I generated new EPICS database files with Daniel's updated parser. This splits the databases into separate files for input records and input/output records. It also splits the records into multiple files to limit the number of records in each. This was necessary because too many channels would keep the EPICS OPC gateway from running. We can also now scan the input and input/output records at different rates. I quit the EPICS OPC gateways that were running, and started new ones for these databases. There are now 17 instances of the EPICS OPC gateway running on h1ecatc1. This seems to allow more channels to connect, but I'm not sure how much it will help performance. As more instances are running they take longer to start up. We may have to try spreading the processes over multiple cores, but I don't believe there is currently a way to have this done automatically.