HEM and PostgreSQL are two separate programs. Two seperate programs can use two different cores.
I don't think it really matters though. Except when doing a bulk initial import, I don't think I've ever seen HEM or PostgreSQL be CPU bound.
If you're having performance problems the bottleneck is probably in the I/O.
There are some ways to tune PostgreSQL but I've never really done it on windows. One thing you can do is make sure the data is stored on a different physical drive than the WAL. A different disk on a separate controller is even better.
I'm trying to remember but I think on Windows the only option is to have the WAL write to the drive where PostgreSQL is installed. The data directory however is easily configurable.
If you could get a really fast disk like a WD velociraptor, or better yet, an SSD to hold the data files for PostgreSQL I'm pretty sure you'd see greater performance gains than you would by getting a top of the line processor.
|