In CUDA, how can we call a device function in another translation unit?

Question

I'm pretty new to CUDA. I use Microsoft Visual Studio 2010 where I don't need to worry about writing a makefile. A problem arose as I tried to call in a .cu file a device function which was declared in the .h file and defined in another .cu file. At the end of building, I received an error message:

1>ptxas : fatal error : Unresolved extern function '_Z22atomicAddEmulateDoublePdd'

This appears in both CUDA 4.2 and 5.0. I'm wondering how should I configure my MVS to avoid this error. Sorry for the nooby questions and thanks for any suggestion!

I have the exact same problem... would be interested to know if you solved it! — Kristian D'Amato, Aug 17 '12 at 12:56
@KristianD'Amato: By having VS generate relocatable device code. — biubiuty, Aug 21 '12 at 18:00
I have attempted that, but it didn't seem to work. Can you give me an idea how you did that exactly? — Kristian D'Amato, Aug 22 '12 at 08:53
Ahhh... the little grey cells must have been asleep! I missed the instruction in the CUDA C++ tab! — Kristian D'Amato, Aug 22 '12 at 11:07

score 3 · Accepted Answer · answered Aug 16 '12 at 19:53

3

CUDA 4.2 and does not support static linking so device functions must be defined in the same compilation unit. A common technique is to write the device function in a .cuh file and include it in the .cu file.

CUDA 5.0 supports a new feature called separate compilation. The CUDA 5.0 VS msbuild rules should be available in the CUDA 5.0 RC download.

answered Aug 16 '12 at 19:53

Greg Smith

11,007
2
36
37

Without separate compilation, I use this method to have a single definition file for both host and device; [link](http://stackoverflow.com/questions/6978643/cuda-and-classes). How to set-up separate compilation? – Ali Naci Erdem May 22 '14 at 17:50

In CUDA, how can we call a device function in another translation unit?

1 Answers1