cannot open shared object file: No such file or directory with cross compiled elf

We cross compile and link our elf from a x86 host against aacrch64 libs and headers. After copying the compiled binary to a TX1 and try to run it, we get:

“error while loading shared libraries: cannot open shared object file: No such file or directory.”

The readelf reports:

0x000000000000000f (RPATH) Library rpath: [/usr/local/cuda/lib64:/usr/local/cuda/lib64/stubs]

which does exist on the TX1.
Does the elf has to be compiled on a TX1 for it to work? Or I am missing something in the RPATH or our build environment?

Thanks for the help!


Thanks for your question.
This is NVMLibrary and this library doesn’t support tegra platform.

Supported products:
Full Support
  NVIDIA Tesla Line:
    S2050, C2050, C2070, C2075,
    M2050, M2070, M2075, M2090,
    X2070, X2090,
    K8, K10, K20, K20X, K20Xm, K20c, K20m, K20s, K40c, K40m, K40t, K40s, K40st, K40d, K80</li>
  NVIDIA Quadro Line:
    410, 600, 2000, 4000, 5000, 6000, 7000, M2070-Q
    K2000, K2000D, K4000, K5000, K6000
    K1, K2, K340, K520
  NVIDIA GeForce Line: None

Limited Support
  NVIDIA Tesla Line: S1070, C1060, M1060 and all other previous generation Tesla-branded parts
  NVIDIA Quadro Line: All other current and previous generation Quadro-branded parts
  NVIDIA GeForce Line: All current and previous generation GeForce-branded parts

Extract from: