I'm trying to run the grid* parse reranking scripts in ~/logon/lingo/redwoods/ to reranking English parses in the jhpstg corpus. Many of the reranker configurations are failing, and I want to understand if their particular mode of failure is familiar to others who have done this work. I am running many different grid* files in parallel using the Condor distributed computing system. On my latest round of jobs, all of my Condor jobs completed, but when I look at the directories created under ~/logon/lingo/redwoods/tsdb/home, I find that most of them have empty score files. I take the empty score files to be a sign that something didn't work. I suspect the problem is that the reranking jobs require more RAM than my Condor nodes have installed. However, were the machine to run out of RAM, I would expect the ./load --binary grid* process to die, and this fact to register in the Condor logs. The fact that this is not happening makes me wonder if there's something else going on that I haven't thought of. Have other people run the grid reranker and ended up with empty score files? If so what was the problem? -- Bill McNeill <a href="http://staff.washington.edu/billmcn/index.shtml">http://staff.washington.edu/billmcn/index.shtml</a> Sent from: Seattle Washington United States.