Take advantage of symmetry in hessian

@KristofferC

As the benchmarks of HyperHessians.jl suggest (@KristofferC), symmetry can probably speed up Hessians quite a bit, even though it is not the only factor. It would be nice to take advantage of it here!

You won't get this clean code

ForwardDiff.jl/src/hessian.jl

Lines 31 to 37 in 569af35

    
           function hessian!(result::AbstractArray, f::F, x::AbstractArray, cfg::HessianConfig{T} = HessianConfig(f, x), ::Val{CHK}=Val{true}()) where {F,T,CHK} 
        
               require_one_based_indexing(result, x) 
        
               CHK && checktag(T, f, x) 
        
               ∇f = y -> gradient(f, y, cfg.gradient_config, Val{false}()) 
        
               jacobian!(result, ∇f, x, cfg.jacobian_config, Val{false}()) 
        
               return result 
        
           end

but it is probably worth doing anyway.

Worst-case scenario, we hide it behind a config option so that it is non-breaking and optional

I don't see why it would be breaking.

Navigation Menu

Repository navigation

Take advantage of symmetry in `hessian` #836

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

	function hessian!(result::AbstractArray, f::F, x::AbstractArray, cfg::HessianConfig{T} = HessianConfig(f, x), ::Val{CHK}=Val{true}()) where {F,T,CHK}
	require_one_based_indexing(result, x)
	CHK && checktag(T, f, x)
	∇f = y -> gradient(f, y, cfg.gradient_config, Val{false}())
	jacobian!(result, ∇f, x, cfg.jacobian_config, Val{false}())
	return result
	end

Take advantage of symmetry in hessian #836

Description

Activity

KristofferC commented on Aug 15, 2026

gdalle commented on Aug 15, 2026

KristofferC commented on Aug 15, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Take advantage of symmetry in `hessian` #836