search

Chia-Network / bladebit

A high-performance k32-only, Chia (XCH) plotter supporting in-RAM and disk-based plotting
Apache License 2.0
340 stars 107 forks source link

Suggestion - don't increase plots made count after finding a failure #424

Open bhorvitz opened 9 months ago

bhorvitz commented 9 months ago

Not something urgent, but it would be convenient if the plots made count did not increase if a plot is removed due to failing check:

Generating plot 9 / 2196: 315cabfb27705e124e8f2ea3be61074a5a7db3c2f65992a71592272dfc863ac6
Plot is OK. It passed the proof threshold of 80.000%
Generating plot 10 / 2196: 4288d67c1217d2fe34dc8e5a6e7d905061b4c448ac8e053f7353abbce68ee8ba
WARNING: Deleting plot '/plots/out/plot-k32-c05-2023-09-30-20-36-4288d67c1217d2fe34dc8e5a6e7d905061b4c448ac8e053f7353abb
ce68ee8ba.plot.tmp' as it is below the proof threshold: 0.760 / 0.800.
Generating plot 11 / 2196: fb57dfec35249608d826dc63d0b82f86f225f97b5a208372283be8f3cdc414ba
Plot is OK. It passed the proof threshold of 80.000%

In this example, we should have made 10 / 2196 again.

sobertram commented 8 months ago

I think the enhancement would be to add a removed metric and not change the created one. Even though a plot was removed it was also created. So it is accurately showing how many were created. Or maybe instead of removed show a net created. Not sure how that would look. Generating 11/2196 net 10, maybe too much.