Sorry it’s taken a bit of time to get back to you with this. It’s not at all clear what is causing this behaviour at the moment.
On my machine I also see the same bombing out with the ‘top 20’ partition, which leads me to ask how many CPU cores you have on your machine?
Also, I tried running the bottom (full) table in a chunk loop, one row at a time, and I see some pretty wierd stuff here too - random rows have empty output, and each time I re-run it, it is different rows. I know there is another thread on this dataset at Good workflow for optimizing structure geometry?, but I am going to ask @greglandrum who has commented over there whether there is some sort of timeout built into the conformer generation? (Adding a random seed seems to make that consistent rather than random, so I’m guessing it is simple conformer generation failure somewhere along the way)
Steve (@Vernalis’ alter-ego!)