sciinstitute / structuredeikonal Goto Github PK

make the current main for an example folder.

inputs would be volume voxel speed assignments and seed point initialization.

With my GeForce GT650M card on OSX 10.11, running Example1 with default parameters fails with message:

CUDA error! 2

If I change the size option to 128 (default is 256), the example runs.

It would be helpful if the error message output by the library was more informative. At the very least, the readme should be updated to address what to do with this error.

I have OSX 10.11, XCode 7.3, GeForce GT650M, CUDA 6.0
With this combination, I get a compile error: 'array' not found
I tried fixing this by setting CMAKE_CXX_FLAGS to --std=c+11 --stdlib=libc++
and then get errors that nvcc doesn't recognize these flags
After upgrading to CUDA 7.5, I still get the compile error with a default cmake build, but the above setting to CMAKE_CXX_FLAGS works.

However, then I get a compile error in helper_timer.cpp at the line StopWatchLinux~StopWatchLinux
which should read StopWatchLinux::~StopWatchLinux

I recommend updating the readme to address the former (both supported CUDA version and CMAKE_CXX_FLAGS settings) and fixing the code to address the latter.