Quote from: jstefanop on January 30, 2023, 08:23:21 PM

Quote from: crypto_curious on January 29, 2023, 11:44:15 PM

Quote from: Polynom42 on January 29, 2023, 10:38:00 AM

Quote from: crypto_curious on January 28, 2023, 11:02:09 PM

Quote from: PakoTor on January 27, 2023, 09:52:28 PM

Quote

Just use ckpool, that's what I'm doing

The point of mining solo directly on our node is it might be a little bit faster.
Any fraction of a seconds counts to get the edge.
ckpool is not the best with one block per 4 months and all the intermediate connections.
FutureBit told me solo mining was not ready for release yet.
Hopefully soon.

Solo mining to your own node may have the opposite effect. ckpool's VPS is in very well-connected paid datacentre, and most likely connected to maximum amount of 125 peers. ckpool can push the block to 125 connected nodes, and to the entire world, in a matter of few seconds. What is your home upload? Most likely in tens of Mbits or slower, no match for 1Gbit upload from datacentres. If your block is not propagated well, then 20ms delay you save by not connecting to ckpool means nothing if you propagate your block in 30000ms (30s) and by then someone faster pushes hit block and yours become an orphan.

Currently my node is connected with 18 other nodes and upload for a few kb nonces cmon u dont need 1Gbit. I just bought this full node package for mining the own node. Sure ck.pool is not bad and the only option right now not supporting the big boys like ant/via etc.

18 connections? That's very low. I connect ONLY via Tor, no IPv4/v6 connections as all, and have 32 peers. I was reading a bit today about block propagation in Bitcoin network and apparently it's still an important factor, especially for big miners who hit blocks often. You must be extremely lucky to hit the block within your lifetime if you own one, or even 10, Apollos. Then, you must be extremely unlucky to have that block orphaned by not propagating fast enough, but it is possible.

Yes this is one of the reasons we have not released it yet, we are building infrastructure on the backend that will allow each individual node to connect to the back bone of the main pools. Number of connections is not as important for new block propagation as how close your node is to the nodes of main pools.

Its in the interest of small miners to send the block to the main pools first, because as soon as the large pools node sees the Block they will stop mining on the previous block header and move work immediately to the next. If a small miner is on the fringes of the node network, by the time the block propagates to the pool nodes they could have very well found another block and will orphan yours.

This would be a horrible outcome, your device actually finds a 6BTC valid block but you lose out because of slow propagation. We are currently spinning up super nodes that are on fast connections and directly connected with major pools, in a future update all futurebit nodes will have our infrastructure nodes as a seed nodes, so once we enable solo mining direct to node if you do find a block the first "hop" will be to our backbone super nodes, so at worst your block will be two hops away from all the major pools.

This should significantly increase odds of small solo miners not orphaning blocks, but want to make it perfectly clear its still worse than solo mining directly to ckpool for example (and the degree of how much worse depends on your upload speed).

If your on dialup or DSL you should not be solo mining on your own node, and would not recommend anything less than a 30-40 mpbs upload speed and direct ethernet only (worst case 500ms delay with ~1.5MB blocks). Not bad but obviously not as good as a 30-50ms upload speed of a few kb pool share to ckpool. So for example to be on par with solo mining to ckpool you would need beat ~50ms share propagation to ckpool + ckpools 50ms block propagation (assuming his setup is near perfect which I would assume it to be), so 100ms for ~1.5MB you would need a 120-150 mbps upload speed.

I also dont want to discourage people and blow the issue out of proportion either, bitcoin has 10 minute blocks for a reason its extremely unlikely anyone will find two blocks within a few seconds of each other...for each second you delay uploading a block there is 1/600 chance you'll orphan the block.

This is also a good exercise in remembering why we won the block size war...if we did not no one on low end hardware or slow internet connections would ever have a good chance at solo mining...

This sounds great! Any idea on an ETA for this new software? Any plans to open it up to testing?