CUDA11.2+PVF14.10环境下OpenAcc加速失败原因

xiaoshuaihao666

各位老师好，我在PGI Visual Fortran14.10环境下在language中打开了OpenAcc 指令，但是会出现这个错误是什么原因？希望能够得到老师的指导。错误指示为：Call to cuModuleLoadData returned error 209 :No binary for GPU

程序如下：

[Fortran] 纯文本查看 复制代码

查看代码

打印?

01	program main

02	use openacc

03	implicit none

04	real,allocatable::v1(:)

05	integer length,idx

06	call acc_set_device_num(2,acc_device_nvidia)

07	length=102410241024

08	allocate(v1(length))

09	v1(1:length)=0.0

10	!$acc data copy(v1(1:length))

11	call add1(v1,length)

12	!$acc update host(v1(1:5))

13	print*,‘v1(1)=’,v1(1:5)

14	do idx=1,length

15	v1(idx)=v1(idx)+10

16

enddo

17	print*,‘v1(1)=’,v1(1:5)

18	!$acc update host(v1(1:5))

19	print*,‘v1(1)=’,v1(1:5)

20	!$acc end data

21	deallocate(v1)

22	end program

23	subroutine add1(vec,length)

24	use openacc

25	implicit none

26	integer,intent(in)::length

27	real,intent(inout)::vec(1:length)

28	integer idx

29	!$acc parallel loop present(vec(1:length))

30	do idx=1,length

31	vec(idx)=vec(idx)+idx

32

enddo

33	end subroutine