Strange cuda compilation error

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,

Anyone knows what this could be about?

1>CudaSkrjabinField00.cu

1>CudaSkrjabinField00.cudafe1.gpu

1>CudaSkrjabinField00.cudafe2.gpu

1>### Assertion failure at line 123 of …/…/be/cg/NVISA/expand.cxx:

1>### Compiler Error in file CudaSkrjabinField00.cpp3.i during Code_Expansion phase:

1>### unexpected mtype

1>nvopencc ERROR: C:\CUDA\bin/…/open64/lib//be.exe returned non-zero status 1

this file CudaSkrjabinField00.cpp3.i looks like:

[codebox]#line 1 “CudaSkrjabinField00.cudafe2.gpu”

#line 7 “CudaSkrjabinField00.cudafe2.gpu”

#line 482 “C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\crtdefs.h”

typedef unsigned size_t;

#line 484 “C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\crtdefs.h”

#line 1 “C:\CUDA\bin/…/include\crt/device_runtime.h”[/codebox]

This is release, debug compiles fine…

Kind Regards,

Daniel Dekkers

2

Hey DoZo,

I have exactly the same problem, did you solve it somehow? Any help appreciated, thanks

–jan

3

Here is the solution nvcc internal error

4

I got the same error when I used pow((double)a,(double)B). The compiling goes well when the function is replaced by pow((float)a,(float)B).