Suppose a subroutine has two loops
loop1
loop2
Which way of writing the OpenACC directives is better? and why?
method 1
!$acc kernels
loop1
loop2
!$end acc kernels
method 2
!$acc kernels loop
loop1
!$acc kernels loop
loop2
Thanks,
Ping
Suppose a subroutine has two loops
loop1
loop2
Which way of writing the OpenACC directives is better? and why?
method 1
!$acc kernels
loop1
loop2
!$end acc kernels
method 2
!$acc kernels loop
loop1
!$acc kernels loop
loop2
Thanks,
Ping
Which way of writing the OpenACC directives is better? and why?
Neither is better than the other. It’s more personal preference and what works better for the particular context.