Don’t you find it ironic that the chia GUI spends so many resources?
0,4% CPU vs 0% for flexfarmer
1.210 MB of RAM vs 11!!!
Sorry but I’m going with Flexfarmer now.
Can you run Chia’s client and the Flexfarmer client at the same time?
Did some reading up so now my understanding is:
Lower CPU on account Flexfarmer does not maintain the blockchain database
Lower Memory on account Flexfarmer does not maintain the mempool
If I’m correct (?) I kind of feel using Flexfarmer is somewhat parasiting on farmers running the chia full node. Flexfarmers ‘enjoy’ a decentralized blockchain/DEFI platform without really contributing to it.
Just an opinion and maybe I didn’t quite understand the way Flexfarmer works, also posible of course;-)
3 Likes
Flexfarmers are delegating the power of their HDD to Flexpool, they’re contributing disk space but somewhat giving up governance.
The Chia GUI client does much more than the flexfarmer client, hence why it uses so much ressources.
But if each and every farmer would switch to the flexfarmer client, the (4 or so) flexfarmer server nodes would become single point of failure and more important a single point of control maintaining the blockchain. Of course they would still payout the farmers for their HDD capacity but could at the same time insert any transaction they saw fit into the chain. Not saying they would, could!
So there is value in running a full node as a farmer. Which takes some memory and cpu cycles.
It’s not that I don’t see the issues with chia’s current client and the problems for farmers to sync and stay synced. I fully understand people going there. But a couple of MBs RAM and some CPU load to save as single reason doesn’t justify going that route in my opinion.
2 Likes
It’s true that the Chia GUI client isn’t optimized. It’s also true that pools re-centralize mining. But in the big picture, (1) a full node runs fine on a NUC (and even a rpi4 with some tweaks) or very cheap computer, (2) no chia pool has the kind of power as some btc pools have (and flexpool has a negligible share of the network).
Keep in mind that Flexpool has made clear that they want to open source their client, they just can’t do it in an economical way yet, and that their work will eventually pay off for the whole community more than it already does right now.
2 Likes
It really depends what node you run (how much resources you have). If your node is not powerful enough to run full chia node (e.g., low RAM and SD based RPis), then running FF is your best option, and potentially the only option. Same, if you have an old box with low power single core performance and have problems moving your blockchain db to SSD/NVMe.
On the other hand, if your node is OK, doesn’t suffer problems with daily syncing, then by switching to FF unfortunately you will give up all logs, or rather log analyzers (e.g., farmr, ChiaDog, PSChiaPlotter). If you have more than a couple of drives, that may not be something that you want. This was the primary reason for me to dump FF. I also had a tad more stale points on my FF harvester, although that could be just bad luck (Flexpool has problems with stale points, when mem pool is even a bit elevated, so maybe FF somehow is more tied to that, but I don’t really see how, so had to be rather bad luck for FF on my setup).
There are also claims that FF reduces HD load, but one of the assessments coming from Flex was that it is so minuscule, that rather the difference could be due to either the current ambient temps or air pressure (IMO) while running their tests - basically not worth to be considered. (To me, that is just optimization for the sake of optimization with already diminished results.)
And of course, for those that struggle with chia for various reasons, FF may be a much better solution, as that is a much better written code (and of course, like it was pointed out, it is a subset of functionality of what the full node does). On the other hand, if you compare FF to just a harvester running via CLI, the difference may not be that great.
As far as being afraid that FF will take over the world, and open up possibilities for manipulation, that is just a farfetched claim not worth to entertain, as FF only works with Flex pool for now, and that is just about 1% of netspace, plus some Flex members (like me) dumped FF for the lack of proper (chia like) logging, and basically no improvements by using it (giving up too much and gaining rather nothing).
Also, FF is just a temporary step, as Flex is working on a full node solution. As Flex’s focus is primarily on making that code working, I have no doubt that it will be much better engineered, so being real competition to what chia produces. If that full node will have compatible logging, so things like farmr will work, I would see really no reason to stick with chia code anymore.
So, FF is a good thing, and really saving a lot of setups, and should be given plenty of credit for that. Although, it is helping farmers, not really netspace (consolidates nodes for those farmers). The original bashing from Chia was either based on ignorance or NIH coming from their side. Both Gene and Bram have already stated that they were misled and therefore wrongly bashed on Flex (kind of don’t buy it, as for once, they are the best expert in this area and I don’t think need advices from new hires, and also it was just one another bashing on another piece of software that basically outperformed chia code hands down - e.g., MM went through the same route, farmr is proving that chia UI is really bad, and provides not much info to operate your node).
1 Like
No, I don’t. You are comparing a truck to a sports car. Chia is a full node, where FF is a harvester with some limited farmer parts. Due to that, you cannot run FF without first fully syncing Chia and creating a plot NFT for flex. You are also limited to Flexpool only. Also, you are giving up on any third-party components that help to fine tune your farm.
Although, if your box is underpowered (doesn’t look like yours is), or you have a hard time getting chia to run on your box, FF is a perfect tool for you.
I just ran flexfarmer - for the sake of experiment - on a Pi Zero 2W : it’s doing ok-ish with a (2.5inch!) 4tb drive. I could never do that with the official Chia GUI/client running a full node, maybe a cli harvester only… You have to give them credit where credit is due : it’s a flex!
3 Likes
When the duststorms started, I didnt wanna buy a SSD for my pi-farmer and switched to flexfarmer.
Now my little buddy can work for months without my help at all.
2 Likes
Yes 
If you don’t know what you are doing, you will be submitting your partials twice, as such one will be dumped. Potentially, the pool may take some actions considering it inappropriate (as it will look like your node is trying to double dip on points - not sure how this side works). Not sure what will happen if your chia harvester points to any pool but Flex, where FF to Flex. It may be a good enough reason to ban from the pool(s).
If you know, then you can disable harvester on the chia side, and have it run just a full node, without a farmer/harvester (so both of your dbs will be syncing / giving you wallet access, while FF will be harvesting).
2 Likes