[julia-users] Re: Modified Gram-Schmidt: getting the best out of Julia

[Ján Dolinský]

Writting line #13 as a BLAS call reduced further memory allocation. Thanks 
for all the tips (BLAS, NumericExtensions, InplaceOps, etc.).

using ArrayViews

function mgs(X)
# mgs speed project

  nobs, nvars = size(X)
  R = eye(nvars)

  for l=1:nvars-1
    v_i = view(X,:,l+1:nvars)
    s = view(X,:,l);

    R[l,l+1:nvars]=BLAS.gemv('T', 1.0/sumabs2(s), v_i, s) # --- line 13
    BLAS.ger!(-1.0, s, vec(view(R,l,l+1:nvars)), v_i) # --- line 14
  end

  X, R

end