Revise 2.0 memory footprint estimates in documentation

[yezr]

Actionable followup to #559. In it, Armin says

tabulating 300,000 CVRs out of 1,500,000 records takes 500mb additional memory (in addition to 350mb used just to launch RCTab). That means we'd be able to support 2.4 million CVRs on a 4GB machine - plus/minus a bit, since the rest of the machine will use some memory, but the JVM would also garbage collect more often.

Let's test and confirm this.

Then update Section 03 - System Hardware Specification in the TDP. This currently has some ballpark estimates about how much memory is required for certain ballot thresholds and how long it will take. Let's review and confirm those.

Once we have new estimates, review priority of #559