Probably 200 MBytes…
Can anyone tells me how many C32 plots does a RTX 3060 ?
Decided to buy 2 4070 Super for plotting. Congratulate me)
4 Likes
Hell i do and absolutely no difference. Plot linux farm windows.I reformat any drive from a windows system to ntfs in Linux (usually to get rid of a buncha junk on said disk). Never had an issue. Matter of fact it writes just as fast if not faster to the drive. Only issue I had with Linux was that I was plotting quicker than my u.2 nvme. Could get rid of them I.e. write it to the disk @ 250 mbs, And the nvme slowly getting filled. Solved that issue with 2 nvme and boom. With my pci-e 3 z840 I plot 72 TB in a day. I’m sure there are way faster but for me this is amazing.
And thank you Max for publishing your software. Some bitch about the fee, but it is well worth it to me. It does exactly what it’s suppose to, work. I dont have to worry about it not working. My harvester stays on for weeks at a time with no crazy behavior,it just works.
1 Like
For those that would like to maybe better understand what drives Max @PoorInvestor-Youtube has a new interview with him - https://www.youtube.com/watch?v=QfUxK4HqtmM.
I was surprised that Max has attributed the loss of GH farmers to NoSSD 2.0 better compression. Of course, better compression was the key. However, IMO the simplicity of NoSSD client cannot be underestimated. It is obvious that not a single NoSSD farmer gives a rat’s ass about the Nakamoto coefficient and would not have any problems if GH had a similar client to jump NoSSD ship for a higher GH compression and potentially a better security.
I was also surprised that MMX was not mentioned at all (wasn’t it a big part of the previous interview). I would really love to hear an update about it, whether Max is still working on it, and the rough ETA (if the project is still active).
Yes true, fast farmer will fill that gap.
I will get back to it after GH 3.0 is done.
1 Like
So, there is a chance that it will hit the mainnet relatively soon? That would be great. 
I would think the majority had never heard of the Nakamoto coefficient, let alone knew what it meant, until recently I hadn’t, but I steered clear of NoSSD for other reasons.
I plot on Linux to a temp drive, I’m plotting around 56 TiB a day using a 3080 (around 110 seconds per plot) and Telsa P4 (around 400). I use my Plot Mover software on Windows to delete old plot and move the new one across the 10Gb network, but I simply can’t plot quick enough, although I have 10 disks setup it’s often only moving two plots at a time, as it ends up waiting for more to complete.
I keep resisting the urge to buy a 3060 TI to replace the P4 for plotting, it doesn’t make financial sense, but the geek in me wants to see how much faster I can replot.
1 Like
With a 3080ti in pci-e 3 the best I got was 1.86 minutes. My inner geek does the same thing. Why i now have two nvme for temp. Saw a bottleneck and HAD to remedy it lol. I’m limited by the bus speed but oh well.
1 Like
I’m also on PCIe 3 - using a Dell T7910, my 3080 usually does a c30 plot in around 110 to 130 seconds, but it seems to be on a go slow today at around 170 seconds, oddly the P4 hasn’t slowed down.
2 Likes
What is “Fast Farmer”? You speak of.
1 Like
Great now I have to watch that Poor Investor guy to see what Max has to say. I find his videos a little like fingernails on a chalkboard… 
2 Likes
1 Like
Great. Thank you for the link. So that is a lightweight farmer. Super.
Wait, wtf lightweight farmer even means? I went through that description before, also checked the github. Kind of not enough info about it. Just one adjective that should be explaining everything.
Hopefully, someone will provide more info about it soon.
1 Like
It means you don’t need to run any CNI code or even a node to farm. You can optionally use your own local node however, instead of a community node.
And it allows to make use of CHIP-22 for GH, you can use an official node + GH fast farmer.
2 Likes
So, as the node use is optional, does it mean that it can run completely without a node (local or one coming from the pool)? How will it connect to bc then?
Also, if Chip 22 is implemented, will it be sufficient to only provide a Chip 22 compliant harvester, and have the rest of the services coming from chia? Or rather both harvester and farmer need to be provided.
1 Like
No you need a node, either local or community node.
When using fast farmer you only need a node, as it already acts as a farmer + harvester (like flexfarmer).
This node has to be an official node yes, a GH node wont work.
1 Like
So which software for windows should we be looking for (will stotiks be working on the windows version)?
1 Like
Then is the node-to-farmer protocol the standard one (looks like it is, as it connects straight to chia node), or some additional proprietary protocol is used to connect to Foxy’s community node. By the way, what that community should imply, basically a shared node run by a non-local farm? Using that extra name (community), in this case implies proprietary protocol to me.
It would be really nice if both that Foxy page and FastFarmer pages were updated to provide a bit more info. Maybe some images similar to what you did for recompute (that would be really cool).
1 Like