PB is out again for [heavy?] maintenance, else you'd be able to see the noon projection chart indicating a new record bashing number of results of around 92-93K, up from last weeks new hi of 85.7K. This is not going to be with a runtime record, as the mean execution time is slightly down to about 11.37 hours, in fact almost an hour down from yesterdays 12.24.

Edit: Turns out, the old style of predicting does not work for FAH2. The workday crunchers start their wave on Monday which then delivers a bigger serve on Tuesday, but then the next run won't complete until Wednesday... so there was 47K validation in the morning and 36K in the afternoon.

What I mean is, this image displays the updated version, whereas yours in the first page doesn't:

5 months and we're still only in the 106 batch test-calibration-optimization phase. The contribution curve keeps slipping, but that's a consequence of there being less generations/step blocks to go around [as batches are completed of the initial set].

Crunch On.

https://bit.ly/WCGFAHB

Keep watching and report changes... With enough data moments there's an endpoint to extrapolate.

Thx for recording. It means that WCG now *does have* an estimated work volume indication to plot an ETOC out in time.