I am struggling to understand the type instability of function foo(x,y,z,r) r_prime = Point3D(x,y,z) result = r - rprime end given as input x = 0.; y = 0.; z = 0.; r = Point3D(1.,1.,1.) @code_warntype foo(x,y,z,r) The output I obtain is MethodInstance for foo(::Float64, ::Float64, ::…

It is very likely that Point3D is somehow an alias for SVector, and the source of type instability is rprime (in result = r - rprime), which is some leftover from the running session (of type Any) and should be r_prime instead.

SVector and Type Stabilty

Specific Domains Modelling & Simulations

Domenico_Lahaye April 23, 2024, 12:56pm 3

const Point3D = SVector{3,Float64}

Avoiding allocations when normalizing a vector

Topic		Replies	Views
Why this program have type issues? General Usage	2	623	January 2, 2018
Revisiting type stability General Usage type-stability	5	724	August 23, 2018
Type stability question New to Julia question	6	718	February 14, 2018
Type instability - two examples: yes/no? General Usage question	4	705	March 15, 2017
Simple and fast bisection Performance	8	2852	September 13, 2018

SVector and Type Stabilty

Related topics