freicoin / freicoin-old

Bitcoin integration/staging tree
http://www.bitcoin.org
MIT License
26 stars 8 forks source link

Decide on (and implement) hashing algorithm (GPU vs CPU) #66

Closed ImpalerWrG closed 12 years ago

ImpalerWrG commented 12 years ago

Litecoin successfully uses a CPU intensive algorithm called Scrypt, this might be a preferable to the standard SHA as it would make the block-chain harder to attack with the large GPU mining rigs that exist. Their may also be considerable upsides in energy consumption, hardware costs and a more evenly distributing mining in the population which may be beneficial for public acceptance.

maaku commented 12 years ago

Litecoin is an altchain created by miners who felt disenfranchised in the transition from CPU to GPU mining. It was a reactionary movement to the industrialization of mining: no longer could your average person participate and turn a profit; mining requires capital investment and technical know-how. That's certainly more true of FPGA mining, and will be true of the transition to ASICs as well.

I understand the desire for mining to be a participatory endeavor. However from a core developer's point of view the industrialization of mining is A Very Good Thing. It raises significantly the capital costs involved in mounting an attack, eliminates the advantage enjoyed by botnet operators, and centralizes security of the network into a federation of professional operators that you can work with. There really isn't a downside, from the network-operators' point of view.

It should be mentioned too that the consumer-friendly, GPU-hostile property of scrypt and related functions is a myth. Worse, it appears to have been a scam: it's been shown somewhat persuasively that one of the originators of the concept was GPU-mining from the start. Regardless, a public GPU miner for Litecoin now exists and it is now longer economical for the average person to CPU-mine. If litecoin has any lasting power, there will eventually be a run of ASIC mining rigs as well.

Finally, the most significant drawback to changing the proof-of-work function is that it eliminates our current path towards ultimate, long-term security: merged-mining with the support of bitcoin's large pool operators. While changing the hash function may or may not gain us security in the short-term, it definitely comes with an opportunity cost in our long-term security plan.

maaku commented 12 years ago

Sufficient time was allowed for dissenting comments; now considering this closed.