Problem with gmx energy

GROMACS version: 2020.1, 2021
GROMACS modification: No

Hello,

I am having problems with gmx energy in 2020.1 and 2021 versions. The simulation runs without problems., but when I run the command I get this error:

(...)
WARNING: there may be something wrong with energy file md_50ns_CD-PRG.edr
Found: step=0, nre=0, nblock=0, time=0.

WARNING: there may be something wrong with energy file md_50ns_CD-PRG.edr
Found: step=0, nre=0, nblock=0, time=0.

WARNING: there may be something wrong with energy file md_50ns_CD-PRG.edr
Found: step=0, nre=0, nblock=0, time=0.
Reading energy frame     90 time    0.000
WARNING: there may be something wrong with energy file md_50ns_CD-PRG.edr
Found: step=0, nre=0, nblock=0, time=0.

-------------------------------------------------------
Program:     gmx energy, version 2021
Source file: src/gromacs/fileio/enxio.cpp (line 339)

Fatal error:
edr file with negative step number or unreasonable time (and without version
number).
If you want to use the correct frames before the corrupted frame and avoid
this fatal error set the env.var. GMX_ENX_NO_FATAL

The output is incomplete or sometimes even empty. Any thoughts?

Thanks in advance.
Best

Hi,
it can be that your energy file (edr) file is corrupted. Good to check the other outputs to understand what happens.
You can try to avoid the corrupted part, by using option -b -e or gmx eneconv gmx eneconv — GROMACS 2021 documentation.
Best regards
Alessandra

Hi,

Thank you for your reply. I checked with the version 2019 and got the same error, also with a different system. So I am suspecting this is a problem with my HD (SSD) of my new device, and/or an I/O issue. I checked my HD for errors and it is OK though.

When I load the .log file of the run, I get lines with NULL value, and a skip in the time step. During the execution there is no error though, and the run keeps going. On the other hand, I need the whole run, so ignoring the corrupt parts is not an option. Also, I did the run with the code compiled with double precision and get the same corrupted files.

Any thought or suggestion will be appreciated.

Best,
-Yasser

It is extremely unlikely that this is due to an issue in GROMACS. I suspect something is wrong on your system. Either the data on disk gets corrupted, or buffers don’t get written correctly to disk.
How long are your runs and at what point in time do you get corruptions?

Hi,

thank you for your reply. My run is 50 ns long and I get the first corruption before 5 ns, sometimes even before 1 ns.

.

I think it is something with my device, because I did some runs with NAMD, and I got corruptions too.

I meant walkclock time, not simulation time.

Ma’am,

can you please help me out to understand the output of gmx eneconv?

actually I did 500ns simulation in 50 batches each of 10ns, now I want to concatenate.edr file using gmx eneconv module. I checked the output file using vi editor and no. of lines in the file were less than the input file as well as the size of the output file was similar to the input files…should’t it be otherwise? how exactly I can check what I did was right or wrong. also I’m confused because of what is written just before the end of the concatenation step which are same despite a huge change in no. of input files. please have a look at the attached file.
123
same was the case when just one 2 .edr were used as well as when 25 .edr files were used

Thank you in advance!
-Sneha

250