FoBiS.py

Appearance

OMP002 accumulation-without-reduction ​

warning enabled by default

A shared scalar is accumulated inside an OpenMP parallel region without reduction.

Why ​

s = s + x(i) in a parallel loop reads and writes the shared s from every thread: updates are lost and the result changes from run to run. A reduction(+:s) clause gives each thread a private partial sum and combines them at the end of the region; atomic is correct but serialises every update.

Reported ​

fortran
!$omp parallel do
do i = 1, n
  s = s + x(i)
end do

Preferred ​

fortran
!$omp parallel do reduction(+:s)
do i = 1, n
  s = s + x(i)
end do

Suppressing ​

For one occurrence, with the reason on the same line (or alone on the line before the statement):

fortran
! fobis: ignore[OMP002] <why this occurrence is fine>

For the whole project, ignore = OMP002 in the fobos [lint] section.

References ​

Sections of the shipped knowledge packs behind this rule:

  • openmp:ch07-data-environment#frameworks-introduced (fobis kb show openmp:ch07-data-environment#frameworks-introduced)