Error when running wrf.exe

[sher]

Hello everyone, I am using the WRF-Solar version 4.5. When I run wrf.exe, the simulation stops after only a few minutes. I did not find any error messages in the rsl.error files. I would like to know what might have caused this issue.
After some adjustments, I managed to run the simulation successfully with a single-layer grid, but the simulation time was extremely long. When I tried running it again with a two-layer or three-layer nested grid, it still failed. I would like to understand the reason behind this. Thank you very much.

Timing for Writing /fs2/home/qiaozhi/software/wrfout/2017/wrfout_d02_2017-01-01_00:00:00 for domain 2: 13.23135 elapsed seconds
WRF NUMBER OF TILES FROM OMP_GET_MAX_THREADS = 56
Tile Strategy is not specified. Assuming 1D-Y
Total number of tiles is too big for 1D-Y tiling. Going 2D. New tiling is 3x 19
WRF TILE 1 IS 1 IE 12 JS 1 JE 1
WRF TILE 2 IS 13 IE 24 JS 1 JE 1
WRF TILE 3 IS 25 IE 36 JS 1 JE 1
...
WRF TILE 49 IS 1 IE 12 JS 17 JE 17
WRF TILE 50 IS 13 IE 24 JS 17 JE 17
WRF TILE 51 IS 25 IE 36 JS 17 JE 17
WRF TILE 52 IS 1 IE 12 JS 18 JE 18
WRF TILE 53 IS 13 IE 24 JS 18 JE 18
WRF TILE 54 IS 25 IE 36 JS 18 JE 18
WRF TILE 55 IS 1 IE 12 JS 19 JE 19
WRF TILE 56 IS 13 IE 24 JS 19 JE 19
WRF TILE 57 IS 25 IE 36 JS 19 JE 19
WRF NUMBER OF TILES = 57

 

[Ming Chen]

Your namelist.input looks fine, except that the time step is way too small . Please change it to 120.

If the case crashed, there should have some error messages in your rsl files. These messages are not necessarily saved in rsl.error.0000. Please check all your rsl files to find possible messages that are helpful for us to figure out what is wrong.

 

[sher]

Thank for your reply. I have checked the rsl.error files. I would like to understand the reason behind this and how to solve the problem. Thank you very much.The information is as follows:

WRF NUMBER OF TILES = 66
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
wrf.exe 00000000032F9E82 for__signal_handl Unknown Unknown
libpthread-2.28.s 0000145D8BD9FCE0 Unknown Unknown Unknown
wrf.exe 00000000025BDCB5 Unknown Unknown Unknown
wrf.exe 00000000025A703F Unknown Unknown Unknown
wrf.exe 000000000259F09C Unknown Unknown Unknown
wrf.exe 000000000259A2EF Unknown Unknown Unknown
wrf.exe 0000000001D72CEC Unknown Unknown Unknown
libiomp5.so 0000145D8C094ED3 __kmp_invoke_micr Unknown Unknown
libiomp5.so 0000145D8C057726 Unknown Unknown Unknown
libiomp5.so 0000145D8C05671C Unknown Unknown Unknown

Stack trace terminated abnormally.

 

[sher]

Your namelist.input looks fine, except that the time step is way too small . Please change it to 120.

If the case crashed, there should have some error messages in your rsl files. These messages are not necessarily saved in rsl.error.0000. Please check all your rsl files to find possible messages that are helpful for us to figure out what is wrong

And there's about 200G of storage left when using 30 cores. I can run the normal WRF4.5 version successfully.

 

[Ming Chen]

Your rsl.error.0029 indicates that this case failed immediately after wrf.exe started with segmentation fault. This often indicates that either the input data is wrong, or the run-time memory is not sufficient for your case.

However, If the same case is able to run using officially released WRFV4.5, then I am suspicious that the case crashed not because of the above two common reasons. There might be something specific in WRF-Solar that leads to the model crash.

Please recompile the codes in debug mode, i.e., ./clean -a and ./configur e-D, then rerun this case. The log file will tell exactly when and where the model crash first. This can give us some hints what is wrong and how to fix.

In addition, can you run this case with use_wudapt_lcz = 0? Please rerun WPS without using LCZ. I ask this because I am not sure whether WRF-SOLAR works fine with LCZ.

 

[sher]

谢谢。在修改 Registry 并重新编译之前，我在使用 namelist.input （包括 aero 设置）时能够成功运行 WRF 模型。但是，在使用 Registry 更改重新编译模型后，我遇到了分段错误，并且仍然没有提供详细的错误消息。


# WRF-太阳能
状态 实际 swddir ij misc 1 - rhd “SWDDIR” “短波表面向下直接辐照度” “W m-2” “”
状态 实际 swddir2 ij misc 1 - rd “SWDDIR2” “来自 FARMS 的短波表面向下直接辐照度” “W m-2” “”
状态 实际 swddirc ij misc 1 - rd “SWDDIRC” “晴空短波表面向下直接辐照度” “W m-2” “”
状态 真实 SWDDNI ij 杂项 1 - rhd “SWDDNI” “短波表面向下直接法向辐照度” “W m-2” “”
状态 实际 SWDDNI2 ij 杂项 1 - RD “SWDDNI2” “来自 FARMS 的短波表面向下直接法向辐照度” “W m-2” “”
状态 实际 SWDDNIC ij 杂项 1 - rd “SWDDNIC” “晴空短波表面向下直接法向辐照度” “W m-2” “”
状态 实际 SWDDNIC2 ij 杂项 1 - RHD “SWDDNIC2” “来自 FARMS 的晴空短波表面向下直接法向辐照度” “W/m^2” “”
状态真实 swddif ij misc 1 - rhd “SWDDIF” “短波表面向下漫射辐照度” “W m-2” “”
状态 real swddif2 ij misc 1 - rd “SWDDIF2” “来自 FARMS 的短波表面向下漫射辐照度” “W m-2” “”

我按照你的建议，在调试模式下（使用 ./configure -D）编译了模型，具体信息显示在 rsl.files 中。
感谢您在诊断和解决此问题时提供帮助。

forrtl：错误 （72）：浮动溢出
Image PC 常规线路源
wrf.exe 0000000000E29A4B5 未知 未知 未知
libpthread-2.28.s 000014CD610BFCE0 未知 未知 未知
wrf.exe 0000000000B3445C4 module_shcu_grims 1149 module_shcu_grims.f90
wrf.exe 0000000000B3431A6 module_shcu_grims 1030 module_shcu_grims.f90
wrf.exe 0000000000B342B3F module_shcu_grims 972 module_shcu_grims.f90
wrf.exe 0000000000910EFA3 module_physics_in 3958 module_physics_init.f90
wrf.exe 000000000090A00DD module_physics_in 1526 module_physics_init.f90
wrf.exe 00000000005C5C865 start_domain_em_ 1067 start_em.f90
wrf.exe 00000000004F482F0 start_domain_ 122 start_domain.f90
wrf.exe 000000000043E6895 med_initialdata_i 229 mediation_wrfmain.f90
wrf.exe 0000000000414244 module_wrf_top_mp 273 module_wrf_top.f90
wrf.exe 00000000000413AD5 MAIN__ 23 wrf.f90
wrf.exe 00000000000413A62 未知 未知 未知
libc-2.28.so 000014CD60B1ECF3 __libc_start_main 未知 未知
wrf.exe 000000000041396E 未知 未知 未知
~