5 ms·

Comparing the two is completely meaningless. Ethereum's hash function (Ethash) is orders of magnitude more expensive than Bitcoin's SHA256. And also, it is ASIC

Comparing the two is completely meaningless. Ethereum's hash function (Ethash) is orders of magnitude more expensive than Bitcoin's SHA256. And also, it is ASIC-resistant.

aserafini 11y ago
Good point, a better way to compare would be the Watts consumed by each network: that gives you the minimum energy costs required for a 51% attack. Unfortunately I can't seem to find any typical "hashes per joule" numbers for Ethereum mining rigs that would allow us to extrapolate this. I still speculate it would be orders of magnitude more expensive to launch such an attack on the Bitcoin network.
- tromp 11y ago
  The mining reward itself is a good estimator of the energy cost (an over-estimate in fact since the reward has to cover both energy costs as well as hardware investment/depreciation, but the former dominates the latter). So attacking bitcoin costs roughly 150BTC per hour, while attacking Ethereum costs roughly 1200ETH (worth about 30BTC currently) per hour. This suggests Bitcoin is 5 times more "secure".
  - aserafini 11y ago
    Mining reward and exchange rate gives us a strict upper bound on the cost of energy it's economical to spend on mining but I'm not sure it provides a good estimate of the actual energy being used. If energy_used = mining_reward * exchange_rate, an increase in hash power would be correlated positively with exchange rate. In reality, Bitcoin hashrate has increased over 30x in the last 2 years and the exchange rate in USD has actually decreased in the same period. That suggests we aren't using anywhere near the upper bound of energy that remains economical. Given that Bitcoin network performs 1 million x as many hashes per second than Ethereum, the energy cost per Ethash must be somewhere like 200,000x more expensive than SHA256 to compute if we are to assume Bitcoin is only 5x as 'secure'. I have't fired up an Ether mining rig to test whether that is the case or not, but it seems highly unlikely so I'm still maintaining a 51% attack against Bitcoin is orders of magnitude more costly to perform than the equivalent attack on Ether. Realistic 'energy use per hash' for SHA256 versus Ethash would be good numbers to have that either refure or back that up. UPDATE: found some numbers from mining rig data. We can use 'hashes per joule' numbers for the most efficient Bitcoin mining rig and Ethereum mining rig. Top Bitcoin mining rig performs 4000 Mhash/J according to [1] Top Ether mining rig performs 0.08 Mhash/J according to [2] So if every miner achieves those numbers, the Bitcoin network is currently running around 300,000,000W versus 20,000,000W on Ethereum network (15x). So, perhaps not orders of magnitude more secure, but probably at least one! (for the time being). [1] https://en.bitcoin.it/wiki/Mining_hardware_comparison https://en.bitcoin.it/wiki/Mining_hardware_comparison [2] http://www.vmxlabs.com/ http://www.vmxlabs.com/

Comparing the two is completely meaningless. Ethereum's hash function (Ethash) is orders of magnitude more expensive than Bitcoin's SHA256. And also, it is ASIC-resistant.

aserafini 11y ago

Good point, a better way to compare would be the Watts consumed by each network: that gives you the minimum energy costs required for a 51% attack. Unfortunately I can't seem to find any typical "hashes per joule" numbers for Ethereum mining rigs that would allow us to extrapolate this. I still speculate it would be orders of magnitude more expensive to launch such an attack on the Bitcoin network.

tromp 11y ago
The mining reward itself is a good estimator of the energy cost (an over-estimate in fact since the reward has to cover both energy costs as well as hardware investment/depreciation, but the former dominates the latter). So attacking bitcoin costs roughly 150BTC per hour, while attacking Ethereum costs roughly 1200ETH (worth about 30BTC currently) per hour. This suggests Bitcoin is 5 times more "secure".
- aserafini 11y ago
  Mining reward and exchange rate gives us a strict upper bound on the cost of energy it's economical to spend on mining but I'm not sure it provides a good estimate of the actual energy being used. If energy_used = mining_reward * exchange_rate, an increase in hash power would be correlated positively with exchange rate. In reality, Bitcoin hashrate has increased over 30x in the last 2 years and the exchange rate in USD has actually decreased in the same period. That suggests we aren't using anywhere near the upper bound of energy that remains economical. Given that Bitcoin network performs 1 million x as many hashes per second than Ethereum, the energy cost per Ethash must be somewhere like 200,000x more expensive than SHA256 to compute if we are to assume Bitcoin is only 5x as 'secure'. I have't fired up an Ether mining rig to test whether that is the case or not, but it seems highly unlikely so I'm still maintaining a 51% attack against Bitcoin is orders of magnitude more costly to perform than the equivalent attack on Ether. Realistic 'energy use per hash' for SHA256 versus Ethash would be good numbers to have that either refure or back that up. UPDATE: found some numbers from mining rig data. We can use 'hashes per joule' numbers for the most efficient Bitcoin mining rig and Ethereum mining rig. Top Bitcoin mining rig performs 4000 Mhash/J according to [1] Top Ether mining rig performs 0.08 Mhash/J according to [2] So if every miner achieves those numbers, the Bitcoin network is currently running around 300,000,000W versus 20,000,000W on Ethereum network (15x). So, perhaps not orders of magnitude more secure, but probably at least one! (for the time being). [1] https://en.bitcoin.it/wiki/Mining_hardware_comparison https://en.bitcoin.it/wiki/Mining_hardware_comparison [2] http://www.vmxlabs.com/ http://www.vmxlabs.com/