JuliaLang/LinearAlgebra.jl

Linear Algebra operations with matrices of abstract element types are promoted incorrectly

Open

#287 opened on Dec 2, 2015

View on GitHub
 (2 comments) (0 reactions) (1 assignee)Julia (62 forks)github user discovery
bignumsbughelp wanted

Repository metrics

Stars
 (77 stars)
PR merge metrics
 (Avg merge 8d 17h) (5 merged PRs in 30d)

Description

Follow up to JuliaLang/LinearAlgebra.jl#285.

Ex.

julia> a = Real[1.0 big(realmax(Float64))+1
       1.0 big(1.0)]
2x2 Array{Real,2}:
 1.0  1.797693134862315708145274237317043567980705675258449965989174768031572607800285e+308
 1.0  1.000000000000000000000000000000000000000000000000000000000000000000000000000000

julia> eigfact(a)
Base.LinAlg.Eigen{Float64,Float64,Array{Float64,2},Array{Float64,1}}([1.3407807929942597e154,-1.3407807929942597e154],2x2 Array{Float64,2}:
 1.0           -1.0
 7.45834e-155   7.45834e-155)

To be correct, this would have to widen the element type in a manner similar to map.

cc: @andreasnoack

Contributor guide

Research direction
Examine the promotion logic for matrices with abstract element types. Trace the code path for eigenvalue decomposition and determine why it promotes to Float64 instead of the common type of elements.
Tech stack
julia
Domain
backend
Issue type
Bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Julia
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
65