madMAx released GH 3.0 today. C29-C33 compression

I haven’t tested it myself yet, but their website says:

They haven’t setup a community node yet… If you use your own node it will work.

I’m farming there myself: SpaceFarmers.io

1 Like

Okay, thanks for the clarification, then I’ll start replotting without fear.

You just need to make sure the pool allows you to set the difficulty manually, not all do.

I’ve never changed the difficulty and I don’t understand why to do it. Now I have GH C19 plots, everything seems to be working fine:


Will it be necessary to change the difficulty for new plots C31?

Yes, just go to settings:

And click Save Settings.

See the Gigahorse chart for the recommended partial difficulty, lower diff will have double the compute in worst case (ie. diff 1).

I’m letting Felix do all the work!!! :rofl: :rofl: :rofl:

What I don’t get is why it appears the size of the farm doesn’t play any part in the difficulty setting. Previously it was based on farm size, so my 583 TiBe of c19 it was 120.

Now for any size of c30 farm its got to be 20,000, my farm would be roughly 756 TiBe, does this mean the gap between partials will be bigger, the effective capacity reported by the farm will fluctuate a lot more???

What if we have a mixture of old and new plots etc.

Nothing seems to be explained in relation to the difficulty, apart from it will halve the workload of the GPU, but surely that relates to how many plots you have.

Yes unfortunately. You can run lower diff but need to handle higher compute load.

It just happens that I found a way to take advantage of higher difficulty. To make the new plot format more efficient.

2 Likes

Just a bit of stats, in case somebody is interested:

RTX 3060 12GB farming max size [PB]

                   		difficulty			
	filter    20	    200	    2000	20000
C30	256	      0.174	    0.18	0.272	0.338
	128	      0.087	    0.09	0.136	0.169
C31	256	      0.095	    0.096	0.124	0.177
	128	      0.047	    0.048	0.062	0.089
C32	256	      0.045	    0.045	0.051	0.079
	128	      0.022	    0.023	0.025	0.039

GTX 1070 8GB farming max size [PB]

               difficulty
	filter	20000	50000
C30	256	    0.148	
	128	    0.074	
C31	256	            0.084
	128		        0.042
4 Likes

The current chia (solo) difficulty is around 12k. That would imply that farming solo will not be as efficient as joining a pool that have adjustable difficulty, as I think it is not possible to specify difficulty for a solo farm. Is there are GH setting that would force the farm to use 20k in such a case?

Also, if we set the difficulty to 20k (for pool setup), if there is a mix of c1X and c3X plots, will those c1X plots be visible / counted by the pool as they will show up really rarely?

I don’t think that FoxyPool algorithm is aware of c3X difficulty requirements, so it will have to be set manually.

Then what would you recomend the Diff setting be when one has C17, C18, C19 and now C30 to C33 plots? (using FoxyPool)

In terms of partial difficulty it’s around 800k. Network diff and partial diff are different.

It will work just fine.

Use what is best for the new plots. It doesn’t hurt the old plots.

2 Likes

Who came up with this link for diff calc?? Can this be modified for C3X?
.
chia static difficulty calculator :: xchdata.io
.

Who the heck is this “Felix” person?

.
.


He is the man…

Auto diff will be fine until you hit your GPU limit, then you can set a high(er) static diff manually to squeeze out a bit more farmable space. You don’t need high diff from the get go, unless your GPU is already at its limit. Foxy supports setting a high static diff.

2 Likes

Hi, i’m Felix, creator of Foxy-Pool

2 Likes

Thank you for replying. I actually wanted to ask you about it. (I am also pooling with FoxyPool.)

I think that I am not yet at the GPU limit (about 50% avg load). Although, I did modify my diff yesterday by about 50% yesterday, and a couple of hours ago bumped it up by ~3x (over the auto-average). What I have noticed is that my lookups improved a tad and looks like my GPU power draw is lower. All looks good so far, even though the Shares chart has more gaps right now.

Based on that, if there are those gaps in the shares due to high diff (effective HD space is at the same level), will that affect the daily payments, or will those payments follow the effective HD space (weighted shares)?

What I am trying to say is that if the GPU power draw drops with higher diffs, why not to use that as a default to save a bit on electricity expenses. Of course, at some point increasing diff will have diminishing returns as far as power consumption, so maybe there is a sweat spot a bit higher than the calculator you linked to suggests.

1 Like