openMP offload

Neo123

如何采用visual studio + oneAPI实现程序的GPU并行？我采用如下程序测试发现omp_get_num_devices（）=0，也就是无法识别GPU,编译报错为未加载openCL.pdb。已经根据显卡驱动版本安装了CUDA 软件

[Fortran] 纯文本查看 复制代码

查看代码

打印?

01	program test

02	use omp_lib

03	implicit none

04

05	!integer:: host, default_device,x

06	!!omp_set_default_device(1)

07	!host = omp_get_initial_device()

08	!x = omp_get_num_devices()

09

!

10	!write(*,*) host

11	!write(*,*) default_device

12	!write(*,*) x

13

14	real(8) :: a(1024), b(1024), c(1024)

15	integer:: i

16

17	!$omp target device(0)

18

19	!$omp teams distribute parallel do simd

20	do i=1,1024

21	c(i)=c(i)+a(i)*b(i)

22

end do

23	!$omp end target

24	end program