How to solve the problem?
the error point to the line of code “ERR = cudaMemcpyAsync(m_DEN(STARTA),DEN(STARTA),NPRT)”
And the detailed information by cuda-memcheck are listed
0: copyout MemcpyAsync (host=0x485360af8, dev=0x986780af8, size=32000, stream=0) FAILED: 11(invalid argument)
========= Program hit cudaErrorInvalidValue (error 11) due to “invalid argument” on CUDA API call to cudaMemcpyAsync.
========= Saved host backtrace up to driver entry point at error
========= Host Frame:/usr/lib64/nvidia/libcuda.so.1 [0x2f31b3]
========= Host Frame:/opt/pgi/linux86-64/2015/cuda/7.5/lib64/libcudart.so.7.5 (cudaMemcpyAsync + 0x1b3) [0x435d3]
========= Host Frame:/opt/pgi/linux86-64/15.10/lib/libcudafor.so (pgf90_dev_copyout_async + 0x38) [0x1a7b0]
========= Host Frame:/opt/pgi/linux86-64/15.10/lib/libcudafor.so (cudamemcpyasync3r8out_ + 0x1c) [0x2346c]
========= Host Frame:./fasMD.exe [0x13de1f]
========= Host Frame:./fasMD.exe [0x13b8f6]
========= Host Frame:./fasMD.exe [0x12a62f]
========= Host Frame:./fasMD.exe [0x51af]
========= Host Frame:./fasMD.exe [0x7a2e]
========= Host Frame:./fasMD.exe [0x3b34]
========= Host Frame:/lib64/libc.so.6 (__libc_start_main + 0xfd) [0x1ed5d]
========= Host Frame:./fasMD.exe [0x3a29]
========= ERROR SUMMARY: 1 error
the codes can be run successful in one machine, but it can not work well in the present machine, I don’t know Why. and what should I do?