What is Autolykos?

Autolykos is a memory-hard Proof of Work algorithm developed by the Ergo Platform team, currently in its second version (Autolykos v2). The algorithm requires miners to solve a puzzle that involves generating a large table of hash values in memory and then finding combinations of these values that satisfy a difficulty target. In Autolykos v2, miners must maintain approximately 2-4 GB of data in fast memory (GPU VRAM) and perform repeated lookups and computations against this dataset. The algorithm uses a combination of Blake2b hashing and modular arithmetic over large numbers, with the puzzle requiring miners to find a set of indices whose corresponding hash values sum to a specific target.

The primary design goal of Autolykos is ASIC resistance through memory hardness, ensuring that mining remains accessible to consumer-grade GPUs. The large memory requirement creates a bottleneck that makes ASIC development economically impractical, as the cost of integrating sufficient high-speed memory into custom chips would negate any computational advantage. Autolykos v2, which replaced the original Autolykos v1 in a hard fork, removed the pool-resistance feature of v1 (which required miners to store private keys) and enabled traditional mining pool support, making it more accessible to smaller miners. The algorithm also features an adjustable memory table size that grows with the blockchain, helping to maintain ASIC resistance over time as hardware evolves.

Autolykos is used exclusively by Ergo, a smart contract platform launched in 2019 that focuses on providing efficient and secure financial contracts. Ergo was founded by Alexander Chepurnoy (known as kushti), a former core developer of Nxt and co-founder of smartcontract.com (Chainlink). The transition from Autolykos v1 to v2 occurred in 2021 and was a significant event that opened up Ergo mining to a much broader community by enabling pool mining. Following Ethereum's transition to Proof of Stake in September 2022, Ergo became one of the most popular destinations for displaced GPU miners, experiencing a significant increase in hash rate as miners redirected their hardware to the network.