diff --git a/src/post_process/m_data_input.f90 b/src/post_process/m_data_input.f90
index 9617766a2..19de8e9df 100644
--- a/src/post_process/m_data_input.f90
+++ b/src/post_process/m_data_input.f90
@@ -104,22 +104,27 @@ end subroutine s_setup_mpi_io_params
     !> Helper subroutine to read IB data files
     impure subroutine s_read_ib_data_files(file_loc_base, t_step)
 
-        character(len=*), intent(in)                :: file_loc_base
-        integer, intent(in), optional               :: t_step
-        character(LEN=len_trim(file_loc_base) + 20) :: file_loc
-        logical                                     :: file_exist
-        integer                                     :: ifile, ierr, data_size
+        character(len=*), intent(in)         :: file_loc_base
+        integer, intent(in), optional        :: t_step
+        character(LEN=path_len + 2*name_len) :: file_loc
+        logical                              :: file_exist
+        integer                              :: ifile, ierr, data_size
 
 #ifdef MFC_MPI
         integer, dimension(MPI_STATUS_SIZE) :: status
         integer(KIND=MPI_OFFSET_KIND)       :: disp
         integer(KIND=MPI_OFFSET_KIND)       :: m_MOK, n_MOK, p_MOK, MOK, WP_MOK, var_MOK
         integer                             :: save_index
+        character(len=10)                   :: t_step_string
 #endif
 
         if (.not. ib) return
 
-        if (parallel_io) then
+        if (parallel_io .and. file_per_process) then
+            call s_int_to_str(t_step, t_step_string)
+            write (file_loc, '(A,I0,A,i7.7,A)') 'ib_markers_', t_step, '_', proc_rank, '.dat'
+            file_loc = trim(case_dir) // '/restart_data/lustre_' // trim(t_step_string) // '/' // trim(file_loc)
+        else if (parallel_io) then
             write (file_loc, '(A)') trim(file_loc_base) // 'ib.dat'
         else
             write (file_loc, '(A)') trim(file_loc_base) // '/ib_data.dat'
@@ -127,7 +132,17 @@ impure subroutine s_read_ib_data_files(file_loc_base, t_step)
         inquire (FILE=trim(file_loc), EXIST=file_exist)
 
         if (file_exist) then
-            if (parallel_io) then
+            if (parallel_io .and. file_per_process) then
+#ifdef MFC_MPI
+                call MPI_FILE_OPEN(MPI_COMM_SELF, file_loc, MPI_MODE_RDONLY, mpi_info_int, ifile, ierr)
+
+                data_size = (m + 1)*(n + 1)*(p + 1)
+
+                call MPI_FILE_READ(ifile, MPI_IO_IB_DATA%var%sf, data_size, MPI_INTEGER, status, ierr)
+
+                call MPI_FILE_CLOSE(ifile, ierr)
+#endif
+            else if (parallel_io) then
 #ifdef MFC_MPI
                 call MPI_FILE_OPEN(MPI_COMM_WORLD, file_loc, MPI_MODE_RDONLY, mpi_info_int, ifile, ierr)
 
diff --git a/src/simulation/m_data_output.fpp b/src/simulation/m_data_output.fpp
index 6c923b334..239d8df08 100644
--- a/src/simulation/m_data_output.fpp
+++ b/src/simulation/m_data_output.fpp
@@ -880,6 +880,7 @@ contains
         integer(kind=MPI_OFFSET_kind)        :: WP_MOK, var_MOK, MOK
         integer                              :: ifile, ierr, data_size
         integer, dimension(MPI_STATUS_SIZE)  :: status
+        character(len=10)                    :: t_step_string
 
         $:GPU_UPDATE(host='[ib_markers%sf]')
 
@@ -890,21 +891,39 @@ contains
         WP_MOK = int(storage_size(0._stp)/8, MPI_OFFSET_KIND)
         MOK = int(1._wp, MPI_OFFSET_KIND)
 
-        write (file_loc, '(A)') 'ib.dat'
-        file_loc = trim(case_dir) // '/restart_data' // trim(mpiiofs) // trim(file_loc)
+        if (file_per_process) then
+            call s_int_to_str(time_step, t_step_string)
 
-        call s_mpi_barrier()
-        call s_delay_file_access(proc_rank)
+            if (proc_rank == 0) then
+                file_loc = trim(case_dir) // '/restart_data/lustre_' // trim(t_step_string)
+                call s_create_directory(trim(file_loc))
+            end if
+            call s_mpi_barrier()
+            call s_delay_file_access(proc_rank)
+
+            write (file_loc, '(A,I0,A,i7.7,A)') 'ib_markers_', time_step, '_', proc_rank, '.dat'
+            file_loc = trim(case_dir) // '/restart_data/lustre_' // trim(t_step_string) // '/' // trim(file_loc)
 
-        call MPI_FILE_OPEN(MPI_COMM_WORLD, file_loc, ior(MPI_MODE_WRONLY, MPI_MODE_CREATE), mpi_info_int, ifile, ierr)
+            call MPI_FILE_OPEN(MPI_COMM_SELF, file_loc, ior(MPI_MODE_WRONLY, MPI_MODE_CREATE), mpi_info_int, ifile, ierr)
+            call MPI_FILE_WRITE_ALL(ifile, MPI_IO_IB_DATA%var%sf, data_size, MPI_INTEGER, status, ierr)
+            call MPI_FILE_CLOSE(ifile, ierr)
+        else
+            write (file_loc, '(A)') 'ib.dat'
+            file_loc = trim(case_dir) // '/restart_data' // trim(mpiiofs) // trim(file_loc)
 
-        var_MOK = int(sys_size + 1, MPI_OFFSET_KIND)
-        disp = m_MOK*max(MOK, n_MOK)*max(MOK, p_MOK)*WP_MOK*(var_MOK - 1 + int(time_step/t_step_save))
-        if (time_step == 0) disp = 0
+            call s_mpi_barrier()
+            call s_delay_file_access(proc_rank)
 
-        call MPI_FILE_SET_VIEW(ifile, disp, MPI_INTEGER, MPI_IO_IB_DATA%view, 'native', mpi_info_int, ierr)
-        call MPI_FILE_WRITE_ALL(ifile, MPI_IO_IB_DATA%var%sf, data_size, MPI_INTEGER, status, ierr)
-        call MPI_FILE_CLOSE(ifile, ierr)
+            call MPI_FILE_OPEN(MPI_COMM_WORLD, file_loc, ior(MPI_MODE_WRONLY, MPI_MODE_CREATE), mpi_info_int, ifile, ierr)
+
+            var_MOK = int(sys_size + 1, MPI_OFFSET_KIND)
+            disp = m_MOK*max(MOK, n_MOK)*max(MOK, p_MOK)*WP_MOK*(var_MOK - 1 + int(time_step/t_step_save))
+            if (time_step == 0) disp = 0
+
+            call MPI_FILE_SET_VIEW(ifile, disp, MPI_INTEGER, MPI_IO_IB_DATA%view, 'native', mpi_info_int, ierr)
+            call MPI_FILE_WRITE_ALL(ifile, MPI_IO_IB_DATA%var%sf, data_size, MPI_INTEGER, status, ierr)
+            call MPI_FILE_CLOSE(ifile, ierr)
+        end if
 #endif
 
     end subroutine s_write_parallel_ib_data