Plot file size can be 2.5G smaller without any farming cpu usage increase

During write the GPU plotter, found there are few place can be optimized for a smaller plot file.
So far , 2.5G space can be saved if use a new plot spec. There are no extra cpu usage when farming. ( actually even lower cpu usage.). but as it’s a updated plot spec, it won’t be compatible with official plotter file.
Please vote that if this enhancement helps community.

  • yes, that’s great , please publish it.
  • no, don’t care 2.5G at all.

0 voters


i think many one working on compress plot as I see flexpool says something in discord


I am so hesitating to implements those compression b/c it’s a pandora’s box, if it’s open, it maybe a disaster for whole community. Smaller plot file will force the whole community to replot for the advantage. As NoSSD already start this battle, nobody can retreat if you don’t wanna lose. So chia the company have this in road map. If this compression war is only one fight then settled, that’s still ok. but what if it never end? eg. every 6 month , few percentage less plot file format created? so the whole community keep replotting? will that be a nightmare for everyone?

1 Like

Not trying to be Debbie downer on plot compression, but the whole idea of saving a few GB/plot seems to make since only for those newly plotting hard drives. Here’s what I mean: say I already have 5000 plots or about 1/2PB of k32 plots now and I see that the ‘new’ plot format can save me 2.5GB/plot.

OK, so 2.5GB x 5000 plots is 12,500 GB or 12.5TB. Wow, so much savings on 1/2PB! (sarcasm). For the cost of plotting new a 12TB drive vs the cost of replotting 500TB** … frankly I’d rather have a root canal done, something along those lines, than replotting 5000 K32s. Further how many XCH will a brand new 12TB drive going to get over the long haul? Not many.

Bottom line…Great Idea…make something that does plot compression. Anyone contemplating plotting new drives can jump at it…the rest of us…not so much.


do we want to make here now every time a vote whether we publish what? come on that really a little childish.

but back to topic. i’m always surprised how seemingly capable people create things, but then on the other hand have a problem with the value or the impact to classify.

a few months ago we saw this with noSSD who valued their development with 100BC and thought now everyone would run him over and switch to their pool.
6 months later they have just 0.6% of the total network size and there the size reduction is over 20%.

here we are talking about 2%, where should there be a negative impact?

it is a pity, this is now the 2nd time, things were developed which are close source, and so the chia team then develops it itself as open souce. in the end, the possible advantages of the clouse souce projects are lost, because they can not flow into the open source.

if i understand this correctly, the GPU plotter and a 20% smaller plot format will be released next by the chia team.

1 Like

NoSSD was not impact too much simple b/c the market situation.
if it’s first 3 month , and 20% less plot file size, you guess what happens? still remember hpool ? lol
Do you ever heard “Free software foundation” ?
Do you know GUN founder Stallman once said that
“Open source is a development methodology; free software is a social movement.”
and even more details for you.

If free software is not domination the community, opensource just make things even worse. Think about after WTO, why US & China still have trade war?

Same situation here. If chia the company really able to get 20% less plot file and opensource it, ( I highly doubt that’s possible without farming cpu usage increase) . what will NoSSD do? they integrated it. then 40% ( for example ) less plot file commertial plotter come out? what you do then? re-plotting twice or still lost 20% revenue to NoSSder ? lol

Only communication will helps you understand more and only vote knows what people really need. we call this blockchain consensus. lol

1 Like

i didn’t want to attack you or have an unnecessary discussion about definitions of open / free software.

i just wanted to show you that in my opinion your 2% savings will have no impact on the chia project no matter if you publish it or not.

as far as i know the chia team is reprogramming the same approach of noSSD (which by the way chia itself provided as proof of concept even before noSSD existed)
so there will never be 40% smaller size by combination of both things because it is the same thing.

have a nice day

1 Like

Just curious that nossd’s compression algo is their core biz secret. How chia the company know which way nossd did?
Does chia the company reverse engineer nosed’s binary?
Btw,if it’s same way, that’s sad. 20% is not free, we have to pay extra hydro bill for it and if every one did 20%, then it’s whole community pay extra bill without any gain.
How stupid this move is. !!!
This is negative sum game for all. All loser no winner.

1 Like

If Chia didn’t pay (the ridiculous) BTC demand by noSSD for the sauce, I’m assuming they did exactly what you said.

The small gain the GPU plotting project has found is in line with what the Chia team has been saying for a while, that there were minor improvements to be had in the plots. I am not surprised someone working with the project managed to find some tweaks along the way. Pretty cool I say :sunglasses:. I’m personally sticking with the official project, but many people are more adventurous (Hpool back in the day, noSSD recently, etc). I guess I’m relaxing my views on alternatives a bit - I used to view this stuff as straight up cheating.

That “ridiculous demand” was rather a sarcastic response to Chia’s ridiculous offer and vicious / baseless attacks that were (I think) never retracted. (I have nothing to do with NoSSD, neither joined their pool.) The fact that Chia did not immediately contact them behind the scenes (PM would do) is just mind boggling.

The sad part is that it was not the first such attack by Chia. More or less, whenever they see NIH they go that route.

As far as the new plotter, I would not be surprised if that would be MM v2 and potential changes on harvester side also thanks to Max.

The fact that Chia published the demo code for that compression over 3 years ago (way before mainnet), and still had issues with NoSSD is also telling.

1 Like

sorry but that’s not the story, if that would have been a “sarcastic response” they wouldn’t have gotten the 100BC later with their workload and the people involved.

chia is definitely not the nice guy next door when it comes to business interests, no question. and the first reaction of Bram Cohen on twitter back then to noSSD was definitely not the right way to go, even if chia says that this was a statement as a private person.

they have a clearly defined busniss plan that they follow which is also largely public and of course they defend this against external negative influences.
developments that are not compatible with the current blockchain are definitely unintentional in my opinion.
if someone does not like this direction, no one is forced to join.

1 Like

Sure, whether that was sarcasm or not is just a guess. My read is that it was. NoSSD didn’t start with that idiotic number, but rather responded when all the crap was dropped on them.

That was my whole point. What NoSSD did had zero negative influences. The algorithm was published over 3 years ago by Chia, and NoSSD just did a good job in testing and implementing it. The fact that MM v2 (with that compression) supports that as well.

Unfortunately, technology moves forward one way or another. Once the cat is out of the bag the game is over. And in this case, the cat was let out 3+ years ago. So, really hardly not compatible.

Also, all that nonsense could be easily sorted out behind the scenes without much drama, and we would be already busy replotting for those compressed plots.

By the way, there are no "statements as a private person’ by the CEO, especially when it is clearly with respect to the company business and is basically a defamation of a party seen as a potential acquisition target.

Although, I would propose to close this exchange, as we are hijacking arc_plot thread for no reason. Sorry for starting this exchange.


please keep going. those opinion is what I exactly want to know. thanks for share.