We are running Telsa M2070 which are Compute Capability 2.0.
One item of concern is the line in the “5.4 USER-CUDA package” documentation:
“Your NVIDIA GPU needs to support Compute Capability 1.3. This list may help you to find out the Compute Capability of your card: http://en.wikipedia.org/wiki/Comparison_of_Nvidia_graphics_processing_units “
Are we good? One would assume LAMMPS would support Compute Capability 1.3 OR higher.
The main issue is when compiling lammps-20Apr12 with USER-CUDA module I keep hitting the same error no matter what I try.
I’m able to build the standard packages without user-cuda just fine but I’m hitting a compilation problem with force.cpp when adding user-cuda.
Using mvapich_intel compiler with CUDA v4.0.17 and fftw v2.1.5
I have the liblammpscuda.a built and followed all the steps for building user-cuda package.
I’m out of ideas other than I compilation problem.