Skip to content

norm performance optimization #351

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

Open
lkdvos wants to merge 2 commits into
base: main
Choose a base branch
from
+9 −2
Open

norm performance optimization #351

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8b788b7
specialize UniqueFusion norm
lkdvos Jan 15, 2026
9a1807b
replace `sum` with manual loop because less allocations for some reason
lkdvos Jan 15, 2026
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
11 changes: 9 additions & 2 deletions src/tensors/linalg.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -269,15 +269,22 @@ function _norm(blockiter, p::Real, init::Real)
return isempty(b) ? init : oftype(init, LinearAlgebra.normInf(b))
end
elseif p > 0 # finite positive p
np = sum(blockiter; init) do (c, b)
return oftype(init, dim(c) * norm(b, p)^p)
np = init
for (c, b) in blockiter
np += dim(c) * norm(b, p)^p
end
return np^(inv(oftype(np, p)))
else
msg = "Norm with non-positive p is not defined for `AbstractTensorMap`"
throw(ArgumentError(msg))
end
end
function LinearAlgebra.norm(t::TensorMap, p::Real = 2)
InnerProductStyle(t) === EuclideanInnerProduct() || throw_invalid_innerproduct(:norm)
# performance specialization:
FusionStyle(sectortype(t)) isa UniqueFusion && return norm(t.data, p)
return _norm(blocks(t), p, float(zero(real(scalartype(t)))))
end

_default_rtol(t) = eps(real(float(scalartype(t)))) * min(dim(domain(t)), dim(codomain(t)))

Expand Down
Loading