Last test pass, so I have narrowed the seg fault to 4 loops. Removed two more from that list for this test.

Author: Daniel Vickers

src/common/m_mpi_common.fpp

@@ -837,38 +837,38 @@ contains
                         end do
                         $:END_GPU_PARALLEL_LOOP()
 
-                        #:call GPU_PARALLEL_LOOP_OLD(collapse=5,private='[r]')
-                            do i = nVar + 1, nVar + 4
-                                do l = 0, p
-                                    do k = 0, buff_size - 1
-                                        do j = -buff_size, m + buff_size
-                                            do q = 1, nb
-                                                r = (i - 1) + (q - 1)*4 + nb*4 + v_size* &
-                                                    ((j + buff_size) + (m + 2*buff_size + 1)* &
-                                                     (k + buff_size*l))
-                                                buff_send(r) = real(mv_in(j, k + pack_offset, l, i - nVar, q), kind=wp)
-                                            end do
+                        $:GPU_PARALLEL_LOOP(collapse=5,private='[r]')
+                        do i = nVar + 1, nVar + 4
+                            do l = 0, p
+                                do k = 0, buff_size - 1
+                                    do j = -buff_size, m + buff_size
+                                        do q = 1, nb
+                                            r = (i - 1) + (q - 1)*4 + nb*4 + v_size* &
+                                                ((j + buff_size) + (m + 2*buff_size + 1)* &
+                                                 (k + buff_size*l))
+                                            buff_send(r) = real(mv_in(j, k + pack_offset, l, i - nVar, q), kind=wp)
                                         end do
                                     end do
                                 end do
                             end do
-                        #:endcall GPU_PARALLEL_LOOP_OLD
+                        end do
+                        $:END_GPU_PARALLEL_LOOP()
                     end if
                 #:else
-                    #:call GPU_PARALLEL_LOOP_OLD(collapse=4,private='[r]')
-                        do i = 1, nVar
-                            do l = 0, buff_size - 1
-                                do k = -buff_size, n + buff_size
-                                    do j = -buff_size, m + buff_size
-                                        r = (i - 1) + v_size* &
-                                            ((j + buff_size) + (m + 2*buff_size + 1)* &
-                                             ((k + buff_size) + (n + 2*buff_size + 1)*l))
-                                        buff_send(r) = real(q_comm(i)%sf(j, k, l + pack_offset), kind=wp)
-                                    end do
+                    $:GPU_PARALLEL_LOOP(collapse=4,private='[r]')
+                    do i = 1, nVar
+                        do l = 0, buff_size - 1
+                            do k = -buff_size, n + buff_size
+                                do j = -buff_size, m + buff_size
+                                    r = (i - 1) + v_size* &
+                                        ((j + buff_size) + (m + 2*buff_size + 1)* &
+                                         ((k + buff_size) + (n + 2*buff_size + 1)*l))
+                                    buff_send(r) = real(q_comm(i)%sf(j, k, l + pack_offset), kind=wp)
                                 end do
                             end do
                         end do
-                    #:endcall GPU_PARALLEL_LOOP_OLD
+                    end do
+                    $:END_GPU_PARALLEL_LOOP()
 
                     if (qbmm_comm) then
                         #:call GPU_PARALLEL_LOOP_OLD(collapse=5,private='[r]')