Topic: [Ultracoin] [Est. Feb 2014] ~ ASIC Resistant & Ultrafast 6 Second Transactions! - page 140. (Read 381003 times)

The opposite of your statement is closer to the reality.

Since most GPUs from high end to low end are pretty much on equal footing at NF15+, with speed more or less dictated by their amount of RAM there is less need to perpetuate the arms race of newer better faster more expensive hardware to remain an effective miner.  Since there is not a continued requirement to invest $ into new hardware there is also less need to dump mined coins to payoff the capital costs.

Some continued arguments for Kracko:

I think the equality in mining higher NFactor chacha is a very marketable feature.  Both for hardware costs and power usage.  At higher N the all GPUs are drawing less power as NF increases, but this is particularly noticed on the lower end cards.  My entire 6x R7-240 4GB system currently pulls only 150w at the wall (I believe it was 180w at NF15).  At NF15+ this system is hashing just as fast as a 6x R9-290X system for a much less ridiculous capital cost and runs quiet and cool as a cucumber.

I was recommending NF15 over 16 since there are many more people who have 2GB GPUs in their computers (or mining systems) than 4GB.  At N15 these are still functioning at decent efficiency.  At NF16+ the 2GB cards take a big hit in mining effectiveness so there would be many less people willing to mine.  Since I have mostly 4GB cards I am making this recommendation not for the benefit of myself but what I think is potentially most beneficial to UTC.  N16 has been treating me very well.

So at NF15 you have a unique currency with 1) the most egalitarian mining hardware requirements of any coin (ie almost anyone with a PC can mine it), 2) has among the lowest, if not the lowest, power usage requirement for GPU mining of any coin and 3) remains asic resistant and 4) is botnet resistant since it remains far more efficiently mined by GPU and not CPU.  That combination of features is the golden egg in my view that so far has eluded others.

All that said NF14 would still be ok in my view...just not as perfect.  But as you stated you are also working toward lite clients to function on android and other weak systems so if that goal is not obtainable at NF15-16 then that is a valid enough point.

It isn't a question of whether YOU are "fine" with miners using their same GPUs or YOU preferring to not make miners lives more difficult. And I say that as an avid miner myself. Mining profitability is really a side point because as a miner, one can mine and dump a coin with low market cap/price vs a coin with high market cap/price just the same. The same principle goes for the block reward. I will mine the most profitable coin (generally) regardless of price, marketcap, block reward.

Let me ask you something... do you think it is a coincidence that when mining advancements came out for bitcoin (sha256), the price of bitcoin increased substantially each time? It is a legitimate question I think.

I was responding to your "Convince me" comment on which NFactor to remain on. The idea of removing the coin from the schedule of NFactor changes is a different topic.

Lastly, I get a sense that your personal attitude is part of your decision. As someone who apparently has the power to single-handedly change the block reward and algorithm of this coin, I feel that would be an issue. In particular, if I told you that NFactor 15 would be more profitable for my mining rigs, I feel like that would influence your decision to stay at NFactor 14. Am I wrong or right about that?

This is very much a coin that benefits miners who are just dumping rather than staking. If we can stop miners dumping then we could start to see a slow steady rise. As an investor I don't see any benefits at all in investing more of my money into UTC. Could you go down Bitbays route and peg it in the future maybe. Making it stable. Stability is what people are crying out for and would attract a lot of investors

I dont get how nfactor causes the slowdown? please can you explain in layman terms.
I understand the security issue as the host device wont be checking the hashes but relying on a third party for that. That wont be good & somewhat centralized.
Also i will suggest that we should try everything before kind of "rolling back" nfactor. Its a new challenge & i hope we can solve it with least compromise & won't take the easy path. Thanks for replying

I think your "starttime" is wrong. Try making it 1388361600, Also make sure you have "gpu-threads" : "1”. And then, cut it down to a bare minimum to test. Get rid of Expiry, Queue, Scan-time, anything other than temp control which is really not vital. If you get it working that way, you can slowly ad them back in.

Good luck!

Not particularly.

I hear what you are saying and understand your position of maintaining the nFactor schedule as it is integral to the original design of our coins.  I just don't believe that the progression of Scrypt-ChaCha's nFactor is worth keeping.  It's not like it has worked so well for YAC.  I like the concept of keeping up with hardware advances, but the schedule leaves much to be desired and causes more havoc with each increment than it's really worth.  Being forced to upgrade hardware to keep up with an arbitrary schedule is actually less profitable for miners. I am actually fine with those people using their same GPUs.  Whether incrementing the nFactor or not, as long as it is profitable there are going to be miners whose sole purpose is to mine and dump your coin.  As non-altruistic as they are, they are still providing a service to the network.  I would prefer to not make their lives more difficult- at least to the point to where ASICs are discouraged.  There will still be an advantage for those that upgrade even staying at one nFactor.  But, they should be competing against each other more than against ever increasing memory requirements.


Being more ARM friendly was only one of the reasons.  But it can be worked around by sacrificing security through trusting and offloading the hashing requirements onto external servers.  It's really the initial synch that's the issue.  Validating hashes on individual blocks for mobile devices isn't that bad.  Hashing over a million of them is.  Things slow down quite a bit around NF12/13.


I'm glad your experience with the Nfactor increases has worked out so well.  I find it's only a momentary hassle to find new settings that work with my hardware as well.  Unfortunately this is not universal.

This isn't a hockey game, this is crypto.  The game and even the rules change all the time.  And many coins have reinvented themselves to successful results.   It's not about adjusting the coin for an "insider" group, it's about making the coin flourish, which benefits all of us.  Stagnating along the same path as YAC is not going to be very good for the future.  We need to differentiate ourselves- carve out our own niche.

It is a very valid point!
However, to play devil's advocate, the marketcap of UltraCoin (particularly before this talk started) is such that changing the rules mid-game doesn't have a lot of downsides because the game is small-time and not a lot of people are playing it--to put it bluntly. If the marketcap or the influence of UTC was much higher, your point becomes very, extremely valid.

At the same time, there is a good reason to not stay at NFactor 14 which I mentioned above. It should be worrisome that what I said is not strongly taken into consideration. I have also warned about reducing the block reward; I have also warned about the pool fee/orphan rate. Alas, my warnings have been and will continue to be tossed aside and labeled as 'trollish'.

I have learned through this whole crypto experiment that 'leadership' and 'management' of a coin is something that will ALWAYS be a huge focal point. Despite the 'decentralized' sell of bitcoin and cryptocoins, humans, even the seemingly intelligent ones, need some higher power, higher authority figure telling them what to do or leading them to some end. I can accept this notion because competition can breed those in leadership positions to make the right decisions or fall off in obscurity.

The question for UltraCoin is... do you think the leadership is strong and wise and has the best interest of the coin long-term for old and new investors. I have my own opinion, and I don't need to share it.

I think you have a valid point there. Tinkering with preset parameters sets a mindset that any change is possible with an excuse. I am not saying that things should not be changed. No one could have seen that NF16 will be a problem on mobile device but we should try other ways to cope rather than tinker with a very central parameter to whole strategy. NF16 is what allows CPU & GPUs to be equal. That is a central aim for this coin. Now we should not drop that to be mobile friendly.

I am not a dev but is it possible to change POS algo so that mobile devices can stake w/o problem or use of latest bitcoin 0.10 way of verifying blocks to ease the problem on ARM processors if we are not already using that?

Regards

Sam

Personally I do not understand at all this hassle with N.  
There was sheduled N change. All know what day and hour next N change. But now, it is like "who knows, it depends weathers"

Now someones want manipulate this N. Perhaps soon some peoples are not satisfied and other day other peoples are not satisfied and then agen N change out from original schedule. Who can trust and what?

YAC coin  have changet to N15 as scheduled, then it have changed to N16 and no problem, works like charm. Just aftrer some days it is going to N17.  

Why N16 is problem with UTC?  Perhaps next day some group of peoples want change algoritm also. After this N factor game who can trust what is coming next. After UTC launch there have been N time schedule. Today it can think this schedule was "perhaps" schedule or just fake schedule. When some "insider" group start adjust coin for them selves it do not promise very good for future.

My opinion is that this game with N value is mistake.

How about hockey where you should change the rules mid-game. Or, if a hockey goal, the location could abruptly change if someone wants.

The newer CPUs and GPUs coming out are shifting to higher cache/vram. When new technology comes out, it will definitely help your coin to be the most profitable for those who are willing to spend more capital on new products. Otherwise, people who have had their GPUs for years will just simply mine and dump your coin, and they won't even have to worry about ROI at this point...

Convinced yet?

One of the things I'm aiming for is an algo that is more friendly to mobile devices.  NF14 is still pretty slow on ARM CPUs just validating blocks on the download.  If it were completely up to me I'd go with an even lower, maybe NF12.  Having a unique algorithm is protection in itself- I'm doubtful that there's enough reward for someone to go through the cost and effort to develop and produce an ASICs device for just us at this point in time.

Any change is going to disadvantage someone.  Why shouldn't those with the higher end cards get a little bit of an advantage?  After all they did pay for it.  But NF15 or any other Nfactor isn't off the table.  Let's hash it out.  Do you have any more reasons to choose NF15 other than not to disadvantage those with low-end GPUs?  Convince me.

+1 on the NF change (heck, I love NF16 though!)

Set it where? In ATI Settings Calculator calculator? It gives me:
yacminer --scrypt-chacha --worksize 128 -g 1 --lookup-gap 8 --buffer-size 2590 -R 5120 -o stratum+tcp://: -u -p --nfmin 4 --nfmax 30 --starttime 1388361600
With that I'm only getting HW errors and less than 1Kh/s per card.

Edit1: ok, will try to set nfmax 14.
Edit2: it seems to be working. Tnx.

You'll have to set your nfactor max at 14 (I think it's --nfmax 14) as it strays from the normal calculation from start time.

cudaminers use --algo=scrypt-jane:14 instead of --algo=scrypt-jane:UTC for the same reason.

Kracko, what are the correct settings for R9 280x? Starttime correct at 1388361600? Trying to mine but yacminer shows NF 16 and getting pretty lame results. Less than 1 Kh/s per one 280x.

That's unfortunate that Nfactor14 was chosen as that shifts the mining back in favour of higher end GPUs rather than the unique niche of Nfactor15 that made low end GPUs on par with high end GPUs.  N15 is also a sweet spot as 2GB cards (high or low end) are still effective unlike N16+ that really starts to favour 4GB cards.

Nice to see the reduced reward to 10 from 30 though.

The test pool is up.  Nfactor14 folks!

Feel free to abuse it- in fact, I encourage it.  If the oscillations continue with the longer smoothing times, we'll tweak it some more.   Hopefully 240 blocks will average out the highs and lows better and also lesson the chance of orphans at the lower difficulty swings.

This is what we'll be rolling with moving forward.  The sources on Github are updated, please see the may2015 branch from http://github.com/Kracko/ultracoin-2  as that is what is currently running behind the test pool.

Here is the link to the pool:  http://test.tumblingblock.com

It's in progress, along with multiple other changes.  For one, we're smoothing out the retarget from 10 blocks to 240.  Hopefully that takes care of the oscillation and high orphans at the low difficulties with the increased averaging time.  Once I have a working prototype ready I'll throw it up on a test pool and let you all have at it.

what miner and what GPUs ?

for the life of me can't mine this coin on any of my 6900's 7900's or r9 290's

just driver crashes or zero accepted shares , sigh