compiler error message

Accelerated Computing CUDA CUDA Programming and Performance
1

I’m attempting to compile with nvcc, and I keep getting the same error message. There is no .obj file written out for the device code. Is there a listing of compiler errors somewhere? Can’t seem to find one. If someone could point me in the right direction for this message, I’d appreciate it very much. I’ve substituted “program_name” the real file name, and “kernel_function” for the real function name. The problem seems to be in my kernel function. I’m using Visual Studio.

1>program_name.cudafe1.gpu
1>program_name.cudafe2.gpu
1>program_name.cpp3.i(0): ### Compiler Error (user routine ‘_globfunc__Z18kernel_functiondddddllP7double2PhS0’) during Global Optimization – Offline value numbering phase:
1>### Invalid machine type FQ in Targ_Is_Integral
1>nvopencc ERROR: C:\CUDA\bin/…/open64/lib//be.exe returned non-zero status 1

Thanks very much

2

I managed to track down the problem. The problem doesn’t seem to relate to the error message. but compiler error messages often don’t. I’d still like to find a reference to the nvcc error messages, though.

3

Can you tell me what the source of your problem was? I am seeing the same error, but the code compiles fine in emulation mode.

4

I’d used a double double literal, i.e. 180.0L, in one of the device functions called by the kernel function.

5

THANK YOU! This fixed my compiler error too.