Random errors generating modules

[nvaneck]

When I buld a project, I randomly get one or more "Error in creating the compiled module file".  For xample, the last occurrence was:

"error #7001: Error in creating the compiled module file.   [OPTSOL__genmod]"

I then just re-issue the build and they usually go away.

Somtimes this doesn't work and I have to compile a file separately, or a different one and build again, until the messages disappear.

Please note this happens without any changes to the modules.  It is not generally reproducible on demand--just occurs randomly.

I am using release 11 of the compiler; is this something that was addressed in future releases of the compiler?. It's not a show-stoppe, but is an annoyance.

[Steven_L_Intel1]

Something on your system is preventing the compiler from writing the module file. I have seen antivirus programs do this. Also, is this just a single project or is it multiple projects in a solution? One experiment would be to boot into Safe Mode and see if you can reproduce the problem.

[nvaneck]

Thanks, Steve

It's just a single project.

Only anti-virus is MS defender.  I don't think it's worth the trouble to try safe mode--I could end up testing a dozen times and not encounter the problem and it wouldn't be definitive.  As I said, it's random and does not happen often enough to pursue this way. I guess I'll just live with it until something worse happens or something eventually sheds light on it. It doesn't really cost me much time to re-compile a routine when it happens, and when it complains about several files, building again clears them up. I was just hoping it was related to something in that compiler version, but since it's not I'll just ignore it all.

[nvaneck]

Steve,

I just noticed as it happened again that it always thows an intrernal compiler error, so I can't see it is anything to do with something on my computer under my control...

>D:\Projects\MicrOsiris\PGMS\RECODE.F90(2429): error #7001: Error in creating the compiled module file.   [SYNTAX__genmod]1>fortcom: Fatal: There has been an internal compiler error (C0000005).

[Steven_L_Intel1]

Does it ALWAYS give you an ICE?  Please provide us with the sources needed to reproduce the error (and compile options used or a copy of the .vfproj).

[nvaneck]

I think it does, but as noted before it is not generally reproducible, but happens often enough upon repeated builds. I'm not sure how best to try and get it to you. It's a large project, but not a giant one, with many routines in two drirectories included in the project.

[Bernard]

This seems like an Access Violation error.It could happen by accessing null pointers or calling unmapped exported DLL routines or passing bad pointers which can point to unaccessible memory.You should obtain a minidump file of the failed process (fortcom) and contact Intel devs.

[Steven_L_Intel1]

Don't bother with the minidump - there is nothing useful we can do with that. I suggest using Intel Premier Support to supply us with what we need to try to reproduce the problem.

[nvaneck]

Can't send it to support; I guess since I don't have the latest compiler version.  Anyway,I'm 'not authorized'

[Steven_L_Intel1]

That means you don't have a current support license. If you want to use Send Author a Message to send me a message, and attach a ZIP of the project there, I'll at least try it out and see what results I get with the current compiler. Please do a Build > Clean first.

[lklawrie]

I'm suddenly having this happen in a CMake generated (with multiple targets) project file.  Currently using the V13 compiler -- but I don't think it's a compiler issue.  I, though another person reported he couldn't do this, can just "rebuild" the targets that fail and they work.  But don't know where to go -- I think it's some kind of VS (2008) or CMake generation problem.

[nvaneck]

I'm using VS 2010 and V12 when I get this problem and I don't use CMake......

[Lorri_M_Intel]

Are there parallel builds happening here? 

I'm wondering if the same-named file is being generated/accessed in two threads

                 --Lorri

[lklawrie]

Lorri,

Yes the same named file is being generated (in my case -- perhaps this should be a different thread) -- I was thinking perhaps there was some kind of collision happening on faster processors (and more cores) that wouldn't show up on slower machines even though they are in different projects.

[nvaneck]

Well, in my case three are no parallel builds unless the compiler is doing something I don't know about. I do run it on a multi-core machine.

Neal

[nvaneck]

FYI, after upgrading to Windows 8.1, with numerous compiles and a rebuild, the problem has not occurred.....