| View previous topic :: View next topic |
| Author |
Message |
chylin
Joined: 19 May 2006
Posts: 4
|
 Posted: Fri May 19, 2006 5:06 pm Post subject: Internal compiler error at forall construct |
 |
|
When compiled using IPA optimization (-fastsse -Mipa=fast,pure -O2), the following code:
| Code: | if (iter2 /= 0) then
! Do some initializations [... Omitted ...]
main_iteration: do iter=iter1,iter2
! [... Omitted ...]
if (iter <= 1) then
! Do some initializations [... Omitted ...]
else ! (iter > 1)
select case (imass)
case (2) ! Axially distributed
! Evaluate the rate of density change:
forall (i = 1:ncie, j = 1:neta) ! Line 427
rhodot(i,j) = dxm_dz_new(i)*dz_dvol(i)
end forall
end select
end if
! Body of main loop [... Omitted ...]
end do main_iteration
end if
|
causes an internal compiler error (edited here for length):
IPA: Recompiling intgrt.obj: new IPA information
PGF90-F-0000-Internal compiler error. rewrite_subroutine: \
unknown stmt found 0 (intgrt.f90: 427)
PGF90/any Linux/x86-64 6.1-3: compilation aborted
If the forall construct in the above code fragment is replaced by the following:
| Code: | ! Evaluate the rate of density change:
do j = 1,neta
rhodot(1:ncie,j) = &
dxm_dz_new(1:ncie)*dz_dvol(1:ncie)
end do
|
then the program compiles with no errors.
There are many forall constructs in the program, scattered over a number of routines, but the one indicated above is the only one that caused any difficulty in the IPA optimization stage. |
|
| Back to top |
|
 |
mkcolg
Joined: 30 Jun 2004
Posts: 7319
Location: The Portland Group Inc.
|
| Posted: Mon May 22, 2006 12:46 pm Post subject: |
|
|
Hi chylin,
Can you please send a report to trs@pgroup.com? I was unable to recreate the error with the given information so if you could include the full source, or a test case which illustrates the problem, we would appreciate it.
Thanks,
Mat |
|
| Back to top |
|
|
chylin
Joined: 19 May 2006
Posts: 4
|
| Posted: Tue May 23, 2006 9:56 am Post subject: |
|
|
Mat,
I'm afraid that I can't give you the full source. I'll see if I can't put together an illustrative test case.
Carl |
|
| Back to top |
|
|
|
Search
Memberlist
Register
Profile
Log in
