Optimal compression. FindChia Pool

Dear Friends, Greetings!

After announcing support for compressed plots in the Chia blockchain network, the farmers of our pool started actively replotting their plots. Large farms and high compression settings led to a significant increase in GPU/CPU load, as a resulting in delays (stales)* and a decrease in the pool’s farming space estimation.

After analyzing the performance of farms with compressed plots, we discovered that by reducing the target partial** value in the account settings to 400, we can systematically reduce the farm’s load by increasing the contract difficulty. To ensure stable operation of farms with compressed plots, it is necessary to reduce the number of plot accesses by increasing the contract difficulty. We conducted several tests together with pool participants, which demonstrated that using the beta version of Chia blockchain with compression support and raising the contract difficulty is the optimal solution for stable farming.

*Stales are partials that are submitted to the pool later than 25 seconds.
** Target partial — the number of expected solutions from the farm within a day. By default, this value is set to 900 (if you are not using compression, 900 is the optimal value).

Follow Us

Linktree: findchia | Twitter, Facebook | Linktree

Twitter: https://twitter.com/FindChia

Reddit: Reddit - Dive into anything

Discord: Findchia.com

Telegram: Telegram: Contact @FindchiaChatENG

Facebook: FindChia Pool | Facebook


I would say that if someone has to lower his partials from 900 to 400 so that he doesn’t get any steals, then this is only a temporary solution.
because in 1 year when the plot filter changes he will have the same problem as 400 is the same as 800 today. (daily partials)

it would make more sense to think about a hardware adjustment if this is a long term project.

for me personally the times are 3-5 sec with C8 (gigahorse) which only increases from 8000 daily partials. (2.1 PB @ NV 4070 TI)
but I use only 1000 that is enough for a relatively constant farm size in the pool.

have a nice day

but i am not sure if the pool partials are not independent from the filter^^

how to change this target partial value/contract difficulty?

.lol. Large farms on a pool… that’s the problem.

You need to change the value of h 400 in the settings/contracts section. There is a pencil icon opposite the contact.

Apparently here we are talking about user farms and their problems, not the pool…

If the farm wasn’t on the pool, it wouldn’t need to do the extra compute to report back to the pool. Understand?

What kind of “additional calculations” are we talking about? The pool does not create any “additional” load on the farm.

Sure it does. The pool needs to track the size of your farm. The pool does this by issuing challanges to the farmers. This is in addition to farming challenges.

You are greatly mistaken. The pool does not send any calls to farmers. The pool determines the size of the farm based on the number of partial operations accepted and their complexity over the last eight-hour period.
I recommend that you familiarize yourself with the basics of farming in more detail. By the way, FindСhia has an excellent FAQ and WIKI

How about providing a link instead of saying, I seen something somewhere that proves my point but you can find it yourself. Here is a link for you Pooling FAQ · Chia-Network/chia-blockchain Wiki · GitHub
Answer me this oh wise one. Since farmers send partial proofs to the pool so the pool can determine their farm size, how is being a large farm on a pool not the problem? If The farm didn’t need to calculate and send partial results to the pool, that removes the problem, correct?

This is not about what I “saw somewhere” - this is about what is written in this article that you comment on.

  1. There is an initial erroneous opinion in your statement, so you should first figure out what a pool is, how it works, and why it is needed at all
  2. The larger the farm, and especially on compressed plots, the more power the farm itself needs to process its volume. If at this moment the farm is connected to the pool, but cannot fully process its volume, then the pool will receive exactly as much as the farm sent. The article only talks about how to optimize the work of the farm by reducing the load by setting - “reducing the target partial** value in the account settings to 400”
  3. It doesn’t matter if the farm is in a pool or solo, it will send the same information to the blockchain.

Thank you for the links, but my wisdom is enough for me))

So do you agree to my original point, that if the farm was NOT on a pool, it would NOT need to do the extra compute? Because nothing you said proves your point. A whole lotta words that aren’t relevant. Again, you never provided a link to back up your claim… Your claim being, a large farm on a pool is NOT the problem of high resource utilization.

If you want to look stupid, argue with a fool. Do you know such an expression ?
Your first comment sounded like this - “lol. Large farms on a pool… that’s the problem.” It was just taken out of context, because the full sentence sounds different))). Further, you stated that the pool sends some kind of calls to the farms, but after you could not operate on my response that the pool does not send any calls to the farmers and does not force the farms to experience additional loads, you switched to insults, showing your obviously unhealthy aggression.
Then you again, in the form of a question, issued a false statement and tried to answer it, trying to retreat. To which I politely replied stating only the facts set out in the article.
I don’t see where you found information in the article that large farms are a problem for the pool. Try to explain this to those who read your comments. For me, everything is obvious.
About some link that you require - do you need someone’s opinion? - so ask the opinion of the author of the article, and I draw my purely objective conclusions from my experience, since I am a miner of 3 different blockchains and a validator of several networks, and I know what I’m talking about.
Further, I consider the dialogue to be exhausted. If you do not agree, read the first sentence of this post.
Good luck.

I broke down the estimated increase pool partials add at certain difficulties here