Topic: Question for GPU miners (Read 7026 times)

... so the ASIC miners are going into the hands of a few people (or they could be used by the ASIC mining manufacturers themselves (mining BTC to offset/lower their production costs) before being released to their customers), and that means that ASIC units are going for $3800 for a 5000 MHash/sec unit on Ebay http://www.ebay.com/itm/Butterfly-Labs-ASIC-5-6-GH-s-Jalapeno-Bitcoin-Miner-IN-HAND-NOT-PRE-ORDER-/330944064052?pt=LH_DefaultDomain_0&hash=item4d0dcdaa34  (as opposed to anywhere near their $274 list prices).

One possible solution is to use the most advanced FPGA technology - 28 nm FPGA - which can get you 2000 MHash/sec for $1000.  http://www.raspberrycoins.com  (The ASIC miners are using 65 nm technology node - they are limited by how much time and technology a Fab will provide them - (and the Fabs give their best customer services to companies like Qualcomm, Nvidia, Altera, and Xilinx - who order hundreds of millions of chips) ).  The FPGA prices will continue to drop over time (Altera recently announced a 14nm and 20nm nodes, and they're a little overdue to announce a Cyclone 6).

There were many times that FPGA's were considered too overpriced, but these companies just keep pushing Moore's law consistently ... (and companies like Altera and Xilinx can afford the 7 figure ASIC NRE's for the newest process nodes), and they make standard products which gets sold to a myriad of applications (and those volumes again - lowers cost) ...

... so between the greed and/or incompetence of certain ASIC mining manufacturers who are over a year late to market, and the steady execution of professional FPGA chip vendors, I wouldn't quite count out the FPGA technology just yet ...

more like asicminer selling bitcoins as soon as they get them.  wouldn't doubt a majority of the people that are wasting money on these crap devices that asicminer has made available (good lord, i wonder why?) trying to recover money in a futile effort

You mean ASICMiner, Avalon and BFL? I think they try each on their own to make their profit. As the bitcoin is quite stable in the last time I think there isn't anyone manipulating the market yet. But I see the 20% of the networkhashrate of ASICMiner as a big problem. There are pools that have a higher hashrate now, but in a few months when most of the GPU miners will have stopped bitcoin mining and ASICMiner bought even more ASICS, then they might be up to 40%. And from 40% it isn't a long way to 51%...

If they realy would want to merge I think nobody could stop them. But I hope that they are smart enough and know that they would kill the bitcoin with that.

We are likely already there. 25 terrahashes is my conservative estimate..  http://www.coinchoose.com/

The gpu miners have been conned into the alt-coins, very much the same way that the Asic Dream did. Just hearing them all trumpeting the promises of whatever crypto-currencey god(s) they believe in, makes me want to throw up. More now because I've learn that 1-2 of those alts are starting up a ... (pardon my foul mouth) BANK.

@SirMintALot

Is there any evidence of collusion between the big three ?

What is to stop them from merging ?

We are going to have a bitcoin oligopoly, with only a few getting the big amount of coins and being able to control the price of the bitcoin. We are now haveing a market that needs a constant flow of bitcoins to keep their bussiness (shops, gambling sites, etc.) running. If a big player like ASICMiner, with already 20% of the hashrate now, decides that they won't put their coins on the market for a while, then the value of bitcoin goes up. They also could decide to throw their stack of coins on the market causing a bitcoin crash. After that they could buy cheap bitcoins and controll the market even more. Just imagine Apple would do the same with their stock shares to get richer and richer... Luckily we have stock market rules that prevent this. I agree that Satoshi probably never thought about the possibilty that an oligopoly might rule the generation of coins and the market and it's now to late to add regulations to keep the market healthy and fair (like a transaction cap: nobody could do more than 5% of all market transactions, if you have to do more transactions then you have to wait. And a coin generation cap, nobody should be able to generate more then 10% of the coins. If one generates more than 10% of the total hashpower, then all generated extra coins are given away to all other miners). We need social market economy instead of turbo capitalism to keep trust into bitcoin  
Let's face it: Now we are in the same situation as mom and pops little general store when the big whole salers (like walmart & co) crushed the market. We can either believe into bitcoin and put all our money into things that generate more hashpower (GPUs, FPGAs and try to get ASICs) or sell our little general store and do another bussiness (Litecoin, PPCoin - not one of the bullshitcoins), which isn't controlled by a hand full of big players yet. But history will repeat when Litecoin mining will become profitable. Sooner or later we will see scrypt FPGAs and ASICs when the Litecoin market is big enough. PPCoin with it's Proof Of Stake is a good concept to prevent pump'n dump of the small miners (why should they invest lots of time for selling their coins when they get an interrest rate for holding them for a longer time) but the big boys will still do pump'n dump. Freicoin on the other hand isn't the real McCoy either: 4.89% demurrage is to much to get the small miner interrested, but the big boys with lots of hashing power will dump a lot and controll the coin. But I believe we are still in the time where lots of concepts are tried on the way to find the ultimate stable and trusted coin. I think we need a new hashing algorythm (like SHA3, but there is already a GPU miner for Copper Lark. Scrypt is no option either as there are people already working on FPGS and ASICs), Proof Of Stake to stabilise the market a bit and market rules (trade cap, coin generation cap) to prevent an oligopoly or even monopoly. Until then history will be repeating....

 This will never happen.

Nor should you ever underestimate the power of one... lowly gpu.

What would happen if everyone set one or more gpus per rig to solo mine ? How many getwork/s will be diverted from the asics if 10,000 nodes popped up all of a sudden ??

It might be like watching a swarm of locusts devour the harvest before the big bad tractors get to it.

Well said. This precisely what I've suspected. ASIC is a curse for bitcoin.

This depends on the general public's perception on things as they are responsible for much of Bitcoin's jump in value. The Cyprus haircut was a big part of that.

Will they continue to invest in BTC when the true scope of the pre-order centralize Bitcoin scam becomes more public ?

The coins and the hashpower have been hoarded. The Crypto movement has been damaged beyond repair.

But what about this.

Bitcoin is only worth something because people say it is.

What happens when all the GPU miners (70%+ of miners) move to scrypt and say Bitcoin is worth nothing?

Then what will the asic guys do?  Trade among themself? Polish there shiny useless hardware? Cry over it and short it out?

In the end they could be left holding the bag and I think that would be great

A fair point - Personally, I expect the block creation rate to be increased by about 4x and the difficulty and block reward adjusted down accordingly (as soon as the network hash-rate grows past some randomly selected number). This needs to happen for the network so that we can actually get to a point to compete with large payment processors. It would also help with adjustments to global hash rate in that the difficulty would re-target much more often.

You are right - I did go off topic.

Even changing the block interval will not solve the problem - it will have less of an impact but there is a limit to how quickly you can make the difficulty adjustments. Also this would mean the difficulty jumps would be way more often and much higher and these guys would just expand their operations even faster.

The thing is this - Never underestimate the power of human greed.

And an explanation has already been given so thanks for that!

But the problem (as you stated it) is not the centralization of power per se. You said the problem was that ASICs will devalue BTC because they will make it easy to produce new coins. So I asked how is this technically possible and you explained to me the theoretical possibility of inflation-spikes in between difficulty adjustments.

I then proposed a code change to alleviate this issue, to which you went on a tangential rant about the centralization of power again...

? errr - I am chilled - I thought you wanted an explanation!

The changing of the code does not stop the centralisation of power - the inability to get your hands on an ASIC is.

Dude chill down. All I'm saying is a change to the code may be a solution to this problem (i.e. influx of "easy" coins into the market in between difficulty adjustments)...

Dude the problem is not the code - it has worked fine till this point.

What has messed it up are ASIC's and the way they have been distributed. Basically only a select few have managed to get their hands on this technology - I would guess around 10% of miners are using ASIC's yet they are probably 70 - 80% of the network hash rate.

It has caused a concentration of power into the hands of a few.

The increase in performance has blown Moores law completely out of the water - Satoshi never envisaged this problem - the idea was Bitcoins were available for anyone to mine - this is no longer the case.

If everyone could order ASIC's next day then maybe it would not have such an impact - like the change form CPU's to GPU's.

These guys developing the ASICs have gone so against everything Bitcoins stand for.

If they had stockpiled enough before making them publicly available (and not using them to mine themselves AVALON) it would not make a difference. But their greed was put before the mission of the Bitcoin project.

ASICminer are a perfect example - like I said in a previous post - One company should not be able to impact the entire network.

If they have a major outage - block times would be dramatically increased.

Look and see how they are expanding as well. ASICMiner are almost like a central bank now.

OK so it seems the devs may need to look into decreasing the required number of blocks before a difficulty re-calculation, in order to avoid this theoretically detrimental issue?

It only adjusts every 2016 blocks - if it takes 1 year or 1 hour to mine them. If you doubled the network hash rate at the beginning of the difficulty change that limit will be reached in 1/2 the time.

The target is every 10 mins - approximately 2 weeks

Not sure what you mean. I thought the sole purpose of BTC mining difficulty adjustments was to keep the influx of new coins into the system steady... Please correct me if I am wrong.

If that is true please give me the exact time and date of the next difficulty adjustment and the difficulty.

If ASICMiner goes offline again - causing a 25TH/s  drop in network performance the difficulty adjustment may take twice as long

http://www.asicminercharts.com/

it doesn't

last difficulty cycle I believe the avg was around 7 per hr? (ed: hmm more like 7.5)

Yea but the rate of new coins remains steady no matter how many ASICs there are and how efficiently they are mining... So applying the "If you can produce something more efficiently it has less value." formula to the value of BTC is wrong.

the available amt of coins on the market is always increasing

and difficulty adjustments dont occur instantly, either

i've talked to a few people and my theory is the recent drop was the ppl with the avalon 2 units selling as fast as they can.  the continuing pressure for the last few months is probably just asicminer selling 25btc every block it gets

No you are wrong - And yes they can - if the Network hash rate grows exponentially it will always be ahead of the difficulty.

If the network hash rate trebled today the block time would be reduced so more coins would be produced.

Umm, it's not like ASICs can increase the amount of available coins on the market so no you are wrong.

The thing is I can order 40 gpu's for next day delivery.

You can not order ASIC's for next day - you are in a queue with no guaranteed delivery date - could be weeks / months and years (BFL). By the time it arrives the difficulty could be so high that they don't even pay for themselves.

It puts such a big question mark over your ROI that you are unlikely to place an order. Only the very early adopters stand to benefit making mining an exclusive venture.

Asics are actually cheaper then GPU's... and therefore more available to the masses.

ASIC's will devalue bitcoin's

If you can produce something more efficiently it has less value.

Look at Gold - Gold increases in price due to mining costs.

If someone came along with a machine that could just suck the gold right out of the ground and did not need to mine - what would happen to the gold price?

The distribution of ASIC's has totally upset the balance.

If scrypt coins require more power and resources to mine - long term they would have more value.


Gold is rare and difficult to mine - hence its value.

If I developed a quantum computer and it mined all the remaining Bitcoins in 1 week then what? Do they still have the same value?

To further the analogy if gold could only be mined by hand tools its value would be substantially more than it is now.

Cryptsy

Just found this... and that.. lol
https://bitcointalksearch.org/topic/would-fpgas-work-for-mining-litecoin-91210

http://blockburner.net/

Where did you hear that?

Makes sense and +1 azw409.

Pardon my noobility but..

Aren't we just a firmware update or fork away from scrypt asics becoming a reality ?

Its very obvious to me how the current gpu exodus/ alt coin frenzy and the gpu sell off of 2012 were both acic induced. Its a conspiracy to centralize BTC and eventually the entire crypto movement.

With the much of the competition shut down or diverted, asic owners are more than happy to see the difficulty stay as artificially low as it is.

@ all alt-coin miners..

If you think 15M diff is high. How are you going to like 30M after most of these useless alts get asic'd and/or become worthless ?

Agreed

But I guess if it helps me sell off the rest of my 5xxx...  I have 4 left out of like 20 a month ago.  

ed: oh, i'll still be profitable next difficulty cycle, assuming price stays > $100..... about $400 after elec.  after figuring in "standard" equipment depreciation, $200-$300?  not worth the effort, but I guess it's not a loss waiting for stuff to sell

I agree, I've been trying LTC, but many of the blocks don't even contain a single transaction so it's an utterly pointless activity. Where's the value in a currency that you can't spend and nobody's using ?

Unconvinced, and first - even if they are, they are unlikely to be a significant performance/power advantage over GPUs, and second - mining, no matter what the means, efficiency, power, resistance, mechanism etc. does not create an ecosystem in and of itself.

If what I've heard is true, Fpga's are already mining scrypt.

I agree - moreover, scrypt coins are doomed to failure by their very design. You cannot design a digital process that someone isn't going to be able to build hardware... to perform faster for less power. So if litecoin (for example) ever saw some decent numbers in it's price (say between 7 and 15 bucks a coin) almost immediately someone would come out with an FPGA to mine it and the entire thing would go tits up.

I chose to sell off my GPUs this month (yes they're already sold) - I'm not interested in mining alt coins with GPU. Frankly 2+ years of that noise should be enough for anyone. I have not (as of yet) replaced any of my lost hash rate with overpriced and under performing (but instantly available) asic devices. I'm debating simply holding the bitcoin for awhile. I do have early orders with BFL for enough singles to not make the missing hash rate matter if bfl delivers this year. Until then (or something else becomes quick to get and has good power profile) I'll be fine with my reduced electric bill.

The only rule for buying mining hardware now is most mh/s/j wins.

We're just getting out of the 'bad old days' of bitcoin mining and I don't understand miners who are willing to put themselves through that again for an alt coin that's clearly inferior to bitcoin.

I agree, and then difficulty will drop until enough people are making a profit, somewhere someone will be mining scrypt profitable but unless you have cheap power and are prepared to optimize your rigs for efficiency I wouldn't bother. Worst case scenario scrypt mining in 6 months consists of people with free power looking for pennies on the bottom of a very deep well, but I doubt it will get that bad.

The thing LTC has going for it is when/if mtgox finally implements trading and fuel another bubble for a few weeks/months, long term I don't see even that saving profitable scrypt mining for most people however.

How is payout time for a GPU 400 days? Wow, that number is really off. At a modest estimate 180 days, at a good estimate 120, anywhere in between is fair game.

If you have serious GPU power, then you just hop on every morning, check how the alt coins are doing, have a scrypt/bat file setup for each type of coin, I just hop on my ssh tunnel from my laptop, 3 lines later I'm mining whatever coin is most profitable, and I ride it till it changes, I usually stay on a coin for 24-48 hours. If some of the alt coins are being flakey with exchange prices (ie digitalcoin) then maybe I'll hop on LTC for some stable income to pass the time, and then when something flys back up in profitability, hop over.

Obviously mining isn't as glorious as it was say 45 days ago, when I could have easily made $120/month per 7950 with LTC, now that number is about $84 if I hop alt coins every day which is pretty sad, but hey, I'm still making plenty.

Especially because I don't pay for electricity and have infinite access to all the power I could ever need lol

Put it this way, $12,500 deep. 5 rigs, x6 cards each, each card hashes at 675. That returns $2500-3000

And even if I was paying for power, that would shave off about %35.

I don't think that's bad at all.

I usually buy new because most people are asking for too much for their used gear or they seem sketchy. lol The warranty and the $$$ from selling the game codes, makes spending the extra 10-30% a no brainer.

Currently payoff time for a GPU is about 400 days. There should be lots of used GPUs for you to buy.

Cablesaurus.com has a cool Bitcoin\Litecoin\USD checkout system. You can check the price in each type and pay in whatever is best at the moment. Its the coolest one I have seen so far.

Yes, of course it will but this isn't about difficulty it's about democracy. I know mining scrypt will be just as competitive and probably only marginally profitable at best but it will prevent the elite ASIC owners gaining a stranglehold and killing bitcoin. I'm not whining that my low-end ATI card isn't making me a profit anymore, I'm concerned about the bigger picture in that thousands of GPU miners will soon be abandoning and basically leaving just BFL and ASICMiner to do all the mining between them. One of them could easily gain a 51% share but even less than this makes bitcoin a dangerous place.

The old argument I always go back to is, show me one place I can spend litecoins besides just trading them for bitcoins or fiat and I'll believe they are a real thing.

There's always the possibility of scrypt-based litecoin being a healthy and/or more stable cryptocurrency than bitcoin...

What if:

• Bitcoin fails because of internal political disputes among devs and/or userbase, or perhaps a critical flaw is discovered which scrypt is not affected by
• Litecoin thrives because of adoption because of some arbitrary organization, social and/or political advocacy, etc. etc. etc.

Either scenario seems plausible to me... Or maybe they could just coexist like VHS and DVD, or sega dreamcast and sony playstation.

The surprise for scrypt miners will come when mining scrypt is not profitable either. Who would trade ASIC earned BTC for GPU earned LTC at an effective loss? Yes I do believe the LTC economy is just an accessory to the BTC economy and can't stand up on its own right.

Lets face it, BTC is going to die sooner or later for GPU miners. As a few have said, anything SHA based is pretty much stuffed.

Then scrypt will face similar problems with the increase of GPU miners switching, difficulty will increase making it harder for small time miners to stay in. Only the heavy hitters will stay afloat.

You think thousands and thousands of GPU's dropping off SHA and switching to Scrypt won't massively increase the difficulty rate of Scrypt coins?

I agree with Hippie Tech, the only salvation for Bitcoin is to prevent ASIC's from getting a stranglehold which is already happening.

This could be done by changing bitcoin clients to allow two proof of work algorithms :-

 sha256 so that all the existing blocks are still valid and those with ASIC's can continue to mine using low power devices.

 scrypt so that those with GPU's i.e. the vast majority can continue mining for blocks using a different difficulty.

The new client will accept either, so it's the miner's choice which to use. This approach would allow those who've invested heavily in ASIC's to realise their investment but it prevents them owning all mining which will happen very soon otherwise. Perhaps it would also encourage the ASIC companies to actually ship their devices to their customers rather than hoarding them for their own benefit.

Altcoins will never wash with the public that have already invested in bitcoin since they'll lose all faith and won't know which coin is the 'official' replacement. I know bitcoin isn't perfect but it would be a shame to see it die now and damage the prospects for any future p2p currency.

I will keep my GPU's on until they are loosing me money, even at the current difficulty I am still making decent money off GPU's. I think if you want to be in the BTC mining game you need to get ASIC asap and keep reinvesting your BTC's in growing your farm to keep up with the hash rate. Eventually things will level out, it's in no ones interest to keep driving it up everyone looses out it the long run due to the increased cost of doing business (power and space) which is universal.

Hey Lucas

I have a hunch that asics were developed with 51% attack mentality in mind. The easiest way to take BTC down would be to centralize it's hashpower.

With so many of these in the hands of the relatively few, the last thing we need right now is another alt coin diverting more gpus.

Yes, the price dip could be temporary. But the difficulty increase isn't, and will only be going up and up and up....Friedcat will continue adding capacity to ASICMiner, and Butterfly Labs sounds like they are actually working through thei backlogs...  Put those miners in peoples hands and lets see where difficulty is in August or September..

For people with big GPU investments (or any GPU investments), the short term course of action would seem like they should keep going for the time being, but when the difficulty flies up even more, they might be forced to take their machines offline, at least until the prices catches up.

Me, I'm crossing my fingers that BFL gets my units tot me in short order. But i simply can't imagine investing in a new Radeon card at this point in time. At least not for mining Bitcoins. The game definetly looks like it's changing.

Another thought I had is, lots of farms take themselves offline, ASICminer might find itself unintentionally in control o 51% of the network. What happens then?

Why are you guys letting a small dip in the price and diff jump scare you ??

Its not hard to imagine that there are many who regret selling off or shutting down their farms during the latter part of 2012.

I plan to expand as soon as the 78/79xx prices drop a little more..

There's no reason why they wouldn't switch to litecoin or some other scrypt coin. Many already have, and as it is currently more profitable even without too many ASICs, I think that's what will happen sooner rather than later.

With this last difficulty increase, I don't see GPU mining as being profitable any longer if you are looking to invest in PC hardware. Will probably be phasing, parting-out, and selling my GPU rigs by Christmas.

Disclosure: Was mining at a loss for a while over the last two years and ended up paying for all my GPU purchases + current ASIC hardware purchases, so YMMV.

I'm keeping 2 of my gpu's to put in a gaming pc for my kids, the other 7 gpu's I'm selling locally unless the buyer backs out

I think I will finally finish Unreal.

With the jump in difficulty and the recent slump in Bitcoin prices, GPU mining seems like it's just about over with. Wondering what everyone's going to do? Sell their GPU's? Mine an alt coin? Which one? Seems any SHA coin can be overtaken by ASIC's, so will everyone be moving over to Litecoin?

Serious question.

I have a jalapeno pre-ordered since last early fall, and some friends were recently saying we should just go and buy a few Radeons while we waited, but I'm now quite happy that we didn't, especially now that Butterfly apparently is making big progress processing their order queue.