Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

WRF Benchmark - different compilers - output validation


New member

I have built WRF using GCC compiler and ran conus 2.5 benchmark. I got results such as wrfout nc files and rsl out file. And the same way, if I built and generate WRF results using another compiler, how do I compare/ validate the results between GCC and other compiler for accuracy/ correctness? Are there any specific variable(s) should I look between two results?

When running an identical test case using a build with two different compilers, we expect there to be some differences, and over time, those differences can grow. For a short case, like the CONUS 2.5 benchmark, the results shouldn't be that far off. I would just ensure they are somewhat similar and there isn't anything wildly different.
Thank you kwerner for the response.

To verify results identical conus 2.5 benchmark case, ncdiff of wrfout files output is enough ?

Last edited:
There is a program available in the WRF/external/io_netcdf directory, called "diffwrf." I would recommend using that. For e.g.,

./diffwrf wrfout_file1 wrfout_fil2

(using the actual names and paths of the files). You can also use the ncview program to visually compare variables.