CPU usage by itself can be pretty unhelpful information. I remember from my mainframe days, an MVS system with nothing to do would run itself ragged looking for deferable tasks to perform while it could.

If you're trying to compare machines, a more relevant benchmark might be something like: how long, in wall-clock seconds, does it take BIAB to regenerate (or whatever, some intensive operation, maybe a chord wizard) a substantial piece? Which might be what you really care about.

Last edited by Mark Hayes; 01/04/22 07:56 AM.