Commit 51ad5b39 authored by Andreas Marek's avatar Andreas Marek
Browse files

bit of cleanup

parent 59df91ce
...@@ -1489,22 +1489,6 @@ ...@@ -1489,22 +1489,6 @@
#endif #endif
hh_tau_dev, kernel_flops, kernel_time, n_times, 0, current_local_n, i, & hh_tau_dev, kernel_flops, kernel_time, n_times, 0, current_local_n, i, &
last_stripe_width, kernel) last_stripe_width, kernel)
!#if COMPLEXCASE == 1
! if (useGPU) then
! call compute_hh_trafo_complex_gpu_&
! &PRECISION&
! &(aIntern_dev, bcast_buffer_dev, hh_tau_dev, 0, current_local_n, i, a_off, dev_offset, dev_offset_1, &
! dev_offset_2, a_dim2, &
! kernel_flops, kernel_time, last_stripe_width, n_times, nbw, stripe_count, stripe_width)
! else
! call compute_hh_trafo_complex_&
! &PRECISION&
! &(aIntern, stripe_width, a_dim2, stripe_count, &
! a_off, nbw, max_blk_size, bcast_buffer, kernel_flops, kernel_time, &
! 0, current_local_n, i, last_stripe_width, &
! kernel)
! endif
!#endif /* COMPLEXCASE */
#endif /* WITH_OPENMP */ #endif /* WITH_OPENMP */
!send_b 1 !send_b 1
...@@ -1638,26 +1622,7 @@ ...@@ -1638,26 +1622,7 @@
current_local_n - bottom_msg_length, bottom_msg_length, i, & current_local_n - bottom_msg_length, bottom_msg_length, i, &
last_stripe_width, kernel) last_stripe_width, kernel)
!#if COMPLEXCASE == 1
!! the complex case and real case diverged here
! if (useGPU) then
! call compute_hh_trafo_complex_gpu_&
! &PRECISION&
! &(aIntern_dev, bcast_buffer_dev, hh_tau_dev, current_local_n -bottom_msg_length, bottom_msg_length, i, a_off, &
! dev_offset, dev_offset_1, dev_offset_2, &
! a_dim2, &
! kernel_flops, kernel_time, last_stripe_width, n_times, nbw, stripe_count, stripe_width)
! else
! call compute_hh_trafo_complex_&
! &PRECISION&
! &(aIntern, stripe_width, a_dim2, stripe_count, &
! a_off, nbw, max_blk_size, bcast_buffer, kernel_flops, kernel_time, &
! current_local_n - bottom_msg_length, bottom_msg_length, i, &
! last_stripe_width, kernel)
!
! endif
!
!#endif
!send_b !send_b
#ifdef WITH_MPI #ifdef WITH_MPI
...@@ -1745,23 +1710,7 @@ ...@@ -1745,23 +1710,7 @@
hh_tau_dev, kernel_flops, kernel_time, n_times, top_msg_length, & hh_tau_dev, kernel_flops, kernel_time, n_times, top_msg_length, &
current_local_n-top_msg_length-bottom_msg_length, i, & current_local_n-top_msg_length-bottom_msg_length, i, &
last_stripe_width, kernel) last_stripe_width, kernel)
!#if COMPLEXCASE == 1
! if (useGPU) then
! call compute_hh_trafo_complex_gpu_&
! &PRECISION&
! &(aIntern_dev, bcast_buffer_dev, hh_tau_dev, top_msg_length,current_local_n-top_msg_length-bottom_msg_length, &
! i, a_off, dev_offset, dev_offset_1, dev_offset_2, &
! a_dim2, &
! kernel_flops, kernel_time, last_stripe_width, n_times, nbw, stripe_count, stripe_width)
! else
! call compute_hh_trafo_complex_&
! &PRECISION&
! &(aIntern, stripe_width, a_dim2, stripe_count, &
! a_off, nbw, max_blk_size, bcast_buffer, kernel_flops, kernel_time, &
! top_msg_length, current_local_n-top_msg_length-bottom_msg_length, i, &
! last_stripe_width, kernel)
! endif
!#endif /* COMPLEXCASE */
#endif /* WITH_OPENMP */ #endif /* WITH_OPENMP */
!wait_t !wait_t
...@@ -1839,22 +1788,6 @@ ...@@ -1839,22 +1788,6 @@
#endif #endif
hh_tau_dev, kernel_flops, kernel_time, n_times, 0, top_msg_length, i, & hh_tau_dev, kernel_flops, kernel_time, n_times, 0, top_msg_length, i, &
last_stripe_width, kernel) last_stripe_width, kernel)
!#if COMPLEXCASE == 1
! if (useGPU) then
! call compute_hh_trafo_complex_gpu_&
! &PRECISION&
! &(aIntern_dev, bcast_buffer_dev, hh_tau_dev, 0, top_msg_length, i, a_off, dev_offset, dev_offset_1, dev_offset_2, &
! a_dim2, &
! kernel_flops, kernel_time, last_stripe_width, n_times, nbw, stripe_count, stripe_width)
! else
! call compute_hh_trafo_complex_&
! &PRECISION&
! &(aIntern, stripe_width, a_dim2, stripe_count, &
! a_off, nbw, max_blk_size, bcast_buffer, kernel_flops, kernel_time, &
! 0, top_msg_length, i, last_stripe_width, &
! kernel)
! endif
!#endif
#endif /* WITH_OPENMP */ #endif /* WITH_OPENMP */
endif endif
...@@ -2263,7 +2196,6 @@ ...@@ -2263,7 +2196,6 @@
#endif /* WITH_MPI */ #endif /* WITH_MPI */
if (useGPU) then if (useGPU) then
!#if REALCASE == 1
! copy q to q_dev needed in trans_ev_band_to_full ! copy q to q_dev needed in trans_ev_band_to_full
successCUDA = cuda_malloc(q_dev, ldq*matrixCols* size_of_datatype) successCUDA = cuda_malloc(q_dev, ldq*matrixCols* size_of_datatype)
if (.not.(successCUDA)) then if (.not.(successCUDA)) then
...@@ -2283,7 +2215,6 @@ ...@@ -2283,7 +2215,6 @@
stop 1 stop 1
endif endif
! endif ! endif
!#endif
endif !use GPU endif !use GPU
! deallocate all working space ! deallocate all working space
...@@ -2291,11 +2222,6 @@ ...@@ -2291,11 +2222,6 @@
if (.not.(useGPU)) then if (.not.(useGPU)) then
nullify(aIntern) nullify(aIntern)
call free(aIntern_ptr) call free(aIntern_ptr)
! deallocate(aIntern, stat=istat, errmsg=errorMessage)
! if (istat .ne. 0) then
! print *,"trans_ev_tridi_to_band_real: error when deallocating aIntern "//errorMessage
! stop 1
! endif
endif endif
deallocate(row, stat=istat, errmsg=errorMessage) deallocate(row, stat=istat, errmsg=errorMessage)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment