OffTarget

Observed issues › fuzz/b1-121-afar

Appears to behave incorrectlyWrong answer · Fortran · OpenMP · Open

amdflang: teams reduction(iand: r) with firstprivate and an empty body sets r to 0

Legality: independently confirmed. 2 independent readings find the program legal and the result required. How this is checked

Affected variants

CompilerMI300A gfx942
unified memory (XNACK 1)
MI300A gfx942
copy (XNACK 0)
MI250X gfx90a
copy (XNACK 0)
MI300X gfx942
copy (XNACK 0)
MI350X gfx950
copy (XNACK 0)
A100 sm_80
separate memory
A100 sm_80
managed memory
AFAR 24.3.0behaves incorrectlybehaves incorrectlyunknown
not run
behaves incorrectlybehaves incorrectlyn/an/a
CCE 19.0.0correctcorrectunknown
not run
unknown
not run
unknown
not run
n/an/a
CCE 20.0.2correctcorrectunknown
not run
unknown
not run
unknown
not run
n/an/a
CCE 21.0.3correctcorrectunknown
not run
unknown
not run
unknown
not run
n/an/a
CCE 21.0.4correctcorrectunknown
not run
unknown
not run
unknown
not run
n/an/a
NVHPC 25.11n/an/an/an/an/acorrectcorrect
NVHPC 26.9n/an/an/an/an/acorrectcorrect
GCC 16.2.0correctcorrectcorrectcorrectcorrectcorrectn/a

"behaves incorrectly" / "not supported": this issue, confirmed in up to four environments (another build directory, another node, a repeat). "fails another way": a different failure, not counted as this issue. "unknown": not run there, or the runs disagree.

Messages

ResultCompiler and GPUMessage
behaves incorrectlyAFAR 24.3.0, MI300A unified
AFAR 24.3.0, MI300A copy
AFAR 24.3.0, MI300X
AFAR 24.3.0, MI350X
show
RESULT checksum=0

Reproduce

AFAR 24.3.0, MI300A unified

$AFAR/bin/amdflang -fopenmp --offload-arch=gfx942 -O2 repro.f90 -o repro
HSA_XNACK=1 OMP_TARGET_OFFLOAD=mandatory ./repro

CCE 19.0.0, MI300A unified (runs correctly)

module load cce/19.0.0 craype-accel-amd-gfx942 rocm/6.4.3
module unload cray-mpich
ftn -fopenmp -O2 repro.f90 -o repro
HSA_XNACK=1 OMP_TARGET_OFFLOAD=mandatory ./repro
Program (43 lines)
! Minimized by compliance.fuzz_min from fuzz seed 121 (fuzz-3).
! toolchain: afar (-O2, HSA_XNACK=1, OMP_TARGET_OFFLOAD=mandatory)
! failure: wrong_code
! expected: the host reference checksum (host reference (cce19) checksum=1048575)
! observed: device checksum=0 ran_on_host=0
! Legal OpenMP 5.2: offload.engine.rules accepts it (offload.engine.prerun.check, rerun: with the
!   race rule and the restriction checks); nvfortran/nvc 26.9 and 25.11 (NVHPC, not under test
!   here) compile it; the check's other witness, the AFAR flang front end, is the compiler under
!   test and is not counted; no gfortran or gcc is a legality witness.
module fuzz_mod
  implicit none
contains
end module fuzz_mod
program fuzz_prog
  use omp_lib, only: omp_is_initial_device
  implicit none
  integer, parameter :: n = 24, m = 3
  integer :: i, j, p1, r1a
  integer(8) :: cs
  p1 = -4
  r1a = 1048575
  ! fuzz: kernels
  !$omp target map(tofrom: r1a)
    !$omp teams reduction(iand: r1a) firstprivate(p1)
    !$omp end teams
  !$omp end target
  ! fuzz: checksum
  cs = 0
  cs = modulo(cs * 31_8 + int(r1a, 8), 2147483629_8)
  print '(a,i0,a,i0)', 'RESULT checksum=', cs, ' ran_on_host=', 0
end program fuzz_prog
! spec (requirement): each private copy of a reduction list item is initialized with the initializer value of the reduction
!   identifier (for iand, all bits on), and at the end of the region the original list item is combined with the private copies,
!   so an empty teams region leaves the original value unchanged.
! "The reduction clause is a reduction scoping clause and a reduction participating clause, as described in Section 5.5.6 and
!   Section 5.5.7. For each list item, a private copy is created for each implicit task or SIMD lane and is initialized with the
!   initializer value of the reduction-identifier. After the end of the region, the original list item is updated with the values
!   of the private copies using the combiner associated with the reduction-identifier." (5.2 §5.5.8 p.134 l.20-24)
! "At the end of a region that corresponds to a construct for which the reduction clause was specified, the original list item
!   is updated by combining its original value with the final value of each of the private copies, using the combiner of the
!   specified reduction-identifier." (5.2 §5.5.8 p.135 l.7-10)
! spec (this program): teams reduction(iand: r1a) firstprivate(p1) with an empty body must leave r1a at 1048575 (iand with the
!   identity); AFAR returns 0, and the program is correct when firstprivate(p1) is removed.

Source in the repository

Something to add? Dispute its legality · Report it fixed in a newer release · Report a mistake on this page

Expected and observed

Observed where it behaves incorrectly

RESULT checksum=0

What the specification requires

The specification text this program relies on and the compiler is expected to support.

OpenMP 5.2 §5.5.8 p.134 l.20-24

The reduction clause is a reduction scoping clause and a reduction participating clause, as described in Section 5.5.6 and Section 5.5.7. For each list item, a private copy is created for each implicit task or SIMD lane and is initialized with the initializer value of the reduction-identifier. After the end of the region, the original list item is updated with the values of the private copies using the combiner associated with the reduction-identifier.

OpenMP 5.2 §5.5.8 p.135 l.7-10

At the end of a region that corresponds to a construct for which the reduction clause was specified, the original list item is updated by combining its original value with the final value of each of the private copies, using the combiner of the specified reduction-identifier.

Evidence

Same compiler, one change

  • firstprivate(p1) removed from the teams construct: AFAR 24.3.0: correct; CCE 19.0.0: correct; CCE 20.0.2: correct; CCE 21.0.3: correct; CCE 21.0.4: correct
    Control program
    program fuzz_prog
      implicit none
      integer :: p1, r1a
      integer(8) :: cs
      p1 = -4
      r1a = 1048575
      !$omp target map(tofrom: r1a)
        !$omp teams reduction(iand: r1a)
        !$omp end teams
      !$omp end target
      cs = 0
      cs = modulo(cs * 31_8 + int(r1a, 8), 2147483629_8)
      print '(a,i0)', 'RESULT checksum=', cs
    end program fuzz_prog
    

Other compilers run it correctly

CCE 19.0.0, CCE 20.0.2, CCE 21.0.3, CCE 21.0.4, NVHPC 25.11, NVHPC 26.9, GCC 16.2.0