I'm working through Professional CUDA C Programming, and in Chapter 3 I run the simpleDivergence program, but it fails immediately. My test is on a Tegra with a GK20A device. Any ideas would be appreciated.
./simpleDivergence using Device 0: GK20A
blocksize = 64; Data size 64 Execution Configure (block 64 grid 1)
warmup <<< 1 64 >>> elapsed 0 sec
Error: simpleDivergence.cu:143, code: 8, reason: invalid device function
And the answer is: the Makefile in the code download is not appropriate; the compile command in the book is correct.