public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug fortran/106566] New: [OpenMP]
@ 2022-08-08 20:35 burnus at gcc dot gnu.org
2022-08-09 18:28 ` [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures burnus at gcc dot gnu.org
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: burnus at gcc dot gnu.org @ 2022-08-08 20:35 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106566
Bug ID: 106566
Summary: [OpenMP]
Product: gcc
Version: 13.0
Status: UNCONFIRMED
Keywords: openmp, rejects-valid
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: burnus at gcc dot gnu.org
CC: jakub at gcc dot gnu.org
Target Milestone: ---
The OpenMP example
https://github.com/OpenMP/Examples/blob/v5.2/SIMD/sources/linear_modifier.1.f90
is rejected as:
13 | !$omp declare simd(add_one2) linear(p: ref) simdlen(8)
| 1
Error: Symbol ‘add_one2’ at (1) has already been host associated
if it is a module procedure. When commenting the 'module ... contains' and 'end
module', it compiles just fine.
(Similar the .2 and .3 in the examples.)
^ permalink raw reply [flat|nested] 4+ messages in thread
* [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures
2022-08-08 20:35 [Bug fortran/106566] New: [OpenMP] burnus at gcc dot gnu.org
@ 2022-08-09 18:28 ` burnus at gcc dot gnu.org
2022-08-16 8:16 ` burnus at gcc dot gnu.org
2022-08-16 15:26 ` burnus at gcc dot gnu.org
2 siblings, 0 replies; 4+ messages in thread
From: burnus at gcc dot gnu.org @ 2022-08-09 18:28 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106566
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |accepts-invalid
Summary|[OpenMP] |[OpenMP] declare simd fails
| |with with bogus "already
| |been host associated" for
| |module procedures
--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Additionally, the following is not diagnosed – at least not for this example.
"For Fortran, a declarative directive must appear after any USE, IMPORT, and
IMPLICIT statements in a declarative context."
(The original example shows this issue. This is reported by other compilers and
is being fixed on the OpenMP examples side.)
Example - *FAILS* ("has already been host associated") but is *VALID*
module m
integer, parameter :: NN = 1023
integer :: a(NN)
contains
subroutine add_one2(p)
implicit none ! <<<< valid - must before declare
!$omp declare simd(add_one2) linear(p: ref) simdlen(8)
integer :: p
p = p + 1
end subroutine
end module
The following is *COMPILING* - as there is no MODULE:
subroutine add_one2(p)
!$omp declare simd(add_one2) linear(p: ref) simdlen(8)
implicit none ! <<<< invalid because after declare.
integer :: p
p = p + 1
end subroutine
Note: This example is on purpose invalid as 'implicit none' has been moved
after 'omp declare'. Otherwise, it would be valid.
^ permalink raw reply [flat|nested] 4+ messages in thread
* [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures
2022-08-08 20:35 [Bug fortran/106566] New: [OpenMP] burnus at gcc dot gnu.org
2022-08-09 18:28 ` [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures burnus at gcc dot gnu.org
@ 2022-08-16 8:16 ` burnus at gcc dot gnu.org
2022-08-16 15:26 ` burnus at gcc dot gnu.org
2 siblings, 0 replies; 4+ messages in thread
From: burnus at gcc dot gnu.org @ 2022-08-16 8:16 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106566
--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
The ordering part (must be after IMPORT/IMPLICIT/USE) is now tracked in
PR106639 - note that forcing it after IMPORT/IMPLICIT/USE is new since OpenMP
5.1.
^ permalink raw reply [flat|nested] 4+ messages in thread
* [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures
2022-08-08 20:35 [Bug fortran/106566] New: [OpenMP] burnus at gcc dot gnu.org
2022-08-09 18:28 ` [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures burnus at gcc dot gnu.org
2022-08-16 8:16 ` burnus at gcc dot gnu.org
@ 2022-08-16 15:26 ` burnus at gcc dot gnu.org
2 siblings, 0 replies; 4+ messages in thread
From: burnus at gcc dot gnu.org @ 2022-08-16 15:26 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106566
--- Comment #3 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Patch for the rejects-valid issue of module variable + a fix
for an absent linear step with OpenMP 5.2 syntax:
https://gcc.gnu.org/pipermail/gcc-patches/2022-August/599829.html
As mentioned, the declarative directive before IMPORT/IMPLICIT/USE issue
is a only invalid since OpenMP 5.1 - and tracked in PR106639.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2022-08-16 15:26 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-08-08 20:35 [Bug fortran/106566] New: [OpenMP] burnus at gcc dot gnu.org
2022-08-09 18:28 ` [Bug fortran/106566] [OpenMP] declare simd fails with with bogus "already been host associated" for module procedures burnus at gcc dot gnu.org
2022-08-16 8:16 ` burnus at gcc dot gnu.org
2022-08-16 15:26 ` burnus at gcc dot gnu.org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).