Juan 15/02/2023:Rules.LXnvhpc2202.mk, put ?= TP/CUDALEVEL/OPT_CPTCAP for tuning on Turpan ARM computer & compile with OPENACC mode_repro_sum.mod (or pressure not converging !!!)

Juan 01/02/2023:Rules.LXgfortran.mk, add MPIVIDE & openacc -foffload=-march=gfx90a for Siemens Compiler MPI250 Code Bench Sourcergy test

Juan 01/02/2023:ZSOLVER/../discretisation.f90, double size of NT1D buffer , to allow lev1 splitting in MG

Juan 01/02/2023:ZSOLVER/tridz.f90, BUG correction, Bypass SegFault for unnecessary computation for BSPLITTING ( of YSPLITTING)

Juan 01/02/2023:ZSOLVER/dotprod.f90, OpenACC OPT, remove update host(ZDOTPROD) + use SUM_DD_R2_ll_DEVICE

Juan 12/01/2023:configure, match login* for ADASTRA & correction for setting cce/15.0.0 et rocm/5.2.3

Juan 03/01/2023:communication.f90, Bypass MPICRAY MultiGrid problem , mpi_cart_create( reorder=.true. => .false. , to have expected rank number un sub-grid computation

Juan 23/12/2022:Correction of PB IO parallel, Rules.LX*,mode_io_field_write.f90: add -DMNH=MNH for problem with mnh_expand & MPI_REDUCE(...,2,...) two iresp return

Juan 17/12/2022:Rules.LXcray.mk , add "-hsystem_alloc" option , for compilation problem CCE/15.0.0 + ECCODES

Juan 16/12/2022:configure: add ADASTRA=login6 automati configure for default VER_MPI=MPICRAY & bypass openmpi-ucx-rocm VER_MPI=MPIAUTO

Juan 16/12/2022:ZSOLVER/* , GPU loop optimization in good order : DO CONCURRENT -> !$mnh_do_concurrent

Juan 16/12/2022:Makefile & CO:New spll_lst/spll_new command for 1 time compilation in master & user mode

Juan 16/11/2022:MNH/tridiag_wind.f90, Cray/14.X optimization , replace DO CONCURRENT with !$mnh_do_concurrent & put Z "acc seq loop" inside parallel 2D_XY one

Juan 16/11/2022:MNH/rain_ice_nucleation.f90,Nvhpc 22.X Bug Bypass , "ZZWMAX=MAXVAL()" dont parallelize any-more in GPU -> replace by "!$acc ... reduction(max:ZZWMAX)"

Juan 7/11/2022:Rules.LXcray.mk, add rule OPENACCNOBITREP, OpenACC without BitReproductibility ( for nvidia70 test )

Juan 04/11/2022:tools.f90, BYpass CCE > 14.X , "acc atomic" needed now "copy(ic)" to get corretly the ic value

Juan 04/11/2022:MNH/rain_ice_red.f90, ATTENTION , for CCE > 14.X, temporarely inhibe compilation of rain_ice_red, MORE THEN 30 minutes

Juan 04/11/2022:ice4_sedimentation_split.f90, Bypass for CCE > 14.X && nvidia70 target <=> TARGET_NV70 , use MODI_GAMMA

Juan 04/11/2022:br_transcendentals.cpp, BUG for CCE > 14.X && nvidia70 target, missing const in "#pragma omp declare target"

Juan 04/11/2022:mode_double_double.f90, CCE Optimization, declare SUM_DD_R2_ll_DEVICE & use "!$acc parallel" to avoid multiple kernel lunch in seq loop

Juan 04/11/2022:MNH/tridiag*.f90, CCE Optimization, use "!$acc parallel" to avoid multiple kernel lunch inside seq loop

Juan 04/11/2022:ZSOLVER/turb_hor_dyn_corr.f90 , CCE > 14.X Optimization , Inhibe BR_P2, not needed for bitreproductibility with Cray Compiler