Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix fallback for scaling blas number with array #17739

Merged
merged 1 commit into from
Aug 2, 2016
Merged

fix fallback for scaling blas number with array #17739

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions base/linalg/dense.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,8 @@ const ASUM_CUTOFF = 32
const NRM2_CUTOFF = 32

function scale!{T<:BlasFloat}(X::Array{T}, s::T)
s == 0 && return fill!(X, zero(T))
s == 1 && return X
if length(X) < SCAL_CUTOFF
generic_scale!(X, s)
else
Expand All @@ -18,6 +20,8 @@ function scale!{T<:BlasFloat}(X::Array{T}, s::T)
X
end

scale!{T<:BlasFloat}(s::T, X::Array{T}) = scale!(X, s)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why does this method exist? really strange to be sometimes mutating the second input, but usually the first

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Perhaps to be symmetric like multiplication?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Because multiplication is not always commutative.

Copy link
Member Author

@KristofferC KristofferC Aug 1, 2016 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I also thought it was weird. I guess it is there for non commutative numbers in the general case.

Edit: Beaten to the punch

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

mutating methods need to be consistent about what they're mutating.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No they don't. We don't agree here and for those of us who use this method, it doesn't cause trouble.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's mutating the array, whichever argument that is.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added to #16772. Being explicit with 3 arguments seems like a much more predictable API when you're writing generic code that has to deal with various combinations of input types.


scale!{T<:BlasFloat}(X::Array{T}, s::Number) = scale!(X, convert(T, s))
function scale!{T<:BlasComplex}(X::Array{T}, s::Real)
R = typeof(real(zero(T)))
Expand Down