Call to NVHPC runtime function not supported - pghpf_rnum_i8

[alexander.pletzer]

Hi,

I’m getting the following error when compiling:
VFORTRAN-S-1058-Call to NVHPC runtime function not supported - pghpf_rnum_i8 (foo_mod.f90: 8)

The code is below. Note that subroutine foo has the directive $acc routine, used to generate code for the device. Removing this comment and the code will compile. However, I need to get foo running on the device; therefore I need a version of the intrinsic subroutine random_number to be compiled for GPU. The error message suggests that it is not possible to run random_number on a GPU (?) What fix or workaround would you suggest?

Thanks for your help.

Environment: nvfortran 24.11-0 64-bit target on x86-64 Linux -tp broadwell

Steps:
cat > foo_mod.f90 << EOF
module foo_mod
contains
subroutine foo(i, array)
!$acc routine
implicit none
integer :: i
real :: array(:)
call random_number(array)
end subroutine
end module
EOF
nvfortran -c -acc=gpu foo_md.f90

[MatColgrove]

@alexander.pletzer

The error message suggests that it is not possible to run random_number on a GPU (?) What fix or workaround would you suggest?

Correct, and nor would you want the basic random number generator on the device. RNGs are inherently problematic when run in parallel given they have a global state which would be shared by all threads, thus causing a race condition. Instead, you’ll need to use an RNG where each thread carries it’s own private state.

I’ve written a few posts about this (links below). The summary is that you need to look to use a device side RNG like cuRAND or DES PRNG.

Random numbers inside OpenACC loop (nvfortran compiler)

Is there a straightforward way to generate random numbers inside an OpenACC parallel loop? I want the seed within each loop/thread to be different. The reference code that demonstrates the required task to be done is as follows: PROGRAM genRandNums implicit none integer :: i double precision :: a(100) !$acc parallel loop DO i = 1,100 call random_seed() call random_number(a(i)) ENDDO END PROGRAM This code works on the gnu’s 'gfortran ’ but throws a…

Random Numbers in OpenACC

I am trying to generate uniform random numbers inside an OpenAcc parallel loop. Using rand() gives an error and I saw several posts in the forum that suggest either using cuRAND or generating the numbers on the host and then passing it to the GPU. The issue with generating numbers on the host is, that it is a little complicated to determine how many random numbers I will need since my code is a bit complex. I all need is a function that can generate uniform random numbers between a range [A, B]…

Hope this helps,
Mat