Quote from: BubbleBoy on June 19, 2011, 08:17:25 AM

GPU miners have 100:1 gain over CPU, and ASICs have 100:1 over GPU. You need 10.000 CPUs to catch up to a single $50 ASIC, so it follows energy prices are not relevant, it's mostly best tech wins. Building a 100.000 or 1000.000 computer botnet is not free, and has a black market value higher than the equivalent ASIC mining. I predict ASICs will arrive in ~ 3 months if the price stays stable.

I read about the LargeCoin ASICs rumors yesterday for the first time. Before that I only heard about experiments with FGPA chips. Indeed if it worked - ASICS would become the primary (most cost efficient and thereby most effective) way of mining - mostly because they are more energy efficient compared to GPUs in Watts per GigaHash. Therefore I would not agree that 'energy prices are irrelevant'. Rule of thumb should be: More efficient machines will crowd out less efficient ones. As soon as a miner figures out that running costs are higher than the market price for him he has to stop mining immediately - even when he believes that the price for bitcoins might rise dramatically - buying at an exchange from somebody who is able to mine at lower costs would be more effective. If holding/investing in BTCs is not an option for such a sub-threshold-miner he has to invest in more efficient hardware or simply will go away...

Very efficient specialized hardware (compared off the shefs products available to the masses) in the hands of a single entity could have a significant impact on the bitcoin ecosystem. If nobody can catch up this entity could reinvest all generated coins into hardware until it would possess >50% of the network's computational power. This would then allow to dictate transaction fees by not accepting blochs of other miners because eventually the longest blockchain would be generated by machines of that entity itself.

If there will still be competition between a high enough number (>oligopoly) energyefficiency will be their primary area of optimization.

Regarding my original post I agree that it might not be possible to run a ASICs super-computer with significant power consumptions hidden in the corner of a youth center. Parasitic mining would make senes only with the latest generation of minig hardware (a factor of up to 1/10 might be ok - but if others have 100 times more effective machines it will not be worth making a profit of 1% of the energy costs paid by your boss).

blueling

Posts

Merits