In the materials provided in the JCP-2019 paper（Locally adaptive pseudo-time stepping for high-order Flux Reconstruction）, the mesh file pyfrm has been divided into 4 blocks for multi GPU run, but I only have one GPU. How Can I change the mesh back ?
You can partition the mesh as per the documentation.
Can I just use “pyfr partition 1 mesh.pyfrm .” command ?
I mean that in the materials provided in the article, the mesh file is in “pyfrm” format, not in msh format, and has been partitioned, but I only have one GPU. What should I do？
Yep, that should work. As with all the commands in pyfr you can do pyfr partition --help and it will tell you how to use it. In the partitioners case, it accepts a pyfr mesh file (ie a pyfrm file) and can optionally take pyfr solution files (pyfrs files) as inputs.