Operation netpyne on Neuron with CoreNeuron and GPU

[Kim1230]

When I install Neuron integrated with CoreNeuron and CUDA(nvHPC). I download the netpyne from https://github.com/suny-downstate-medical-center/S1_netpyne/tree/coreneuron. And I try to operate command: nrnivmodl -coreneuron mod. It has the bug.
The bug is shown as:

NVC++-W-1053-External and Static variables are not supported in acc routine - _ZN10coreneuron11nrn_threadsE (x86_64/corenrn/mod2c/ProbAMPANMDA_EMS.cpp: 527)
NVC++-W-1053-External and Static variables are not supported in acc routine - _ZN10coreneuron11nrn_threadsE (x86_64/corenrn/mod2c/ProbGABAAB_EMS.cpp: 516)
NVC++/x86-64 Linux 22.3-0: compilation completed with warnings
NVC++/x86-64 Linux 22.3-0: compilation completed with warnings
NVC++-W-1057-Static variables are not supported in acc routine - _ZZN54_INTERNAL_32_x86_64_corenrn_mod2c_StochKv_cpp_b3aebc776gammlnEdE3cof (x86_64/corenrn/mod2c/StochKv.cpp: 814)
NVC++-W-1057-Static variables are not supported in acc routine - _ZN54_INTERNAL_32_x86_64_corenrn_mod2c_StochKv_cpp_b3aebc7710coreneuron12_tmin_tratesE (x86_64/corenrn/mod2c/StochKv.cpp: 524)
NVC++-S-1058-Call to variadic function not supported - fprintf (x86_64/corenrn/mod2c/StochKv.cpp: 618)
NVC++/x86-64 Linux 22.3-0: compilation completed with severe errors
make: *** [/home/yjy/test/install/share/coreneuron/nrnivmodl_core_makefile:243: x86_64/corenrn/build/StochKv.o] Error 2
make: *** Waiting for unfinished jobs....

Steps to reproduce the behavior:

1. Go to ./sim
2. input nrnivmodl -coreneuron mod

System information

• OS and version: Linux 2288HV5-47-26 4.19.90-24.4.v2101.ky10.x86_64
• NetPyNE version: https://github.com/suny-downstate-medical-center/S1_netpyne/tree/coreneuron
• NEURON version: it seems to be 8.02
• Python version:3.8
• CUDA: 11.6 /NVhpc: 22.3

[Kim1230]

And I find that in this branch https://github.com/suny-downstate-medical-center/S1_netpyne/tree/coreneuron. The cfg.py doesn't have parameter : cfg.gpu. If I want to test netpyne with gpu. Which package I can download to test ? Thanks!

[Kim1230]

And when I comment the x86_64/corenrn/mod2c/StochKv.cpp: 618. It can be compiled. But when I operate x86_64/special with Coreneuron. It directly calls the CUDA and GPU. And the get the error: The error is "call to cudaGetSymbolAddress returned error 13:"