fortcom: Fatal: There has been an internal compiler error (C0000005).

[aliwin]

Hi,

I'm getting the following compiler error:
Compiling with Intel Visual Fortran 11.0.074 [IA-32]...
read_orbit_sit.for
fortcom: Fatal: There has been an internal compiler error (C0000005).
From an old posting, looks like it's compiler bug ..?
Any idea how I can avoid this error with IVF 11.0 ?

Thanks,
Ali

[bmchenry]

You'll need to provide some sample code that when compiled produced that error for recommendations on eliminating it. Thereare a plethora of reasons why that error may come up, some due to compiler issues, some due to coding problems.

[Steven_L_Intel1]

That error message can have many possible, unrelated causes. It is always a compiler bug, but not always the same one! Can you attach a short (if possible) source file that gets this error, along with the command line used to do the compile?

For more information, see the Knowledge Base article Internal Compiler Error

[aliwin]

Quoting - Steve Lionel (Intel)

That error message can have many possible, unrelated causes. It is always a compiler bug, but not always the same one! Can you attach a short (if possible) source file that gets this error, along with the command line used to do the compile?

For more information, see the Knowledge Base article Internal Compiler Error

Thank you Steve & all for your input,

It's the fact that I commented the only existing 2 lines inside an "IF" statement that caused this compiler error.
When I commented the "IF" and "ENDIF" lines the compile went fine.

Regards,
Ali

[Steven_L_Intel1]

Would you please attach the source file? You should never get an internal compiler error, even if you have a syntax error in your source. I want to make sure we fix the compiler bug.

[aliwin]

Quoting - Steve Lionel (Intel)

Would you please attach the source file? You should never get an internal compiler error, even if you have a syntax error in your source. I want to make sure we fix the compiler bug.

I don't think it's a syntax error to have an IF block without any instruction inside it, eventhough it doesn't make sense to have this code (and it shouldn't exist).
The file I had this "issue" with is over 500 lines. I tried to keep just the lines that caused the compiler error but the problem wasn't reproducible. Unfortunately I can't send the file without breaking the "Reserved Rights". I tried other combinations to reproduce the problem but wasn't successful.

[Steven_L_Intel1]

You are correct - an IF construct does not require any executable statements inside it.

If you can find a way to provide us with a test case, we'd really appreciate it.

[antony_muthu__Anto_J]

Hi Steve,

I have find the same error that aliwin found some time ago.

I've got:

                 Intel(R) Visual Fortran Compiler Professional for applications running on IA-32,

                 Version 11.0    Build 20090609 Package ID: w_cprof_p_11.0.075
                 Copyright (C) 1985-2009 Intel Corporation.  All rights reserved.

                 fortcom: Fatal: There has been an internal compiler error (C0000005).

The code I am having the error with is FAST by NREL. The code is opensource and can be downloaded from http://wind.nrel.gov/designcodes/simulators/fast/alpha/ It comes with a Visual Studio project and a .bat that can be used to compile it. They say they have compiled their code with Intel® Visual Fortran Compiler XE 12.1.3.300 [IA-32]. I have compiled the previous version of FAST with the Visual Fortran V11.0.075.

Any idea how I can avoid this error with IVF 11.0 ?ç

Thanks,

Jose.

[Kevin_D_Intel]

Jose,

The internal error you and others experienced may not share the same root cause. It is an indicator of an internal failure within the compiler but unless the code you and others compiled is the same or shares something in common it is likely the root causes are different.

The 11.0 release is very old at this point. I will download your code and see whether I can offer any help avoiding it and whether this is already fixed in our latest 14.0 compiler version.

Please stand by...

[Steven_L_Intel1]

Sorry, 11.0 is no longer supported and I won't be able to try to find a workaround for you. Please keep in mind that "Internal compiler error" is a very generic error that can have countless causes.

What I usually do when trying to track down one of these issues is selectively commenting out code in the offending source until I find the minimum I need to reproduce the error. That can sometimes give me a clue as to what I might be able to change to do differently, but this is not always possible.

[Kevin_D_Intel]

Jose,

Here is a mockup reproducing test case that I distilled from the FAST code that suffers the internal error. The internal error occurs in relation to LEN(message) in the message2 declaration and the variable's use in the contained subroutine.

SUBROUTINE FOO(  message )
   CHARACTER(*)              :: message
   CHARACTER(LEN(message))   :: message2

   CONTAINS                                
      SUBROUTINE BAR()
         CALL A_MESSAGE(message2)              
      END SUBROUTINE BAR
END SUBROUTINE FOO

There is no convenient work around available because this sort of usage appears throughout the original source. Your best course of action to resolve this error is to upgrade to at least the Intel® Visual Fortran Compiler XE 12.1.3.300 [IA-32] version NREL notes using for FAST. The internal error is fixed in that version.