Add the total and compute times on amd

GYSELAX/testcase_gpu/README.md

@@ -7,11 +7,15 @@ Case\_MPI8 on 1 nodes:
  - 8 MPI tasks
  - 8 MPI tasks per node  
  - 14.7 Go per MPI task
+ - Total time:  957.56 s
+ - Total time (without init & diag): 900.44 s
  
 Case\_MPI64 on 8 nodes:
  - 64 MPI tasks
  - 8 MPI per node  
  - 13.7 Go per MPI task
+ - Total time:  325.34 s
+ - Total time (without init & diag): 230.47 s
 
 The total time and the compute time (total time without initialization and diagnostics) are the two key indicators for the porting on GPU.