Malformed trajectory file output

[Robert Welch]

Here is an excerpt from the trajectory file in question

#!

...
Blob 0, Conformation 0, step 13110001
DYNAMIC
1.021036e-08 1.559304e-08 -2.748892e-08 -2.280400e+03 2.222771e+03 1.266563e+03 0.000000e+00 -2.280400e+03 2.222771e+03 1.266563e+03
7.665759e--1.958530e-08 -2.128941e-09 -1.847364e-08 -4.333484e+02 1.221210e+03 -6.379302e+02 0.000000e+00 -4.333484e+02 1.221210e+03 -6.379302e+02
-1.983657e-08 -9.398398e-10 -2.016620e-08 -9.169688e+02 7.167968e+02 -1.589561e+03 0.000000e+00 -9.169688e+02 7.167968e+02 -1.589561e+03
....

The line beginning with 7.66 contains a very strange number, 7.665759e--1. The line also contains 11 columns instead of 10..

This frame is actually missing around 200 nodes. The previous frame contains 1646 nodes (the correct number) whereas this one only contains 1461 nodes.

In all other regards, the file is normal. Whilst it will load in by script.load.trajectory() (giving a 'expected %f %f %f warning when it reaches the end of the frame early), any attempts to perform any other operation will result in a list index out of range error.

I will not attach the full trajectory files as it's around ~300mb. But I can scp it to you.

Edit: I also have the output of FFEA in a text file and did not see it crash or give any errors at this step.

[Robert Welch]

Update: the same thing has happened on the other simulation I was running on livia. I'm not 100% sure but I'm 95% or so sure that it happened at the exact same time.

Edit: they happened within 30 minutes of one another. But not at the exact same time. Somehow, that makes things more confusing!