Hi Matt,
Configuration: Driver Version: 550.90.07 CUDA Version: 12.4
I am encountering a strange issue during compilation. I have a program with the following barebones structure:
1. SUBROUTINE some_subroutine
2. IF (condition) THEN
3. !$omp parallel
4. !$acc stuff
5. !$omp end parallel
6. ELSE
7. !$omp parallel
8. !$acc stuff
9. !$omp end parallel
10. END IF
11. END SUBROUTINE some_subroutine
When compiling (-Minfo=accel,mp
), only information about the first openmp parallel section (and the contained openacc directives, line 4) is displayed, and nothing about the openacc directives in the 2nd openmp section (line 8). When I comment out lines 2-6, 10, optimization information for line 8 does get displayed.
Is this a compiler bug or am I not setting some compiler flags to display all the acc
and omp
optimization info?
Cheers,
Jyoti