Author Topic: June '19 - Proposal #3 - Change PoW algo again?  (Read 3591 times)

SomethingGettingWrong

  • Newbie
  • *
  • Posts: 9
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #15 on: June 25, 2019, 01:31:44 PM »
Randomx would be fine but Graft Currently doesn't have checkpointing. If it was implemented that would be fine. But Randomx will assuredly have Nicehash and with Nicehash 51 percent attacks come. With checkpointing nicehash would be okay and in certain instances good for the network.

CN-GPU is a good algo I think its more FPGA resilliant then any other. However the lack of documentation will probably negate the team from doing it and that's a logical thing to do. That's why either conceal … or checkpointing and Randomx is the best case.   But I prefer Conceal and Checkpointing because at the end of the day.... GPU Hashrate will outshine Randomx Hashrate on a small network and the difficulty on CONCEAL with GPUS will be higher in equivalency then RANDOMX with Gpus and CPUs and that dictates a larger difficulty and thus bigger minimum cost to mint a coin. Thus making the price more.

Something people don't understand is that although the algo doesn't technicly dictate a price.. it does on an unexploited algo dictate the minimum cost to mint a coin and since you don't sell for less then you spend in electricity it very well does give incentive for a higher price!
« Last Edit: June 25, 2019, 01:34:00 PM by SomethingGettingWrong »

jagerman

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #16 on: June 25, 2019, 06:52:31 PM »
Randomx would be fine but Graft Currently doesn't have checkpointing. If it was implemented that would be fine. But Randomx will assuredly have Nicehash and with Nicehash 51 percent attacks come. With checkpointing nicehash would be okay and in certain instances good for the network.
RandomX is actually designed to be easily tweakable in incompatible ways.  If Graft were to adopt RandomX, it would twist some of the easier configuration knobs which would make it totally incompatible with Monero's RandomX hashes.

vlkaliskatezi

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #17 on: June 28, 2019, 11:09:00 AM »
Conceal or cn-gpu would be great RandomX maybe in future.

yidakee

  • Administrator
  • Jr. Member
  • *****
  • Posts: 56
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #18 on: June 29, 2019, 05:24:54 PM »
Randomx would be fine but Graft Currently doesn't have checkpointing. If it was implemented that would be fine. But Randomx will assuredly have Nicehash and with Nicehash 51 percent attacks come. With checkpointing nicehash would be okay and in certain instances good for the network.
RandomX is actually designed to be easily tweakable in incompatible ways.  If Graft were to adopt RandomX, it would twist some of the easier configuration knobs which would make it totally incompatible with Monero's RandomX hashes.

 But, as Patrick pointed out, we've still to implement checkpointing. So another perspective is that, if we move to RandomXG, the network hashrate will plummet in favour of CPU.

 So, initially … how decentralised would the network really become? And who would be mining Graft's chain? Also, what sort of infrastructure would you need to compromise the integrity of the blockchain with a 51%?

 Since 51% in this scenario it is an extremely low-cost risk vector under RandomX, until we have checkpointing, should we consider another algo?

« Last Edit: June 29, 2019, 05:37:18 PM by yidakee »

SomethingGettingWrong

  • Newbie
  • *
  • Posts: 9
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #19 on: June 29, 2019, 06:31:47 PM »
I think that RandomX is a good algo avaialbe to be changed easily at the core algo level. 

I however feel it is way too new as there is no GPU Software and it is untested in the mining community on this aspect.
 However The algo itself has been tested and audited. Please don't get the two confused! It is only a matter of time before the mining
community decides if they will support RandomX are not.

The GPU software is out of grasp currently. Leaving only CPUS.

TeamRedMiner gave me a flat out nope!
https://github.com/todxx/teamredminer/issues/68

XMRSTAK has hinted they wont be implementing anything unless Monero is swapping to it. They feel that RANDOMWOW is a play on algo selection.
https://github.com/fireice-uk/xmr-stak/issues/2441

with that being Said... being CPU only is not an Issue at ALL!
For Loki they have a separate Configuration of Randomx called RandomXL as well as Checkpointing enabled.
This is the perfect Algo for them if they didn't want a FPGA GPU version.

IF Graft were to fork XMRIG as well as do a separate Configuration of Randomx I don't see a problem if they can implement Checkpointing in time.
However I currently feel the timeline isn't going to allow for it. Unless it is given a priority.  There should defintley not be a Normal RandomX version..
 It should definitely be custom version of it!  This will be to get them off of nicehash. RandomX will be on nicehash if Monero accepts it.

Now this creates another problem. RandomX will eventualy be on GPUs. The codeing is avaialable but incomplete
https://github.com/SChernykh/RandomX_OpenCL
https://github.com/SChernykh/RandomX_CUDA

Once the Gpu's code is finished even if not profitable. It wouldn't be hard to create a private version of these variants of RandomX from the sourcecode. Even if  a GPU version of default RandomX becomes common place then someone will fork it for graft in private! I hope Graft takes the time To fork a GPU version if the other Opensource community doesn't do it once its common place. This would lead the opportunity to create a 51 percent without checkpointing.

If you go RandomX please implement Checkpointing ASAP.

I have no doubt that as more and more coins come along there will be GPUS that move over to Conceal and Cn-GPU as well as the other most profitable FPGA reistant coins. There will be a niche that needs feeling. IF monero for example Chooses RandomX where would all the GPUs Go?   one thing to consider when you don't include GPUS. Cn-GPU included Nvidia and AMD. But I digress. I believe the team has made their choice. Im not saying its a bad choice. Its not. I just think the timing might be wrong since the lack of Mining Software GPU support and  No checkpointing.
« Last Edit: June 29, 2019, 06:36:45 PM by SomethingGettingWrong »

Raybl

  • Newbie
  • *
  • Posts: 6
  • Karma: +1/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #20 on: June 30, 2019, 06:45:13 AM »
A change to a randomx variant is not a bad move, but there are some unknown factors coming along with it and as several people here and on telegram pointed out the timing might be suboptimal. Nevertheless a imminent change to a FPGA resistant algo is needed.
Randomx is a cpu optimzed algo and it lacks a amd gpu implementation at the moment to mine it. I wouldnt expect gpus to be profitable on this algo. There is an overall change coming in the mining scene. With Monero forking to randomx in october all the gpu friendly coins will be flooded with massive hashrate and their profitablity will plummet. So a change to randomx or better variant is a good move from a miners profitability view but comes along with investment costs for many miners to even be able to play a role.  Monero once had a cpu mining algo and moved to gpu, why? What were the problems with cpu mining? And will there be new whale miners in form of e.g. datacenters? I dont understand Moneros intentions to move back to cpus. I can imagine that randomx will be further developed to a cpu and gpu friendly algo, that would be optimal.
All these headaches are reliefed with the implementaion of checkpointing! For the moment a change to a gpu friendly but fpga resistant algo like conceal or cn gpu and another change to randomx variant shortly before monero is forking could be taken into consideration.

Sundoshi_Nakamoto

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #21 on: July 03, 2019, 02:05:37 PM »

 No brainer for to me. PoW algo change is an urgent matter. So is RTA which is around the corner. Without checkpointing RandomX is too dangerous.

 CN-GPU or Conceal as remedy for now - launch RTA/new PoW on mainnet ASAP, then go full on checkpointing. From then on, the world is our oyster.

spaolo

  • Newbie
  • *
  • Posts: 1
  • Karma: +0/-0
Re: June '19 - Proposal #3 - Change PoW algo again?
« Reply #22 on: July 10, 2019, 03:41:58 AM »
It is true that FPGA have an unfair advantage in terms of hash/watt but is also true that FPGA entry point is 20 times higher than GPU [1 RX580 cost $200, 1 FPGA $4000, last year $300 vs $6000], warranties are usually valid for 6 months [compared to 3 years for GPU] therefore risk of ownership [of failures] is about 6 times greater and from a ROI standpoint [not mining GRAFT but mining a profitable coin for GPU] we are looking at the same period [of course after the card is paid is mostly unfair but again, warranty is limited and haven't heard yet of FPGA working for years without problems]. Having said that, a single miner owning 10% of the hashrate isn't healthy [I have switched my miners now and own 10% of another coin mining hashrate and in only a couple of months, converted into FIAT, it would be 10 times what mining Graft produced for me in the whole of last year]

If it is true that a FPGA make $14 a day it still take 280 days [9+ months] to break even [without electricity] which is more than the warranty itself. If you ask me, that miner is not dumping, is accumulating. But I may be wrong if he is somehow trying to control the price, to accumulate [by buying at bottom] even more. Also 380K GRAFT a week will give another 50K extra GRAFT a month if supernodes are ran.

I have been mining GRAFT since the beginning of 2018, employing 100's of GPUs. I have been questioning the developer's work publicly and for that I've been banned from the GRAFT Twitter account. Bummer. Now the *** is hitting the fan and they are asking the help of the "community" which in fact does not exist [I was the only person commenting on Twitter for GRAFT major announcements. Now I can't see it but maybe still the same?]

How many of you knows about their digital credit card crowfund? I imagine all of you since there is only a bunch on here. I bought one cause I thought it was an amazing idea. One month later I got the money refunded [about $60] cause they couldn't even get $60K thousand in fundings.. that low were/is GRAFT reach. I still own all the GRAFT but it is becoming obvious it wasn't my greatest decision and its sad.

The coin / supernodes concept and progress is/was amazing up to a certain point, when they started worrying about its valuation rather than improving it. Like a race when you are given very low chances of success and think that by changing external factors, it would increase odds. And here we are again. Graft is marketed as an open-platform, open-source, community project but it act as a company and the development is controlled. Seeking "community" feedback this way, this late, after they spent years NOT creating a community but thinking they knew exactly what they were doing and they didn't need anyone cause they knew it all, perfectly, well.. is not just a sign of weakness at this point it seem like a panic attack after the attack is over, a car that lost its wheels. For goodness sake, get your sh*t together and start coding again. There you go, that's my suggestion. Don't have enough money to keep going? Sell the idea, get investors onboard. Owning 30% of something is better than owning 100% of nothing. Have enough fundings? What the heck are you doing then!!! Putting some timid announcement at the bottom of your site as in we are open, seeking investors, accepting donations etc does look cheap. Once again it does make it appear like you are closed into your tiny office and do not know how to tap into investments; when, in reality, all they do these days is to seek to invest into Crytpo.

Yes, I am upset, I don't care about the money and time I wasted, I am irritated at a project that could have really changed the way world-wide transactions are made but instead is stuck worrying / blaming a single individual. Don't blame a freaking miner for the coin valuation, if you were churning out good updates, nobody would care. Actually there would be dozens like him if it was that good. This guy is just wasting his time, like I did last year and he will be going in a bit when he realises there isn't enough buying to support his selling [if that's what he is doing]
Want to PREVENT this from happening? You need to have contingency plans in place: A, B, C & D, what if, what if what if & ready to go, not suffer the consequences and still thinking what to do months later! Bitcoin, ETH etc have to do that, not a company!

Do you have a monetary policy? A graph, a curve that explain the release of the coin from now until the end of its distribution? Show it so that everyone understand what you are trying to achieve & how. Mining is a distribution mechanism. How long do you like to keep doing that? How many miner would you envision should be mining/distributing GRAFT [every person in the world? 10,000? 100,000? 1M? 10M? Currently I believe +/- 100 on the entire planet]. How much GRAFTS in total, ever? Is that simple. Don't care about how many ppl should mine / receive coins to distribute them? Then I don't see the problem if one person gets them all. Oh wait but one person is not right. Ok, then, what is the right range then? How do we achieve that? Or are we arguing the case that until I get it, is fine but if someone steal my cookies then I throw a tantrum?
Before there was the problem of mining farms with 1,000's of rigs. Now this. He is not getting it cheap, he is getting it fast. Why would someone want to spend $4,000, to get them back slowly in 9 months? With all the risks involved, not to mention the time he/she have to spend to keep it going.

Also one factor ppl are not considering is that in certain countries, electricity is soo cheap is practically free. Some other miners have free electricity, as part of their contracts, or from solar, some others just steal it.. and is quite common, so the argument FPGA is getting weaker and weaker. I am not saying it doesn't exist, I am saying that the policy need to be clearer. I.e. we will allow FPGA when they will be more affordable by everyone [entry point being X] until then we will keep changing algo. Mining rigs aren't cheap either. so the entry point consist of what? What we can afford? Hash/watt? So, only hardware with hash/watt below certain specs is allow? That doesn't make much sense from an energy efficiency standpoint but hey, anyone can decide whatever they want. Anyone can buy one of those rigs these days, $30K is all you need. Not long ago ASIC miners were sold for $10K each, and now the same are sold for $500.

Want to make it fairer? Limit the amount of GRAFT a wallet can hold. Link the wallet to a fingerprint/facial recognition [almost every phone has a camera these days] hash output [indirectly linked to the person but in a way that it cannot be reverted, identifying the person but also in a way to make it hard for a person to get more wallets, or at least keys to wallets]. Put a serial number on each GRAFT coin. Limit the number of wallets anyone can hold. Is mining still settling the ledger on the blockchain? Why not moving the ledger operations to proof of stake and/or supernodes completely and leave just the coin distribution to the miners. Call me crazy. Be bold, be different. Whatever, just do something. Come on people!

Where is your public policy for ASIC & FPGA miners? FPGA miners are not new, they have been around for more than a year now, it cannot be a surprise, they must have been put into the plan with ASIC already months ago. Are they welcome? Yes/no. If no, how are we going to contrast them. If they still make it through, how are we going to respond?

Someone was talking about trust. Definitely but you don't instil trust by saying what you are doing today for tomorrow, that's just hope, hope that tomorrow shit doesn't hit the fan again for some other reason. Trust is injected by leading from the front, not by trying to avoid obstacles at the back.
« Last Edit: July 12, 2019, 03:19:30 PM by spaolo »