Changing the hashing algorithm
-
I read the graph that Calem posted wrong, the lower score is better with Blake2 coming out best. Blake is actually the evolved version of Salsa that we are currently using. I’m happy to take Ghostlander’s advice on what is most suitable as he is the most informed person on this and he came up with this ingenious idea. Right now I do not believe that this is urgent, GPU miners will only start looking for alternatives when the Scrypt hash starts climbing rapidly and their kits stop generating a profit like we saw with Bitcoin.
If we can come up with a working specification then we can take this to the altcoin forum on Bitcointalk to gain interest of other altcoin devs.
-
OK , Can we consider adding POS system ? POW + 25%POS . This is another way can be selected.
-
the NeoScrypt
-
I was thinking more about names last night. Another suggestion that’s much less obvious. “Idavoll”, from what I remember it’s a place in norse mythology untouched by the battle of Ragnarok (end of the world). Those that meet up there take the seeds of Idavoll and use them to reseed the scorched lands.
If you think of ASIC landing as an ‘end of the world’ event, it works.
-
I was thinking more about names last night. Another suggestion that’s much less obvious. “Idavoll”, from what I remember it’s a place in norse mythology untouched by the battle of Ragnarok (end of the world). Those that meet up there take the seeds of Idavoll and use them to reseed the scorched lands.
If you think of ASIC landing as an ‘end of the world’ event, it works.
Nice!
-
It should be possible to create installable binaries together with a basic configuration… :D
-
OK , Can we consider adding POS system ? POW + 25%POS . This is another way can be selected.
If you mean Peercoin PoS, it breaks existing FTC coin distribution model and offers unlimited coin stake generation over time. Peercoin controls inflation by destroying transaction fees.
-
Or implementing unique PoW that really solves some problems that require processing power ( Primecoin like )
-
Improved algorithm relies on the CPU . XPM is a reference.
-
If you mean Peercoin PoS, it breaks existing FTC coin distribution model and offers unlimited coin stake generation over time. Peercoin controls inflation by destroying transaction fees.
I have seen that sometimes the destroyed transaction fees are greater than the stake generated.
Something that occurs to me, it is presumed that we are attacked by resources that are intended for Litecoin. Warren from Litecoin said as much when he took the time to attack us on the Litecoin forum when he thought we were going to be removed from BTC-e. Moving to a new algo stops a huge amount of hash easily being thrown our way to toy with our difficulty. We need to get out of Litecoin’s shadow as the pioneer’s for altcoins are not altcoin friendly. Oh the irony.
On a side note Warren appears to be an excellent dev who has made good progress with Litecoin, he may not like us but we can still like him :)
-
I kinda changed my mind about changing the algo. I think if we do a switch to a new algo which isnt present at the moment, we could come out big of this.
But Ive got some concerns about that:
At the time we switch to a new algo, all hashing power will be rejected, everyone needs to update to a new miner to continue mining
-
How would we calculate the new difficulty?
-
How long would it take, cause in this switching time, nobody could sent/receive any Feathercoin at that certain point.
-
If we stay at the same diff ~200 and not all are switching to our new algo or not fast enough, it would roughly take a good bunch of minutes to solve the first block to proceed transactions again.
-
Next Question is, a new algo means a new/different hashingpower - or would it stay the same? If it wouldnt, its needed to recalculate the whole difficulty to transport it to the same level as we left it on the algo switch.
BUT if that can be managed that well, that the transfer is one smooth transition, we gonna rule the #world (jk but yeah that would be awesome)
Just my opinion.
-
-
very valid points, ChekaZ
I think, we are just in the initial phase, where an idea is born. Then there will be a developement, a test and an implementation phase.
Some of your questions can’t answered right now, but will be later.
Regarding the new difficulty calculation,with the new algorithm, that will be implemented with 0.8.6.1 the difficulty adapts much faster to hash rate changes and therefore it should be no problem, even if we would start with a very low hashrate, let’s say 50.
If we assume, that the real hashrate would be ~3Ghash after the change, the difficulty would adapt in 3-4 blocks and we would start with 2blocks mined very fast and then come back to the 2.5 minutes interval again.
The bigger proble I see, is that we may end up with 2 different branches of the blockchain, one with the old version and another one with the new one.
How could a wallet decide,which one to use, when sending ftc?
-
@CheckaZ
When the new algo hits everyone would need to stop their miners and start them back up again. We would probably want some miners running the new algo before hand which would generate rejected shares until the change. The difficulty is probably the hardest thing to get right. Some of the large mining pools may not follow us to start with and there will be no multipools switching on to us anymore so our difficulty will need to drop drastically. We could drop our difficulty 10 fold and it might still not be enough! It is probably safer to drop our difficultly too far than have it too high, there may be some fast blocks it will work out the new difficulty quick enough that way, leave the difficulty too high and we may be waiting an hour a block!
Ghostlander’s suggestion on how to change algo would give us a solution that behaves in the same way as the current one. I look forward to seeing a lower diff without attackers toying with it and multipools stripping us,
I would not to consider changing the hashing algorithm without something like ACP to cover the vulnerabilities during that period. Hopefully in the long term this change means that we can disable ACP by default but during the change we need this protection more than ever. An attacker could really run riot while everyone tries switching algo.
-
I’m not against ACP and I agree, that we should keep it during the first phase of the new hashing algorithm.
Regarding difficulty, I don’t see a realy big problem here. We just should start with a lower than expected difficulty, so the fist blocks with the new algorithm are found faster than the 2.5 minutes.
I did a quick simulation starting at diff 10 with 3 Ghash and after 12 minutes or 15 blocks we’d reach a block spacing of 1.9 Minutes, after 18 blocks or ~ 19 minutes we would reach the 2.5 minutes target spacing.
The trick is to start on the low side, so the first blocks are found faster than 2.5 minutes.
But I’m definitively talking about future problems. Before we should worry about difficulty, we should do a lot of other things, if we really change the hashing algorithm 8)
-
Switch point can be set to UTC rather than block number, so the people know when to switch their miners.
How could a wallet decide,which one to use, when sending ftc?
Not a problem. Old nodes will be disconnected after the switch by protocol version.
-
How soon is this possibly going in to effect ? Have stayed mining FTC only since I first started mining the beginning of this year, and am looking forward to the change. The manipulation lately has had me thinking of changing to a different cryto, but I really believe in this community and refuse to give up.
-
Glad to hear, that you like the community :)
I think, it’s a bit confusing, so let me try to explain.
Currently feathercoin suffers from large pools jumping in at low difficulties, mining a lot of blocks before the blockchain can react and increase the difficulty. As soon as the difficulty reaches a given level, they move on to the next coin to mine, leaving the other miners at high difficulty with low hashrate with the result, that less blocks are found and the miners have less income.
In addition it takes much time to bring the difficulty down, as the time between blocks it high then and the difficulty is aligned every 126 blocks only.
As soon as the difficulty is down again, the pool come back and so on and so on.
This problem is now and will be addressed in version 0.8.6.1 of the wallet which probably will come in the next weeks.
The other problem -and that is what this thread is about- is, that sooner or later ASICs will be developed to mine scrypt based coins like FTC. Asics provide high hashing power at low power consumption, but are expensive.
The typical result is that only a smaller number of people/groups can affort the invest in ASICs. These people generate high hash rates causing the difficulty to go up, and GPU mining is not provitable anymore.
What is discussed in this thread is the pro and con to change the hashing algorithm used for feathercoin in order to delay the upcome of ASICS as long as possible, as it is the intention of feahtercoin to rely on a distributed infrastructure, consisting of thousands of miners, rather than a realively small group of ASIC owners.
This second problem is a long term problem, as currently there is no proof of a comercially usable ASIC miner for scrypt based coins like feathercoin.
-
I think we have a lot of things need to be prepared.
1, Need to develop new cgminer .
2, Need to develop new client software.
3, All mine pool update stratum protocol.
4, A new coin switching scheme.
-
Would also be nice to get a cuda miner for those who have deviated to the Nvidia dark side. >:D
-
Not confused, just asking a simple question. I was just asking how long to implement a more robust algo like DRK without the bad guy invisible persona… Understand pool hoppers “manipulation” , while leaving us all to work down the difficulty. I am someone who decided to get on the “FTC” boat to help row and not willing to jump over board just yet. Thank you for taking the time to answer my question in great detail.
