search

madMAx43v3r / chia-gigahorse

224 stars 31 forks source link

Recommended Difficulty? #334

Open spleen911 opened 5 months ago

spleen911 commented 5 months ago

I have 4 harvesters and 2 are double the physical size of the others. All have 3060Ti GPUs and I am running a static DIFF of 10,000.

The C18/C29 systems AVG 6 & 7sec and MAX 13 & 19sec over 12H respectively.

SPLEEN has 54 single SATA3 over USB3 across 7 controllers DEEPDISH has 120 SAS2 in RAID0 (11 arrays of 6-12-24HDDs) across 3 controllers

The C30/C31 systems AVG 5s and MAX 12s over 12H.

Wondering if I should increase my DIFF to offset the C30 & C31 plots

I'm hoping to not have to get better GPUs and chill for a while @256 filter.

Harvester C# GPU PLOTS Physical Effective Gain % More Space
SPLEEN C18 3060Ti (8GB) 2815 576.26 1149.95 573.69 99.55
  C29 -- 8795
DEEPDISH C18 3060Ti (8GB) 2100 525.47 1059.84 534.37 101.69
  C29 -- 8600
PLOTTER C30 3060Ti (8GB) 5300 224.18 524.82 300.64 134.11
LASAGNA C31 3060Ti (8GB) 2700 101.83 267.36 165.53 162.56
      30310 1.394 2.932 1574.23 110.26
madMAx43v3r commented 5 months ago

It looks like everything is good, no worries.

spleen911 commented 5 months ago

Thank you. As a test, I increased DIFF to 12,500 and did not notice much of a change (as expected). I am sending 2-3 PPH with an occasional hour of no partials. No impact to points or SP times. Since I still have some C18 to replot to C29 and another ~10TB of C31 to plot, I'll leave it as-is unless the C1 harvester starts logging warnings.

madMAx43v3r commented 5 months ago

Yep, just move to 20k if there's too many stales.