Scrypt Jane Research - Post Ideas Here
-
“Certainly we want an efficient hashing solution. We want something that will resist Scrypt ASICs and still be GPU friendly. I believe we can set the maximum of N. I will soon setup a testnet with four nodes and 5,000KH power to test out some different solutions. I’m currently involved in other high priority tasks so if anyone want to do some research on Scrypt-Jane and report what is possible in a new thread for the community that would be appreciated :)”
As Bush mentioned, we don’t want to be a CPU-centric coin. We want to be GPU accepting and ASIC/FPGA resistant. :)
-
For our grass roots community we would want to stay away from dedicated mining hardware and stay in the remit of GPU miners. This would probably take the form of a move to Scrpyt-Jane. At that point Bitcoin and Litecoin could rule SHA-256 and Scrypt respectively while we could rule Scrypt-Jane. This is something to consider in the future but it would be wise to consider all our options as soon as possible. We want to be prepared for the future :)
Ghostlander pointed out that GPU mining get diminishing returns as N goes up. I understand that certain variables in Scrypt-Jane can be tweaked. I would like to do some testing on Scrypt-Jane to see if it is a viable option later on.
Another thing to consider is how this may this may help protect us against attacks.
-
-
I understand that you want to stay away of that ASIC madness taking place in the Bitcoin world currently. I’m not into it either. Please understand that GPU mining has hit its limits nearly. Top performance video cards are expensive, and they also consume a real lot of power. The upcoming generation of Radeon hardware isn’t going to be much faster than the current one while keeping the prices high (Curacao GPU is expected to feature 2304 shader pipelines compared to Tahiti GPU with 2048 shader pipelines, clock speeds are about the same). The market is expanding constantly and it wants new solutions. The only problem is how to produce fast ASIC hardware in large quantity without long pre-orders and make the pricing affordable to average user.
[quote]Another thing to consider is how this may this may help protect us against attacks.[/quote]
Unfortunately, it won’t give us any protection but trouble. Those who want to attack us will find a solution anyway. They have experience and resources far above average. Custom scrypt settings will break compatibility with popular mining software, and I’m afraid many their developers will ignore Feathercoin like they have ignored Yacoin. We need to make mining easier for average user, not the opposite.
-
If Scrypt-Jane was less efficient, couldn’t you change the difficulty algorithm to match the lack in efficiency?
And ghostlander, if we ever come out with a client that includes mining software (don’t know how realistic it is), using Scrypt-Jane wouldn’t cause inconveniences at all as long as you had the Feathercoin client
-
[quote name=“ghostlander” post=“14813” timestamp=“1371221337”]
I understand that you want to stay away of that ASIC madness taking place in the Bitcoin world currently. I’m not into it either. Please understand that GPU mining has hit its limits nearly. Top performance video cards are expensive, and they also consume a real lot of power. The upcoming generation of Radeon hardware isn’t going to be much faster than the current one while keeping the prices high (Curacao GPU is expected to feature 2304 shader pipelines compared to Tahiti GPU with 2048 shader pipelines, clock speeds are about the same). The market is expanding constantly and it wants new solutions. The only problem is how to produce fast ASIC hardware in large quantity without long pre-orders and make the pricing affordable to average user.[quote]Another thing to consider is how this may this may help protect us against attacks.[/quote]
Unfortunately, it won’t give us any protection but trouble. Those who want to attack us will find a solution anyway. They have experience and resources far above average. Custom scrypt settings will break compatibility with popular mining software, and I’m afraid many their developers will ignore Feathercoin like they have ignored Yacoin. We need to make mining easier for average user, not the opposite.
[/quote]Agreed on ease of use. If this was a move we were going to make then we would commission Scrypt-Jane support in cgminer as current miners for Scrypt-Jane are not stable. I actually wonder if cgminer support is something I should get going now. The more mature it is the better it will be for us if we decide to use it as a solution. Scrypt-Jane would not be suitable right now.
For all that we will avoid some potentially malicious redirected Scrypt hashing we will also lose a number of miners who script mining to be the most profitable, this now includes some large auto pools. If the hacker actually has a warehouse full of 7970s then we would only be more vulnerable after the move.
There are a lot of obstacles which would require a lot of preparation to handle properly. This may not be a sensible move now but it could still be viable in the future.
-
What ghostlander said If I am tbh - Make/Try-to-make something attack proof and you will find a better attacker, in fact you may awake a better attacker ( a bit like giving a soggy box of matches to an arsonist).
Not sure about ASICS though, I agree with the energy factor but if the value increases in line with the difficulty, but granted - what not [b]if[/b]. :-\
-
Lots of valid points, thank you for the contribution so far.
"Custom scrypt settings will break compatibility with popular mining software, and I’m afraid many their developers will ignore Feathercoin like they have ignored Yacoin. "
I’m not as worried about this statement. We would work directly with those involved to integrate whatever solution we eventually move to. I have nothing against Yacoin, in fact, I’m grateful we were able to open up communication between the community. With that being said, they didn’t have the same resources that are available to us in terms of #s, developers and financing.
" Please understand that GPU mining has hit its limits nearly. Top performance video cards are expensive, and they also consume a real lot of power."
I would agree with this statement. I would worry more about the big rigs coming from Bitcoin to Scrypt-based coins in the near future. If you take a look at Bitcoin’s difficulty, we’re already starting to see the jumps.
“We need to make mining easier for average user, not the opposite.”
We’re in agreement with this statement as well. As a general statement, no matter what path we take in the future, this needs to be a prime focus. I know the guys in the tutorials sections are doing a fantastic job moving that direction. We need to hit a point where we have a pretty little apps where you can mash buttons and easily get rigs rocking and rolling. I’m confident if any coin will get there, that will be us. As someone mentioned in one of the Linux threads, we’re like the Ubuntu of Linux.
Good info, keep it coming. I know Zero talked about a dynamic range for N, I’d like to hear more about it.
-
The ASIC hobgoblin is a potentially liberating technology that could be in the hands of all the people but was executed poorly.
What should have happened is the sales of multi purpose barebones + programmable control boards. What we got was specifications that were too tight and put the average person at the mercy of scammers through ghost bids and fraud pre-orders.
The Klondike Blades and ASIC Miner block erupters are the way to go, but you’re stuck with product that can’t be repurposed and you have a company with a ginormous investment which is self defeating.
If we do get into ASIC hardware we have to do as ghostlander suggests: make it as easy to acquire as possible. As far as simplicity of distribution and power consumption, ASICs cannot be beat. But the potential early earnings have made them difficult to acquire because people want easy money. No one in the bitcoin community is willing to reconfigure their machines from 3x 60 Ghs to 2x 80 Ghs + 1x 20 GHs units.
It’s all ADHD droolers.
-
Litecoin gained a lot of value when ASICs started to appear. This was predicted and was once of the reasons I spent my time mining Litecoin. I expected it would become a big thing with GPU miners.
I’m wondering if the same thing would happen to the next coin with a different hash algo after Scrypt ASICs arrive. I expect this could happen again if there is an attractive coin for GPU miners to move to.
-
I know most of this discussion is above my pay grade but I’m for an alternate scrypt. You won’t lose my measly 3MH hashes if you switch(as long as there is a miner). I feel like the time is now to more differentiate us from others. We already have a base set of users though do know that such a change will most likely drop the nhr(network hash rate) to around 200 MH or whatever the equivalent is for jane or alternatives.
Heck such a move has me more confident in the coin. I might actually build another rig. It seems to me every sort of legitimate coin that comes out has a similar copycat to steal it’s thunder. With so many just trying simple tweaks to litecoin to see what works, we have a flood of coins.
I’m just a lowly miner 49er…
-
I think this might have drawbacks.
I am not so sure that GPU computing, as it stands, will be viable and available long term for end users. Limiting ourselves to devices created by AMD/Nvidia, we really are limiting ourselves to the whims of 2 companies. This isn’t much different than avalon/bfl/etc etc other than, FOR NOW, that hardware is available, is that always going to be true or in 3-4 years will we need specialized hardware anyways?
In addition, mining started on CPU’s, then to specialized hardware (GPU) and now more specialized hardware. Why is that such a problem?
Just my 2 cents.
-
[quote name=“jeremiel” post=“14845” timestamp=“1371223908”]
I know most of this discussion is above my pay grade but I’m for an alternate scrypt. You won’t lose my measly 3MH hashes if you switch(as long as there is a miner). I feel like the time is now to more differentiate us from others. We already have a base set of users though do know that such a change will most likely drop the nhr(network hash rate) to around 200 MH or whatever the equivalent is for jane or alternatives.Heck such a move has me more confident in the coin. I might actually build another rig. It seems to me every sort of legitimate coin that comes out has a similar copycat to steal it’s thunder. With so many just trying simple tweaks to litecoin to see what works, we have a flood of coins.
I’m just a lowly miner 49er…
[/quote]As we hash out (he he he) technical details, I want to make sure this idea isn’t lost in the shuffle. As long as we’re under Litecoin’s shadow, we’ll continue to be seen as it’s knockoff. It amazes me how much credit they get for implementing scrypt, but alas, that’s a different discussion.
As I read more up on scrypt jane, it appears to have fairly modular design. I’d like to hear more about that from the community.
-
I’m also interested in the idea of a different algorithm. My concern I guess is how much of our hashing power is community based and how much is transient across coins. From my own perspective when I was switching coins I was mining a month or so ago I found myself avoiding going back to SHA256 ones just because it was easier to stick with my scrypt set up. I wonder how much we would lose and not recover.
But saying that maybe it’s not a problem if the approach is very much that we are building scrypt-jane from the ground up and are unconcerned that initial hash will take a heavy hit when changed. A good opportunity could be missed by not taking the risk with a new algo.
-
[quote name=“randomdef” post=“14852” timestamp=“1371224411”]
I think this might have drawbacks.I am not so sure that GPU computing, as it stands, will be viable and available long term for end users. Limiting ourselves to devices created by AMD/Nvidia, we really are limiting ourselves to the whims of 2 companies. This isn’t much different than avalon/bfl/etc etc other than, FOR NOW, that hardware is available, is that always going to be true or in 3-4 years will we need specialized hardware anyways?
In addition, mining started on CPU’s, then to specialized hardware (GPU) and now more specialized hardware. Why is that such a problem?
Just my 2 cents.
[/quote]Change doesn’t come for free. Abrupt change reconfigures and adds stresses to the community. It’s not pure math. It’s economic physics and chemistry.
There isn’t a sufficient diversity of economic philosophy and strategies to sustain it. ASICs will lead to the Bitcoin Alzheimers Constipation Kerfluffle, at which point it will crash to around $20. The more coins one group of people make reduces (via difficulty) the amount of coins others can make. This splits the community into people who crave more coins and want to buy more and people who give up and go elsewhere demanding to be treated as kings in the alt currencies. The ones who have then cut the flow driving up the price. This works until Wordpress says this is too weird, Amazon Gift Card prices go random, Charities have no idea where the price will be, and the banks start lobbying Congress to have them manage the Bitcoin madness, because clearly the geeks can’t.
That’s the problem.
-
What happens if ATI/AMD alters their APP in the future and no longer is efficient at scrypt mining?
-
[quote name=“zerodrama” post=“14867” timestamp=“1371225797”]
[quote author=randomdef link=topic=1839.msg14852#msg14852 date=1371224411]
I think this might have drawbacks.I am not so sure that GPU computing, as it stands, will be viable and available long term for end users. Limiting ourselves to devices created by AMD/Nvidia, we really are limiting ourselves to the whims of 2 companies. This isn’t much different than avalon/bfl/etc etc other than, FOR NOW, that hardware is available, is that always going to be true or in 3-4 years will we need specialized hardware anyways?
In addition, mining started on CPU’s, then to specialized hardware (GPU) and now more specialized hardware. Why is that such a problem?
Just my 2 cents.
[/quote]Change doesn’t come for free. Abrupt change reconfigures and adds stresses to the community. It’s not pure math. It’s economic physics and chemistry.
There isn’t a sufficient diversity of economic philosophy and strategies to sustain it. ASICs will lead to the Bitcoin Alzheimers Constipation Kerfluffle, at which point it will crash to around $20. The more coins one group of people make reduces (via difficulty) the amount of coins others can make. This splits the community into people who crave more coins and want to buy more and people who give up and go elsewhere demanding to be treated as kings in the alt currencies. The ones who have then cut the flow driving up the price. This works until Wordpress says this is too weird, Amazon Gift Card prices go random, Charities have no idea where the price will be, and the banks start lobbying Congress to have them manage the Bitcoin madness, because clearly the geeks can’t.
That’s the problem.
[/quote]Issue is, in my mind, ASIC’s hasen’t been abrupt, just very kind to early adopters. If anything, the introduction of this alt-coin and many others, has been abrupt.
I didn’t understand anything else you wrote, you need to be a little more dumb and layman’s terms for me.
-
Coblee believes that there is only room for one coin under each hashing algo. Any other coin will see its difficulty rollercoaster which seems true for us right now as we swing between 41% swings up and down. For now we need to share Scrypt with Litecoin but when GPU miners need somewhere else to go we might want to be that place. I’ve always thought that Litecoin had some miners with deep pockets prep the market when the miners started moving over.
Perhaps we would be better of becoming the dominant Scrypt alt :)
-
https://bitcointalk.org/index.php?topic=234287.msg0;topicseen#new
Check it out!!
??? -
[quote name=“Bushstar” post=“14886” timestamp=“1371228649”]
Coblee believes that there is only room for one coin under each hashing algo. Any other coin will see its difficulty rollercoaster which seems true for us right now as we swing between 41% swings up and down. For now we need to share Scrypt with Litecoin but when GPU miners need somewhere else to go we might want to be that place. I’ve always thought that Litecoin had some miners with deep pockets prep the market when the miners started moving over.Perhaps we would be better of becoming the dominant Scrypt alt :)
[/quote]Maybe when this project started, being an add-on for Litecoin was fine. Obviously, Feathercoin quickly became stone soup… we attracted some amazing talent and here we are 1500+ members deep.
“Coblee believes that there is only room for one coin under each hashing algo.”
This says it all. It’s bad enough to watch my investment fall under 2.2 without any massive market pushes, it’s worse to not even be accepted for the strengths Feathercoin brings to the crypto community.
- How many new miners have we helped (me being one of them) ?
- How many people that would have been shunned away from other forums found an interest in crypto because of Feathercoin?
- How many people have worked on projects because they enjoyed the resources and helpful attitude of this community?
- How many giveaways did we have to get people involved?
etc… etc… I get it, we’re not ‘nerd cool’. Let’s stop pretending that being that other coin is good enough… that’s my 00000010 cents.
[b]I say we compete. [/b]Keep our doors wide open for future cooperation, but compete in a new category. Ubuntu is hated by the hardcore Linux guys because it introduces the platform to the novice user, it’s loses it’s ‘nerd cool’. Feathercoin also has a very similar community, one that is open to new ideas and works hard at making things ‘friendly’ and ‘easy’ for the non geeks. I know some developers have trouble understanding how that adds value, let’s show them.
