Parallel processing using FLoops

jling · May 11, 2022, 4:13pm

JuliaStats/Distributions.jl/blob/dd6ae8f4eac304f404b0069540a6c3bb1c667f92/src/matrix/wishart.jl#L206


      
          function _rand!(rng::AbstractRNG, d::Wishart, A::AbstractMatrix)
              if d.singular
                  axes2 = axes(A, 2)
                  r = rank(d)
                  randn!(rng, view(A, :, axes2[1:r]))
                  fill!(view(A, :, axes2[(r + 1):end]), zero(eltype(A)))
              else
                  _wishart_genA!(rng, A, d.df)
              end
              unwhiten!(d.S, A)
              A .= A * A'
          end
          
          function _wishart_genA!(rng::AbstractRNG, A::AbstractMatrix, df::Real)
              # Generate the matrix A in the Bartlett decomposition
              #
              #   A is a lower triangular matrix, with
              #
              #       A(i, j) ~ sqrt of Chisq(df - i + 1) when i == j
              #               ~ Normal()                  when i > j
              #

you might be correct.

no, as Oscar said, the C * C' is already multi-threaded because BLAS is multi-threaded, so dividing like this shouldn’t give you linear speed up

Topic		Replies	Views
Something faster than for loops General Usage	27	6494	May 8, 2019
Julia Threads vs BLAS threads Internals & Design	16	11085	July 26, 2018
For loop in function and multiplication of larger matrices, slow speed in parallel Performance performance , parallel , loops	3	1319	November 22, 2019
Elementwise multiplication of arrays across many cores General Usage parallel	5	2370	April 14, 2017
Parallelizing for loop in the computation of a gradient Performance question	19	2640	February 26, 2018

Parallel processing using FLoops

Related topics