The stats you are reporting are figures from the view of the operating system, so it would be possible to see a BUSY_TIME that was much larger than then CPU time reported by Oracle in (say) the Top Timed Events. Depending on your platform you might see the same OSSTATS reporting sockets, cores and CPUs – although the titles may vary a little; the average busy time is total busy time divided by the number of CPUs (Personally, though, for CPU-intensive workloads I tend to worry more about CPU time / number of cores).

Also confirm if cpu time reported is over all time or time used by database ?

SQL> select distinct value from     v$osstat where stat_name='NUM_CPU_CORES';

     VALUE
----------
         2

1 row selected.

SQL>
SQL> SHOW PARAMETER CPU

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
cpu_count                            integer     4
parallel_threads_per_cpu             integer     2
resource_manager_cpu_allocation      integer     4
SQL>
SQL> select snap_id,max(busy) secs_all_cpu,max(avg_busy) avg_secs_percpu from
  2  (
  3  select snap_id,
  4  (case when stat_name = 'BUSY_TIME' then value/100 else null end) busy,
  5  (case when stat_name = 'AVG_BUSY_TIME' then value/100 else null end) avg_busy
  6  from dba_hist_osstat where snap_id> 3)
  7  where busy is not null or avg_busy is not null
  8  group by snap_id
  9  order by 2
 10  /

   SNAP_ID SECS_ALL_CPU AVG_SECS_PERCPU
---------- ------------ ---------------
         4       912.15          227.56
         5      2924.04          729.83
         6       4526.2         1129.49
         7       5519.8         1376.92
         8      8303.98          2072.1
         9     11594.98         2893.92
        10        16097         4018.53

7 rows selected.

SQL>