for singular system solvers - Adds convert routine from Triangular to dense Matrix - Clean up bidiagonal Matrix tests
- Tests BigFloat and Complex{BigFloat}
- Tests naive substitution solver for all float types
- Wraps LAPCK.trrfs! which computes forward and backward errors
More specialized routines for Triangular matrices
This branch provides additional specialized routines for Triangular matrices
naivesub!)eigvals,eigvecs,eigfactcond{T<:BlasFloat}, condition number estimate, wrappingLAPACK.trcon!, reciprocal of condition number estimate for triangular matricestranspose,ctranspose,istril,istriuAdditionally:
zeros,ones,infsandnansmethods taking an AbstractMatrix argumentcondmethod on dense matrices