What is this Windows Plot Mover you’re using?
Its a program I wrote myself, but its available on Github
2 Likes
This is more of a general point of view, not directed at any one individual considering NoSSD.
The big drawback of NoSSD is they originally obfuscated libtorrent in the code so they can remote update/execute code on your machine, they at one point promised to remove it after enough found out about it, but here we are months later its still present despite those promises.
“Its okay we added an option to disable it” but how could we prove it, they still obfuscate the code.
You now have an extra level of concern on your local network not only from what one of them might do some day, but what some foreign entity may do when they want a botnet and have a source for one with far less security experience and knowledge to keep those thousands of machines safe.
So what are the drawbacks for the couple percent gain, here are some of the big ones…
- Anonymous developers who have had a history of public outbursts casting doubt on them
- Your plots are locked into their pool, meaning zero portability if something goes wrong
- They sign and create blocks, meaning the decentralization of the network drops down quick
- Should another coin use Chia plots, they can create blocks for that network without telling you
- Libtorrent inside NoSSD giving them anonymous remote update/execute to install anything
- You now must trust them with access to your machine and the entire local network as a result
- The remote update/execute becomes a valuable target to someone who wants a botnet
I would say a couple percent isn’t worth the above risks, they have to be right 100% of the time, if they screw up once, you’re stuck replotting or worse, someone using your machine for something else.
(I’m sure the last thing anyone wants is an officer at the door with a warrant to seize equipment used in something you didn’t do, how much will you have to spend in attorneys fees for an incident like that, if thats what a couple percent is worth to someone, then hey, I guess its their roll of the dice.)
7 Likes
Hi, so Foxy Farmer can farm both BB c7 and GH plots? That’s my dilemma right now, I have BB c7 plots, and would like to switch to GH. But I will not be able to farm my BB c7 plots as I replot…
PS - can I keep running the chia full node and run foxy farmer at the same time to farm the new GH plots? If I use foxy farmer, do I need to join foxy pool? I assume so, but wanted to confirm. Thanks!
Or can I run GH Farmer and the Chia Full Node at the same time (same base version as GH Farmer)?
I just closed to the Chia GUI, then run Foxy-Farmer, and then Foxy-GH-Farmer, they both run at the same time, and this creates two new harvesters on Foxy Pool. As long as you have been using the GUI on the same machine/account it will pick up everything from that.
I’ve no idea if you can run the GUI alongside say Foxy-GH-Farmer, although I think it may work. Currently I just look on Foxy’s pool to see how things are going.
Still think NoSSD is the better option unless you want to solo.
How much extra electricity does the new compression use vs the old? For example c8 vs c18 ? Has this been measured yet?
You’re obviously not taking into account @ChiaPizza post above, that is exactly why I haven’t installed it. They are staying anonymous for a reason, and the torrent software is also there for a reason, and was initially hidden.
Give me a week and I’ll tell you the difference between BB C7 and GH c19.
1 Like
Not sure on efficiency of farming the new C18 plots,
But with the new 2.0 GH client, I know the old C8 will be more efficient to farm with.
So overall, whether you replot to the newer smaller plots, or stick with the older GH C1 through C9 plots, you will see overall better efficiency when updating to the GH 2.0 client.
2 Likes
While the new plots will be more “efficient” prospectively, I’m wondering if we should also factor in the cost of replotting. I noticed my electricity use spiked up by a lot when I replotted regular chia plots to BB c7…so now, I am trying to take that into account in decided to replot to GH plots. I already have the hardware so no extra cost there unless I’m going for K33 plots which I might do to ‘“future proof” a bit. And that would mean getting 512gb ram if I want to do all in RAM.
How do others think about this? Of course, so of this is for fun too given chia price right now
1 Like
I am still undecided on which way to go for a re-plot from OG k32’s.
It seems that as quickly as I research and plan a re-plot regime, things change again.
Looks like there are plenty of people re-plotting not for the first time, but perhaps second time at least!
1 Like
I agree that’s why I am on the side watching, 
and letting the dust blow by a little.
1 Like
500W plotter. 24 hours is 12kwh. Say 0.2 USD power cost. It is $2.4.
250 plots a day (plotter has 1080 GUP). That is one cent a plot.
1 Like
We should do, but I’m just replotting regardless, but lets do the exercise.
My plotter uses about 600w, so that’s 14.4kWh a day, times six days = 86.4 kWh.
If I was paying the UK 30p per kWh it would cost £26, but my costs are lower, a combination of off peak electric, solar, and some exported at peak - I have a large battery bank.
So lets take off peak, which is just under 17p, that brings the cost down to around £14.57, but it will be lower than that due to solar (not that there is much at the moment) and peak exports.
I’m plotting roughly 5800 plots, so divide £14.57 by 5800 = £0.002 per plot.
I will also have roughly 1576 more plots, so an increase of roughly 38% over the 4224 I had before.
As a bonus the upstairs of my house is a nice and toasty 22c, which means no need to burn gas to heat it.
It also seems that K32 is going to be with us for a long time, and plotting K33 will increase the plotting time a lot, and also the compute needed to generate the full plot, so more power usage, so I’ve stuck with K32’s.
NoSSD CAN be remote farmed or plotted - using NFS mounts on the farmer/plotter machine, NFS server on the drive storage machine.
This is both simpler and more complicated in ways from setting up recompute_server, as Windows doesn’t always support NFS (need Win10 Pro or an add-in 3’d party add-in) and there’s a bit more work to set up the shares themselves while NFS is normally built into LINUX.
It DOES make the plotting noticeably slower (I’ve not seen any effect on the farming to date) - better to plot to a local drive then use (needs mild modification for the file name difference) plow to move the plots, as that doesn’t slow the plotting down as much.
Max compression level on NoSSD is a little smaller, but any other level GH can match (give or take tenths of percent, they’re NOT exact matches).
There is the “only one pool” limit, of course - but the overall fees work out to be about the same.
You can BARELY do both farming AND max compression with 12GB of video ram, but you have to lower the “farming” ram limit some from the default.
If you run the GigaHorse farmer on Gigahorse plots, you can use any pool with it except NoSSD.
NoSSD is similar that way to Flexfarmer, no “full node” needed but specific to it’s owning pool.
it runs like shit
source
4PB
Gigahorse can now also farm Bladebit compressed plots. Max being Max, he’s also fixed a bug that caused 3% proofs to be dropped.
Thanks for the heads up! Now, I have no reason to not replot to GH!
In this chart:
What is the difference between the yellow C1-C9 and C11-C20? Is it simply CUDA is needed for C11 to C20?
Is Max farm size what I can farm with a single 3080 Ti? If so, why would someone plot C9 when C16 appear to have the same compression, but with higher Max farm size? Assuming one has a capable enough CUDA card.