jgp@cscs_ch
New Member
Dear CCSM3 experts
I am trying to install CCSM (ccsm3_0_1_beta29) on a LinuxOnPOWER5 IBM cluster using xlc/xlf compilers :
* release SUSE_10_2
* xlf 11_01_0000_0003
* xlc 09_00_0000_0000 IBM XL C/C++ Advanced Edition for Linux, V9.0
* unames Linux ppc64
* unamer 2_6_16_600_25_PTF_411256_0ppc64
I have some difficulties building the lnd model : the following compilation command
will give the following error msg :
A temporary fix is to modify /apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c :
Anyhow, I do not believe it should be kept as it is...
Is it the right forum to ask ? Do you have experience with Linux on PWR ?
Any suggestions or previous experiences will be greatly appreciated.
Thanks in advance.
JG
I am trying to install CCSM (ccsm3_0_1_beta29) on a LinuxOnPOWER5 IBM cluster using xlc/xlf compilers :
* release SUSE_10_2
* xlf 11_01_0000_0003
* xlc 09_00_0000_0000 IBM XL C/C++ Advanced Edition for Linux, V9.0
* unames Linux ppc64
* unamer 2_6_16_600_25_PTF_411256_0ppc64
I have some difficulties building the lnd model : the following compilation command
Code:
17:06:42 @blanc161:/scratch/piccinal/ccsm3/lnd $
mpcc -q64 -qarch=pwr5 -qtune=pwr5
-I.
-I/apps/netcdf/4.0.0/xlf11/64/include
-I/scratch/piccinal/ccsm3/lib/include
-I/apps/sandbox/ccsm/ccsm3/SourceMods/src.share
-I/apps/sandbox/ccsm/ccsm3/SourceMods/src.clm
-I/apps/sandbox/ccsm/ccsm3_/models/lnd/clm2/src/main
-I/apps/sandbox/ccsm/ccsm3_/models/lnd/clm2/src/biogeophys
-I/apps/sandbox/ccsm/ccsm3_/models/lnd/clm2/src/riverroute
-I/apps/sandbox/ccsm/ccsm3_/models/lnd/clm2/src/biogeochem
-I/apps/sandbox/ccsm/ccsm3_/models/csm_share/cpl
-I/apps/sandbox/ccsm/ccsm3_/models/csm_share/shr
-I/apps/sandbox/ccsm/ccsm3_/models/utils/timing
-I/apps/sandbox/ccsm/ccsm3_/models/utils/esmf_wrf_timemgr
-Wp,-DLINUX -Wp,-DNO_SHR_VMATH
-Wp,-Dcoupled -Wp,-DNPROC_X= -Wp,-DNPROC_Y=
-Wp,-D_MPI
-c /apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c
will give the following error msg :
Code:
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 79.17: 1506-045 (S) Undeclared identifier OptionName.
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 79.39: 1506-045 (S) Undeclared identifier Boolean.
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 115.14: 1506-045 (S) Undeclared identifier MAX_CHARS.
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 115.14: 1506-1324 (S) Array size must have integer type.
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 126.14: 1506-045 (S) Undeclared identifier MAX_CHARS.
* "/apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c", line 126.14: 1506-1324 (S) Array size must have integer type.
A temporary fix is to modify /apps/sandbox/ccsm/ccsm3_/models/utils/timing/f_wrappers.c :
Code:
12c12,17
< #include <gpt.h>
---
> #include <string.h>
> #include <stdlib.h>
> //#include <private.h>
> #define MAX_CHARS 15
> #include <gptl.h>
76c81,82
< t_setoption ( (OptionName) *option, (Boolean) *val);
---
> t_setoption ( (int) *option, *val);
Anyhow, I do not believe it should be kept as it is...
Is it the right forum to ask ? Do you have experience with Linux on PWR ?
Any suggestions or previous experiences will be greatly appreciated.
Thanks in advance.
JG