Issue while running VecPlay test on GPU

[pramodk]

When we run simple VecPlay test on GPU then we get following error:

.....
 Start time (t) = 0

Memory (MBs) :  After mk_spikevec_buffer : Max 221.0977, Min 221.0977, Avg 221.0977

MPT ERROR: Rank 0(g:0) received signal SIGSEGV(11).
	Process ID: 11846, Host: r2i3n2, Program: /gpfs/bbp.cscs.ch/home/kumbhar/tmp/testcorenrn/x86_64/special-core
	MPT Version: HPE HMPT 2.22  03/31/20 16:17:35

To Reproduce

Build the latest master with GPU enabled:

module load unstable nvhpc cuda/11.1.0 hpe-mpi cmake python-dev
cmake .. -DCMAKE_INSTALL_PREFIX=`pwd`/install -DCORENRN_ENABLE_GPU=ON
make -j install

And then run vecplay test from testcorenrn repo then we get:

kumbhar@r2i3n2:~/tmp/testcorenrn$ ./x86_64/special-core -d coredat --mpi --gpu -e 1
 num_mpi=1
 num_omp_thread=1

 Info : 4 GPUs shared by 1 ranks per node
....

 Start time (t) = 0

 Memory (MBs) :  After mk_spikevec_buffer : Max 221.0977, Min 221.0977, Avg 221.0977

MPT ERROR: Rank 0(g:0) received signal SIGSEGV(11).
	Process ID: 11846, Host: r2i3n2, Program: /gpfs/bbp.cscs.ch/home/kumbhar/tmp/testcorenrn/x86_64/special-core
	MPT Version: HPE HMPT 2.22  03/31/20 16:17:35

MPT: --------stack traceback-------
MPT: Attaching to program: /proc/11846/exe, process 11846
MPT: [New LWP 11857]
MPT: [New LWP 11856]
MPT: [Thread debugging using libthread_db enabled]
MPT: Using host libthread_db library "/lib64/libthread_db.so.1".
MPT: 0x00007fffed6d6179 in waitpid () from /lib64/libpthread.so.0

System (please complete the following information)

• BB5
• Compiler: nvhpc/20.9
• Version: master
• Backend: [e.g. GPU]

Additional context

I believe the error was introduced in #283. In that PR we said I manually ran tests but I believe the PR got changed afterwards and broke some stuff (?).

(#308 is important to have in place!)