I suggest you look at page 9. A total power output of ~300 MEGAJOULES (~83KWH) from a piece of palladium 100 mm long and 2 mm in diameter.
As to "...they didn't get excess power until after 64 days"...", this work is from 1996. Since that time, other researchers have fixed the problem of a long "lead-in" time. Several researchers have devised means of generating excess power at or shortly after startup.
I looked at page 9 and what I saw is that out of 7 experiments, 4 were total failures, a fifth was a partial failure, and of the two that appeared to produce excess power, they weren't consistent. So the results were still pathetic, and P&F were unable to exonerate themselves in spite of several millions of dollars of funding.
And the watt is the proper unit for reporting energy rates. Joules are used when you get pathetic results over a long period of time, but still want to impress the gullible.
In addition to pallidium they were replacing solvent and applying input power. And let's not forget the considerable input power when they weren't seeing any excess power, including the power applied during the 5 failed experiments.