So once people get their hands on ASICs, is mining on GPUs dead?

[XX55XX]

The increased difficulty is making me reconsider whether or not I even want to continue mining or not. Especially with people getting their hands on Avalon ASICs...

[miningdude]

Just switch to Litecoin mining, feathercoin mining, or whatever the latest speculation coin is for quick, but short-lived profits. 

Oh, and make sure you have a way of exchanging whatever you mine into Bitcoin or Litecoin, since they're the most stable currencies at the moment. Some would argue for USD, but I think Bitcoin has more of a chance at appreciating in value than the good 'ol greenbacks.

[GodHatesFigs]

for anyone paying for electricity, yes

[jhansen858]

It didn't die when FPGA's came out?

[Amph]

my suggestion is to mine until the end of may, then switch to litecoin while you wait to buy one of the asic thing that will come in july

my 0.02btc

[jundesirehd]

I think I'll switch to ltc mining then trade ltc for btc

[Frizz23]

...  one of the asic thing that will come in july

July which year? 2014 or 2015?

[computerparts]

...  one of the asic thing that will come in july

July which year? 2014 or 2015?

I think it was 20neverteen

[Teka]

I don't think that GPUs will die as soon as ASICs are out simply because you can always mine other crypto currencies such as Litecoin. Litecoin mining at the moment is more profitable than Bitcoin mining. Yes the difficulty of Litecoin will go up but still due to move of Bitcoin miners to Litecoin the trade of Litecoin will increase and therefore the price might increase.

[Bonam]

It's still a while til GPU mining will be dead. All the ASIC introduction is slower than expected, and the ASICs you can actually buy (ASICminer) aren't any better in terms of MHash/$ than GPUs. And electricity costs are basically negligible right now (I mine ~$200/month and spend $25/month in electricity with GPUs). If BTC price stays the same, you need about a 10x increase in difficulty to make GPU mining no longer profitable. It will take a while yet for 1000 more TH of ASICs to come online. And by then we'll also have newer more efficient GPUs too. By year end, we'll likely have 9970s mining ~1.1 GH for a price of ~$500 USD, a way better ratio than any currently available ASIC, and only a factor of ~7 worse than BFL's promised products (I bet once they are actually shipping for real they will raise prices making it even closer, given what people have been willing to pay for ASICMiner products).

Also, unlike GPUs, the long term reliability and longevity of ASIC products is yet to be demonstrated. I design PCBs and electronic products in my field of work and there's a big difference between something that works on a bench and something that is proven to work continuously for years on end. ASIC products are all basically first run prototypes, none of them having been tested for years of operation. It is likely they will encounter longevity issues. GPUs can also be resold for 50-90% of their original purchase cost to buyers outside the BTC community, whereas obsolete ASICs will have no resale value.

[bitman442]

The profitability of GPU mining isn't driven solely by difficulty, rallies and corrections in coin prices are the other factor. Are you really going to care if your card is only making .25 btc per month if they are worth $1000 a piece? Not saying that's what will happen, a little bit of it comes down to your faith in bitcoin itself.

[ThatDGuy]

The profitability of GPU mining isn't driven solely by difficulty, rallies and corrections in coin prices are the other factor. Are you really going to care if your card is only making .25 btc per month if they are worth $1000 a piece? Not saying that's what will happen, a little bit of it comes down to your faith in bitcoin itself.

This right here.  People a year or two ago probably weren't thrilled that their cards were only producing 1 BTC a week.

[Bitsaurus]

The amount of math and logic fail in this thread is extraordinary.  I have access to "free" electricity and even I will acknowledge at some point it makes sense to sell the cards and put that towards ASIC or buying BTC outright.

9990 for $500 by year end mining 1.1GH/s - what are you people smoking?  It took a few months just to get GCN 7970s up to 650+.  I'll make a 100BTC bet right now with escrow that no retail GPU will mine 1.1GH/s for $500 by end of 2013.

I have been mining for well over 2 years now and when the price dropped down to $2 I did in fact stop mining.  But I didn't give up on Bitcoin, instead I bought coins.  There's no point mining and spending more money to earn a coin than it costs to outright buy it.  At no point does it ever make sense to spend more on electricity mining than it does to buy coins.

Once the cards are making minimal BTC there will be a rush to sell GPUs (probably to other miners who will be mining alt coins).  The market prices will drop.  The only people who will not be in a rush will be those with free electricity 

[Dyaheon]

It didn't die when FPGA's came out?

That's cause FPGAs weren't really better. Their only point was lower power consumption, which was a big deal when BTC value was at the bottom and profit margins were small. Now that they've improved, FPGAs aren't that lucrative anymore. They were always more or as expensive as GPUs.

ASICs, on the other hand, improve over GPUs on all fronts. You get far more GH/s per BTC than you do with any other option, and power usage is greatly improved again. So yes, ASICs will kill bitcoin-mining GPUs definitely, but non-SHA256 cryptocurrencies will likely remain an option.

[ThatDGuy]

It didn't die when FPGA's came out?

That's cause FPGAs weren't really better. Their only point was lower power consumption, which was a big deal when BTC value was at the bottom and profit margins were small. Now that they've improved, FPGAs aren't that lucrative anymore. They were always more or as expensive as GPUs.

ASICs, on the other hand, improve over GPUs on all fronts. You get far more GH/s per BTC than you do with any other option, and power usage is greatly improved again. So yes, ASICs will kill bitcoin-mining GPUs definitely, but non-SHA256 cryptocurrencies will likely remain an option.

I agree with almost everything you said about ASICs improvements but think they differ from GPU and CPU mining in one aspect: being application specific.

Up until very recently, Bitcoin mining has been driven at a basic level by the ability to get meaningful results on equipment that people already had, or could easily purchase.  They could also be used for either computing or graphics and could also be resold for those purposes, typically at a decent resale value.

ASICs by nature of their specificity, completely change that part of the equation.  As you mentioned, non-SHA256 cryptos will still be an option, and I think it will be fascinating to see how that difference plays into the relationship between Bitcoin and the other cryptocurrencies as ASICs are introduced.

[Malawi]

It didn't die when FPGA's came out?

FPGA's did not offer the same increase in hashing power per BTC spent. They were mainly about lowering electricity use.

[glendall]

Ya I really don't see how GPU mining will be profitable.

Big fallacy #1 =  GPU  mining is fine because you can mine LTC when the ASICS hit.

This isn't logical, because naturally all the GPU miners will be doing the same thing, thus LTC difficulty will rise a massive amount in much the same way as BTC mining.


Big fallact #2 = the 9970 will save us

I don't believe this. It's possible, but I don't think likely.

An ASIC is a specialized product for bitcoin mining. It is just naturally an exponential gain in power/performance efficiency . The next gen AMD stuff will gain the likely the same increase as it did from the 6000 series to the 7000 series. If it's more than that , there will be little chance of it matching the 2nd gen ASICs that'll be coming out before not too long (likely before the 9970 itself comes out) .