- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hello,
The following code raises an internal error with ifort 15.0.2 and 16.0.1:
catastrophic error: **Internal compiler error: segmentation violation signal raised** Please report this error along with the circumstances in which it occurred in a Software Problem Report. Note: File and line given may not be explicit cause of this error.
program main !$ use omp_lib integer :: a a = 42 !$omp parallel !$omp master call foo(a) !$omp end master !$omp end parallel contains subroutine foo(a) !$ use omp_lib integer :: a !$omp task depend(in:a) write(*,*) '--->',a !$omp end task end subroutine foo end program
The code compiles if the depend clause is removed. It also compiles if the code in the foo subroutine is inserted directly in the main program body, or, inversely, if the parallel region is created in the foo subroutine.
Thanks,
Theo
ETA: Another compiler error (which may or may not be related to the first one)
I believe I have found another problem with the following code:
program main !$ use omp_lib integer :: a a = 42 !$omp parallel !$omp master call foo(a) !$omp end master !$omp end parallel contains subroutine foo(a) !$ use omp_lib integer, target :: a integer, pointer :: p p => a !$omp task firstprivate(p) depend(in:p) write(*,*) '--->',p !$omp end task end subroutine foo end program
The compiler raises the following error: /tmp/ifort0sm6zA.o:(.data+0xf0): undefined reference to `foo$P$_4.0.2'
Interestingly, the code compiles without problem when the flag -O0 is used, so it looks the error might be due to a too aggressive optimization.
Theo
コピーされたリンク
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
The first problem I can reproduce in 16.0.1 but not in 16.0.2, which should be out within a week or so.
The second problem I can reproduce in our latest build and I will send that on to the developers - thanks. Issue ID DPD200381389.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
I expect the second issue to be fixed in Parallel Studio XE 2016 Update 3. I'll also note that Update 2 has been delayed about a week.
