NoSSD Chia Pool update! GPU mining, 50.3GiB plots, >200% reward, new CPU & GPU plotters

so, nossd plot is not user specific. They can be used by other user. yes?

I am using a 3060 TI. You can do both, but when plotting, it doesn’t mine very well. Gets about 50% shares.

Correct. I could sell hard drives full of FPT files and you could just start farming them and point the rewards to your wallet address.

1 Like

So eventually there will be plot on the fly and all depending on the GPU power. CHIA becomes Bitcoin brother.

Appaently not. ( 20 chrs )

I wouldnt be angry about chia becoming bitcoin brother :slight_smile:

Better a “dependent couple”…
Of course, as long as bitcoin performs well.
No twists and betrayals in a relationship… :wink:
CC2

That post didn’t age well.
Took MadMax about a month to mostly match the new compression levels.

Up side - he has NOT matched the performance, quite, per the figures both of you have posted and per my own NoSSD experience - GH does seem to take a little more GPU to mine a given farm size and a comparable compression amount (going by plot size, not C number).

You can do remote computing on NoSSD, using NFS mounts.
I’ve got an old SuperMicro 6047 with 71 3.5" drives (one of the dual-drive caddies has the 2’nd drive mount broken) that’s being farmed by 2 remote machines with GTX 1070s - can’t mount anything current in that machine due to the combination of 8 bit PHYSICAL PCI-E slots and half-height limitation.
Mix of drives in it, mostly 14GB or 10GB with a few 12GB and 8GB, total space close to 800 GB.

Run a NFS server on the “to be farmed” machines, run NFS client(s) on the “remote farmer machine(s)”, and run the NoSSD software on the “remote farmer machine(s)” pointed to the NFS mounted drives (and can mix in local drives too).

I suspect you could do the same in Windows using SAMBA mounts, but haven’t had any reason to try it.

1 Like

Dual E5-2680 v1 with 128GB ram manages 7.5-8 minutes to plot to a 3-drive SAS 10k raid-0 array.
It was doing the same times when I had 384GB in the machine and was using 256GB as a ram drive to plot to, so CPU limited.
Older HP 1u server, can’t get a GPU in that machine - so it uses a modified plow.py script to send the end plots out via the 10GB lan in it (2.5 GB is fast enough to keep up, 1GB IS NOT).

1GB LAN IS fast enough to keep up with farming remotely via NFS though.

I removed some HDDs from my NoSSD farm and the software, even when restarted, is still looking for those folders. How do I get it to update its cache to stop looking there?

You can get a carpet knife and open the end of the 8 lane pcie slot or even use an extension cable. 800TB?

Do you have those directories listed in your command line?

No I didn’t. I had to delete the software and re-download it to make that issue go away.

Guys, Lets celebrate. NOSSD is at 5 EiB!!!
Do u believe this ? in 09.09.2023 was at 500 PiB.

Congrats to NoSSD team, Lets close the GAP to Space POOL!!!

2 Likes

It’s going to number 1. No stopping it.

Here’s something of a report. I’m GPU plotting C15 into some spare SSD space. On my desktop system 128GB, x570 mobo w/ 5800x and 3060 ti using --no-temp… Can’t break 9 minute plots. On my two node Epyc system, I get 1.4 minute C15 plots on each node.

No substitute for 8 channel RAM it seems.

Update: The plotting has settled to 8 minutes, 40 sec each C15 plot. Just letting it run in the background until it fills that 3.5 TiB ssd. Set it, forget it… “Mining” it via NFS as it builds.

Something is wrong, I have a similar specced machine with a MUCH WORSE gpu than you do and I get 3 min plots. (my gpu is 8 years old, yours is 2)

check if its cpu minng, check your cpu usage, etc.
something is wrong here.

I dunno… The Samsung PM983 SSD is hosted via a m.2 nvme to a SFF-8643 interface and cable to u.2. Right next to the GPU and that’s on the PCIE4 x16 slot. But that shouldn’t be a problem. The SSD is only written to in Phase 3 and/or 4 and it is shared via NFS, again shouldn’t be fighting the GPU too much.

Oh well, again it’s set and forgotten. Should be filled up by tomorrow morning.

O krap, nvtop says the gpu is on a gen 4 1x interface. That SSD seems to be gumming up the works. Don’t know if I’m up to messing with it at this point.

Every few days my farmer crashes with hundeds of these messages a minute.

14:46:57 ERROR: an illegal memory access was encountered (skipped 9 messages)

There are that many i can not see anything before the meesages start.

I am running 2 identical windows 11 PC’s.
AMD 3700X CPU
64G Ram
GTX1070 8Gb GPU

The one with 10277 C13 plots runs fine

The PC with 8293 C13 plots keeps spitting out errors.

I checked event viewer and it shows no issues.

I have uninstalled and reinstalled the latest video drivers.

I had no issues with previous version.

Command used is

client.exe -a xch1… -d C:\Plots\1\nplot -d C:\Plots\2\nplot … -w RobFarm1 --no-benchmark --no-plotting --no-rescan -v

I would appreciate some clues on where to look next.