WRF¶
简介¶
WRF(Weather Research and Forecasting Model)模式是有美国环境预报中心(NCEP), 美国国家大气研究中心（NCAR）以及多个大学、研究所和业务部门联合研发的一种统一的中尺度天气预报模式。 WRF模式适用范围很广，从中小尺度到全球尺度的数值预报和模拟都有广泛的应用。
WPS是预处理WRF运行数据的工具。
可用版本¶
版本 |
平台 |
构建方式 |
模块名 |
|---|---|---|---|
4.2.1 |
|
源码 |
wrf/4.2.1-oneapi-2021.4.0 思源一号 |
4.5.1 |
|
源码 |
wrf/4.5.1-oneapi-2021.4.0 思源一号 |
4.2 |
|
源码 |
wps/4.2-oneapi-2021.4.0 思源一号 |
4.5 |
|
源码 |
wps/4.5-oneapi-2021.4.0 思源一号 |
4.3.1 |
|
源码 |
wrf/4.3.1-intel-2021.4.0 Pi2.0 |
4.5.1 |
|
源码 |
wrf/4.5.1-intel-2021.4.0 Pi2.0 |
4.4 |
|
源码 |
wps/4.4-intel-2021.4.0 Pi2.0 |
4.5 |
|
源码 |
wps/4.5-intel-2021.4.0 Pi2.0 |
4.5.1 |
|
源码 |
wrf/4.5.1-gcc-10.3.0-openmpi ARM |
4.5 |
|
源码 |
wps/4.5-gcc-10.3.1 ARM |
算例位置¶
思源一号 : /dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/wrf-4.2/wrf_data
π2.0 : /lustre/opt/contribute/cascadelake/wrf_cmaq/wrf_data
ARM : 同π2.0
算例目录
[hpc@node234 wrf-4.2]$ tree wrf_data/
wrf_data/
├── fnl_20161006_00_00.grib2
├── fnl_20161006_06_00.grib2
├── fnl_20161006_12_00.grib2
├── fnl_20161006_18_00.grib2
├── fnl_20161007_00_00.grib2
├── fnl_20161007_06_00.grib2
├── fnl_20161007_12_00.grib2
├── fnl_20161007_18_00.grib2
└── fnl_20161008_00_00.grib2
思源一号和π2.0两个集群上的数据均是模拟2016年10月06日00点至2016年10月08日0点的气象数据
geog_data_path的位置¶
思源一号 : /dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/geo/geog
π2.0 : /lustre/opt/contribute/cascadelake/wrf_cmaq/geo
集群上的WRF和WPS¶
思源一号上的WRF和WPS¶
自定义编译WRF和WPS¶
思源一号上已部署所依赖的库及版本¶
hdf5-1.10.5
libpng-1.6.37
netcdf-c-4.8.0
netcdf-fortran-4.5.3
zlib-1.2.13
jasper-1.900.1
自定义编译WRF¶
申请计算节点用于编译
srun -n 20 -p 64c512g --pty /bin/bash
编译WRF前导入所需的环境变量
module load oneapi
export CC=icc
export FC=ifort
export F90=ifort
export CXX=icpc
export CPP="icpc -E"
export DIR=/dssg/share/sample/wrf/WRF_4.1.1_Intel/wrflibs
export PATH=$DIR/netcdf/bin:$DIR/hdf5-1.10.5/bin:$DIR/libpng-1.6.37/bin:$DIR/zlib-1.2.13/bin:$PATH
export LD_LIBRARY_PATH=$DIR/netcdf/lib:$DIR/hdf5-1.10.5/lib:$DIR/libpng-1.6.37/lib64:$DIR/zlib-1.2.13/lib:$LD_LIBRARY_PATH
export NETCDF=$DIR/netcdf
export HDF5=$DIR/hdf5-1.10.5
拷贝WRF源码并进行configure配置选择
cp /dssg/share/sample/wrf/WRF_4.1.1_Intel/src/v4.2.2.tar.gz .
tar xvf v4.2.2.tar.gz
cd WRF-4.2.2/
./configure
可根据所需选择相应的参数，使用intel编译器编译WRF，可以选择15，可以多节点并行运行。
Please select from among the following Linux x86_64 options:
1. (serial) 2. (smpar) 3. (dmpar) 4. (dm+sm) PGI (pgf90/gcc)
2. (serial) 6. (smpar) 7. (dmpar) 8. (dm+sm) PGI (pgf90/pgcc): SGI MPT
3. (serial) 10. (smpar) 11. (dmpar) 12. (dm+sm) PGI (pgf90/gcc): PGI accelerator
1. (serial) 14. (smpar) 15. (dmpar) 16. (dm+sm) INTEL (ifort/icc)
1. (dm+sm) INTEL (ifort/icc): Xeon Phi (MIC architecture)
2. (serial) 19. (smpar) 20. (dmpar) 21. (dm+sm) INTEL (ifort/icc): Xeon (SNB with AVX mods)
3. (serial) 23. (smpar) 24. (dmpar) 25. (dm+sm) INTEL (ifort/icc): SGI MPT
4. (serial) 27. (smpar) 28. (dmpar) 29. (dm+sm) INTEL (ifort/icc): IBM POE
5. (serial) 31. (dmpar) PATHSCALE (pathf90/pathcc)
6. (serial) 33. (smpar) 34. (dmpar) 35. (dm+sm) GNU (gfortran/gcc)
7. (serial) 37. (smpar) 38. (dmpar) 39. (dm+sm) IBM (xlf90_r/cc_r)
8. (serial) 41. (smpar) 42. (dmpar) 43. (dm+sm) PGI (ftn/gcc): Cray XC CLE
9. (serial) 45. (smpar) 46. (dmpar) 47. (dm+sm) CRAY CCE (ftn $(NOOMP)/cc): Cray XE and XC
10. (serial) 49. (smpar) 50. (dmpar) 51. (dm+sm) INTEL (ftn/icc): Cray XC
11. (serial) 53. (smpar) 54. (dmpar) 55. (dm+sm) PGI (pgf90/pgcc)
12. (serial) 57. (smpar) 58. (dmpar) 59. (dm+sm) PGI (pgf90/gcc): -f90=pgf90
13. (serial) 61. (smpar) 62. (dmpar) 63. (dm+sm) PGI (pgf90/pgcc): -f90=pgf90
14. (serial) 65. (smpar) 66. (dmpar) 67. (dm+sm) INTEL (ifort/icc): HSW/BDW
15. (serial) 69. (smpar) 70. (dmpar) 71. (dm+sm) INTEL (ifort/icc): KNL MIC
16. (serial) 73. (smpar) 74. (dmpar) 75. (dm+sm) FUJITSU (frtpx/fccpx): FX10/FX100 SPARC64 IXfx/Xlfx
修改configure.wrf
DM_FC = mpiifort
DM_CC = mpiicc
编译em_real
./compile em_real 2>&1 | tee -a compile.log
编译完成输出
---> Executables successfully built <---
-rwxrwxr-x 1 hpcxdy hpcxdy 48398000 Oct 22 10:19 main/ndown.exe
-rwxrwxr-x 1 hpcxdy hpcxdy 48388496 Oct 22 10:19 main/real.exe
-rwxrwxr-x 1 hpcxdy hpcxdy 47734360 Oct 22 10:19 main/tc.exe
-rwxrwxr-x 1 hpcxdy hpcxdy 52544272 Oct 22 10:18 main/wrf.exe
==========================================================================
自定义编译WPS¶
导入如下环境变量
export WRF_DIR=../WRF-4.2.2/
export JASPERLIB=$DIR/jasper-1.900.1/lib/
export JASPERINC=$DIR/jasper-1.900.1/include/
拷贝WPS源码
cp /dssg/share/sample/wrf/src/v4.2.tar.gz .
tar xvf v4.2.tar.gz
cd WPS-4.2/
./configure
根据个人所需选择所需版本，选择20多节点并行运行。
Please select from among the following supported platforms.
1. Linux x86_64, gfortran (serial)
2. Linux x86_64, gfortran (serial_NO_GRIB2)
3. Linux x86_64, gfortran (dmpar)
4. Linux x86_64, gfortran (dmpar_NO_GRIB2)
5. Linux x86_64, PGI compiler (serial)
6. Linux x86_64, PGI compiler (serial_NO_GRIB2)
7. Linux x86_64, PGI compiler (dmpar)
8. Linux x86_64, PGI compiler (dmpar_NO_GRIB2)
9. Linux x86_64, PGI compiler, SGI MPT (serial)
10. Linux x86_64, PGI compiler, SGI MPT (serial_NO_GRIB2)
11. Linux x86_64, PGI compiler, SGI MPT (dmpar)
12. Linux x86_64, PGI compiler, SGI MPT (dmpar_NO_GRIB2)
13. Linux x86_64, IA64 and Opteron (serial)
14. Linux x86_64, IA64 and Opteron (serial_NO_GRIB2)
15. Linux x86_64, IA64 and Opteron (dmpar)
16. Linux x86_64, IA64 and Opteron (dmpar_NO_GRIB2)
17. Linux x86_64, Intel compiler (serial)
18. Linux x86_64, Intel compiler (serial_NO_GRIB2)
19. Linux x86_64, Intel compiler (dmpar)
20. Linux x86_64, Intel compiler (dmpar_NO_GRIB2)
21. Linux x86_64, Intel compiler, SGI MPT (serial)
22. Linux x86_64, Intel compiler, SGI MPT (serial_NO_GRIB2)
23. Linux x86_64, Intel compiler, SGI MPT (dmpar)
24. Linux x86_64, Intel compiler, SGI MPT (dmpar_NO_GRIB2)
25. Linux x86_64, Intel compiler, IBM POE (serial)
26. Linux x86_64, Intel compiler, IBM POE (serial_NO_GRIB2)
27. Linux x86_64, Intel compiler, IBM POE (dmpar)
28. Linux x86_64, Intel compiler, IBM POE (dmpar_NO_GRIB2)
29. Linux x86_64 g95 compiler (serial)
30. Linux x86_64 g95 compiler (serial_NO_GRIB2)
31. Linux x86_64 g95 compiler (dmpar)
32. Linux x86_64 g95 compiler (dmpar_NO_GRIB2)
33. Cray XE/XC CLE/Linux x86_64, Cray compiler (serial)
34. Cray XE/XC CLE/Linux x86_64, Cray compiler (serial_NO_GRIB2)
35. Cray XE/XC CLE/Linux x86_64, Cray compiler (dmpar)
36. Cray XE/XC CLE/Linux x86_64, Cray compiler (dmpar_NO_GRIB2)
37. Cray XC CLE/Linux x86_64, Intel compiler (serial)
38. Cray XC CLE/Linux x86_64, Intel compiler (serial_NO_GRIB2)
39. Cray XC CLE/Linux x86_64, Intel compiler (dmpar)
40. Cray XC CLE/Linux x86_64, Intel compiler (dmpar_NO_GRIB2)
修改configure.wps
DM_FC = mpiifort
DM_CC = mpiicc
编译安装
./compile
思源一号上使用预编译的WRF和WPS¶
先用WPS处理数据¶
由于WPS处理数据需要复杂的文件依赖关系，可先拷贝WPS目录中的文件到本地
mkdir ~/data && cd ~/data
mkdir WRF && cd WRF
cp -r /dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/wrf-4.2/WPS-4.2 ./
拷贝数据到WPS目录中进行数据处理
cd WPS-4.2
cp -r /dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/wrf-4.2/wrf_data/* ./
namelist.wps文件内容设置如下：
&share
wrf_core = 'ARW',
max_dom = 1,
start_date = '2016-10-06_00:00:00'
end_date = '2016-10-08_00:00:00'
interval_seconds = 21600
io_form_geogrid = 2,
/
&geogrid
parent_id = 1,
parent_grid_ratio = 1,
i_parent_start = 1,
j_parent_start = 1,
e_we = 515,
e_sn = 515,
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!! IMPORTANT NOTE !!!!!!!!!!!!!!!!!!!!!!!!!!!!
! The default datasets used to produce the MAXSNOALB and ALBEDO12M
! fields have changed in WPS v4.0. These fields are now interpolated
! from MODIS-based datasets.
!
! To match the output given by the default namelist.wps in WPS v3.9.1,
! the following setting for geog_data_res may be used:
!
! geog_data_res = 'maxsnowalb_ncep+albedo_ncep+default', 'maxsnowalb_ncep+albedo_ncep+default',
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!! IMPORTANT NOTE !!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
geog_data_res = 'default','default',
dx = 12000,
dy = 12000,
map_proj = 'lambert',
ref_lat = 31.00,
ref_lon = 120.00,
ref_x = 351
ref_y = 208
truelat1 = 30.0,
truelat2 = 60.0,
stand_lon = 120.0,
geog_data_path = '/dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/geo/geog/'
/
&ungrib
out_format = 'WPS',
prefix = 'FILE',
/
&metgrid
fg_name = 'FILE'
io_form_metgrid = 2,
/
运行geogrid.exe程序定义模型投影、区域范围，嵌套关系，对地表参数进行插值。
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 1
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wps
geogrid.exe
5.根据模拟时期选择文件
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 1
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wps
link_grib.csh fnl_2016100*
cp ungrib/Variable_Tables/Vtable.GFS Vtable
6.从grib数据中提取所需要的气象参数
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 1
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wps
ungrib.exe
7.将气象参数插值到模拟区域
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 1
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wps
metgrid.exe
WRF运行¶
由于WRF运行数据需要复杂的文件依赖关系，可先拷贝WRF目录中必要的文件到本地
cd ~/data
cd WRF
mkdir WRF-4.2.1 && cd WRF-4.2.1
cp -r /dssg/opt/icelake/linux-centos8-icelake/oneapi-2021.4.0/wrf_cmaq/wrf-4.2/WRF-4.2.1/run/* ./
拷贝WPS生成的met文件到WRF-4.2.1目录
cp -r ../WPS-4.2/met_em.d01.2016-10-0* ./
namelist.input文件内容设置如下，参数需要与wps的namelist.wps参数一致：
&time_control
run_days = 2,
run_hours = 0,
run_minutes = 0,
run_seconds = 0,
start_year = 2016,
start_month = 10,
start_day = 06,
start_hour = 00,
end_year = 2016,
end_month = 10,
end_day = 08,
end_hour = 00,
interval_seconds = 21600
input_from_file = .true.,.true.,
history_interval = 60, 60,
frames_per_outfile = 12, 12,
restart = .false.,
restart_interval = 5000,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
/
&domains
time_step = 60,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 1,
e_we = 515, 112,
e_sn = 515, 97,
e_vert = 33, 33,
p_top_requested = 5000,
num_metgrid_levels = 32,
num_metgrid_soil_levels = 4,
dx = 12000,
dy = 12000,
grid_id = 1, 2,
parent_id = 0, 1,
i_parent_start = 1, 31,
j_parent_start = 1, 17,
parent_grid_ratio = 1, 3,
parent_time_step_ratio = 1, 3,
feedback = 1,
smooth_option = 0
/
&physics
physics_suite = 'tropical'
mp_physics = 6, -1,
cu_physics = 16, -1,
ra_lw_physics = 4, -1,
ra_sw_physics = 4, -1,
bl_pbl_physics = 8, 8,
sf_sfclay_physics = 1, 1,
sf_surface_physics = 2, -1,
radt = 12, 30,
bldt = 0, 0,
cudt = 5, 5,
icloud = 1,
num_land_cat = 21,
sf_urban_physics = 0, 0, 0,
/
&fdda
/
&dynamics
hybrid_opt = 2,
w_damping = 0,
diff_opt = 1, 1,
km_opt = 4, 4,
diff_6th_opt = 0, 0,
diff_6th_factor = 0.12, 0.12,
base_temp = 290.
damp_opt = 3,
zdamp = 5000., 5000.,
dampcoef = 0.2, 0.2,
khdif = 0, 0,
kvdif = 0, 0,
non_hydrostatic = .true., .true.,
moist_adv_opt = 1, 1,
scalar_adv_opt = 1, 1,
gwd_opt = 0, 1,
/
&bdy_control
spec_bdy_width = 5,
specified = .true.
/
&grib2
/
&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/
运行real.exe程序，脚本如下：
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 1
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wrf
ulimit -s unlimited
real.exe
运行wrf.exe程序，脚本如下，该部分是最终也是最耗时的执行程序。
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=64c512g
#SBATCH -N 4
#SBATCH --ntasks-per-node=64
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load oneapi
module load wrf
ulimit -s unlimited
mpirun wrf.exe
π2.0上的WRF和WPS¶
π2.0上先用WPS处理数据¶
由于WPS处理数据需要复杂的文件依赖关系，可先拷贝WPS目录中的文件到本地
mkdir ~/data && cd ~/data
mkdir WRF && cd WRF
cp -r /lustre/opt/contribute/cascadelake/wrf_cmaq/packet_1/WPS-4.3.1 ./
拷贝数据到WPS目录中进行数据处理
cd WPS-4.3.1
cp -r /lustre/opt/contribute/cascadelake/wrf_cmaq/wrf_data/* ./
namelist.wps文件内容设置如下：
&share
wrf_core = 'ARW',
max_dom = 1,
start_date = '2016-10-06_00:00:00'
end_date = '2016-10-08_00:00:00'
interval_seconds = 21600
io_form_geogrid = 2,
/
&geogrid
parent_id = 1,
parent_grid_ratio = 1,
i_parent_start = 1,
j_parent_start = 1,
e_we = 515,
e_sn = 515,
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!! IMPORTANT NOTE !!!!!!!!!!!!!!!!!!!!!!!!!!!!
! The default datasets used to produce the MAXSNOALB and ALBEDO12M
! fields have changed in WPS v4.0. These fields are now interpolated
! from MODIS-based datasets.
!
! To match the output given by the default namelist.wps in WPS v3.9.1,
! the following setting for geog_data_res may be used:
!
! geog_data_res = 'maxsnowalb_ncep+albedo_ncep+default', 'maxsnowalb_ncep+albedo_ncep+default',
!
!!!!!!!!!!!!!!!!!!!!!!!!!!!! IMPORTANT NOTE !!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
geog_data_res = 'default','default',
dx = 12000,
dy = 12000,
map_proj = 'lambert',
ref_lat = 31.00,
ref_lon = 120.00,
ref_x = 351
ref_y = 208
truelat1 = 30.0,
truelat2 = 60.0,
stand_lon = 120.0,
geog_data_path = '/lustre/opt/contribute/cascadelake/wrf_cmaq/geo/'
/
&ungrib
out_format = 'WPS',
prefix = 'FILE',
/
&metgrid
fg_name = 'FILE'
io_form_metgrid = 2,
/
运行geogrid.exe程序定义模型投影、区域范围，嵌套关系，对地表参数进行插值。
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
geogrid.exe
5.根据模拟时期选择文件
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
link_grib.csh fnl_2016100*
cp ungrib/Variable_Tables/Vtable.GFS Vtable
6.从grib数据中提取所需要的气象参数
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
ungrib.exe
7.将气象参数插值到模拟区域
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
metgrid.exe
π2.0上运行WRF¶
由于WRF运行数据需要复杂的文件依赖关系，可先拷贝WRF目录中必要的文件到本地
cd ~/data
cd WRF
mkdir WRF-4.3.1 && cd WRF-4.3.1
cp -r /lustre/opt/contribute/cascadelake/wrf_cmaq/packet_1/WRF-master/run/* ./
拷贝WPS生成的met文件到WRF-4.3.1目录
cp -r ~/data/WRF/WPS-4.3.1/met_em.d* ./
namelist.input文件内容设置如下，参数需要与wps的namelist.wps参数一致：
&time_control
run_days = 2,
run_hours = 0,
run_minutes = 0,
run_seconds = 0,
start_year = 2016,
start_month = 10,
start_day = 06,
start_hour = 00,
end_year = 2016,
end_month = 10,
end_day = 08,
end_hour = 00,
interval_seconds = 21600
input_from_file = .true.,.true.,
history_interval = 60, 60,
frames_per_outfile = 12, 12,
restart = .false.,
restart_interval = 5000,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
/
&domains
time_step = 60,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 1,
e_we = 515, 112,
e_sn = 515, 97,
e_vert = 33, 33,
p_top_requested = 5000,
num_metgrid_levels = 32,
num_metgrid_soil_levels = 4,
dx = 12000,
dy = 12000,
grid_id = 1, 2,
parent_id = 0, 1,
i_parent_start = 1, 31,
j_parent_start = 1, 17,
parent_grid_ratio = 1, 3,
parent_time_step_ratio = 1, 3,
feedback = 1,
smooth_option = 0
/
&physics
physics_suite = 'tropical'
mp_physics = 6, -1,
cu_physics = 16, -1,
ra_lw_physics = 4, -1,
ra_sw_physics = 4, -1,
bl_pbl_physics = 8, 8,
sf_sfclay_physics = 1, 1,
sf_surface_physics = 2, -1,
radt = 12, 30,
bldt = 0, 0,
cudt = 5, 5,
icloud = 1,
num_land_cat = 21,
sf_urban_physics = 0, 0, 0,
/
&fdda
/
&dynamics
hybrid_opt = 2,
w_damping = 0,
diff_opt = 1, 1,
km_opt = 4, 4,
diff_6th_opt = 0, 0,
diff_6th_factor = 0.12, 0.12,
base_temp = 290.
damp_opt = 3,
zdamp = 5000., 5000.,
dampcoef = 0.2, 0.2,
khdif = 0, 0,
kvdif = 0, 0,
non_hydrostatic = .true., .true.,
moist_adv_opt = 1, 1,
scalar_adv_opt = 1, 1,
gwd_opt = 0, 1,
/
&bdy_control
spec_bdy_width = 5,
specified = .true.
/
&grib2
/
&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/
运行real.exe程序，脚本如下：
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
ulimit -s unlimited
real.exe
运行wrf.exe程序，脚本如下，该部分是最终也是最耗时的执行程序。
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=cpu
#SBATCH -N 1
#SBATCH --ntasks-per-node=40
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.3.1-intel-2021.4.0
ulimit -s unlimited
mpirun wrf.exe
ARM上运行WRF¶
由于WRF运行数据需要复杂的文件依赖关系，可先拷贝WRF目录中必要的文件到本地
cd ~/data
cd WRF
mkdir WRF-4.3.1 && cd WRF-4.3.1
cp -r /lustre/opt/contribute/cascadelake/wrf_cmaq/packet_1/WRF-master/run/* ./
拷贝WPS生成的met文件到WRF-4.3.1目录
cp -r ~/data/WRF/WPS-4.3.1/met_em.d* ./
namelist.input文件内容设置如下，参数需要与wps的namelist.wps参数一致：
&time_control
run_days = 2,
run_hours = 0,
run_minutes = 0,
run_seconds = 0,
start_year = 2016,
start_month = 10,
start_day = 06,
start_hour = 00,
end_year = 2016,
end_month = 10,
end_day = 08,
end_hour = 00,
interval_seconds = 21600
input_from_file = .true.,.true.,
history_interval = 60, 60,
frames_per_outfile = 12, 12,
restart = .false.,
restart_interval = 5000,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
/
&domains
time_step = 60,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 1,
e_we = 515, 112,
e_sn = 515, 97,
e_vert = 33, 33,
p_top_requested = 5000,
num_metgrid_levels = 32,
num_metgrid_soil_levels = 4,
dx = 12000,
dy = 12000,
grid_id = 1, 2,
parent_id = 0, 1,
i_parent_start = 1, 31,
j_parent_start = 1, 17,
parent_grid_ratio = 1, 3,
parent_time_step_ratio = 1, 3,
feedback = 1,
smooth_option = 0
/
&physics
physics_suite = 'tropical'
mp_physics = 6, -1,
cu_physics = 16, -1,
ra_lw_physics = 4, -1,
ra_sw_physics = 4, -1,
bl_pbl_physics = 8, 8,
sf_sfclay_physics = 1, 1,
sf_surface_physics = 2, -1,
radt = 12, 30,
bldt = 0, 0,
cudt = 5, 5,
icloud = 1,
num_land_cat = 21,
sf_urban_physics = 0, 0, 0,
/
&fdda
/
&dynamics
hybrid_opt = 2,
w_damping = 0,
diff_opt = 1, 1,
km_opt = 4, 4,
diff_6th_opt = 0, 0,
diff_6th_factor = 0.12, 0.12,
base_temp = 290.
damp_opt = 3,
zdamp = 5000., 5000.,
dampcoef = 0.2, 0.2,
khdif = 0, 0,
kvdif = 0, 0,
non_hydrostatic = .true., .true.,
moist_adv_opt = 1, 1,
scalar_adv_opt = 1, 1,
gwd_opt = 0, 1,
/
&bdy_control
spec_bdy_width = 5,
specified = .true.
/
&grib2
/
&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/
运行real.exe程序，脚本如下：
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=arm128c256g
#SBATCH -N 1
#SBATCH --ntasks-per-node=128
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.5.1-gcc-10.3.1
ulimit -s unlimited
real.exe
运行wrf.exe程序，脚本如下，该部分是最终也是最耗时的执行程序。
#!/bin/bash
#SBATCH --job-name=test
#SBATCH --partition=arm128c256g
#SBATCH -N 1
#SBATCH --ntasks-per-node=128
#SBATCH --output=%j.out
#SBATCH --error=%j.err
module load wrf/4.5.1-gcc-10.3.1
ulimit -s unlimited
mpirun wrf.exe
运行结果(单位为：秒，越低越好)¶
思源一号上WRF的运行时间¶
wrf/4.2.1-oneapi-2021.4.0 |
|||
|---|---|---|---|
核数 |
64 |
128 |
256 |
Exec time |
0:36:21 |
0:18:05 |
0:10:44 |
π2.0上WRF的运行时间¶
wrf/4.3.1-intel-2021.4.0 |
|||
|---|---|---|---|
核数 |
40 |
80 |
160 |
Exec time |
1:10:28 |
0:42:22 |
0:26:01 |
ARM上WRF的运行时间¶
wrf/4.5.1-gcc-10.3.1 |
|||
|---|---|---|---|
核数 |
64 |
128 |
256 |
Exec time |
1:48:51 |
1:00:02 |
0:50:49 |
参考资料¶
最后更新:
2024 年 11 月 22 日