Internal compiler error (C0000005) with Floating-Point Exception Handling

[stankoz]

Hello, I've been reading about this error in the forums and it appears to be a bug.

I am certain that it is not the code because it had been compiling just fine and I never change it. I don't even recall changing any of the floating point configuration settings either but I've been able to isolate the error to the flag for "Underflow gives 0.0; Abort on other IEEE exceptions (/fpe:0)." Everything compiles fine with the the default project setting of "Produce NaN, signed infinities, and denormal results"

I'm using Visual Studio 2010 version 10.0.40219 with Visual Fortran version 2013.1.119

I'm trying to compile a USGS groundwater model that is available here: http://water.usgs.gov/ogw/seawat/

[Yuan_C_Intel]

Hi, Stankoz

I have reviewed your issue. I just downloaded the package and built it with 2013 Update 1 compiler.

I have reproduced the internal error for x64, release build, after I switch the flag to "Underflow gives 0.0; Abort on other IEEE exceptions (/fpe:0)."

But this error is not reproducible with latest 14.0 update 3 compiler or 15.0 Beta compiler. Only some name convention and project configuration problem need to be adjusted.

Could you upgrade your compiler version to have a try?

Thank you