FutureBit tests Bitcoin mining with a digital fly brain

FutureBit tests Bitcoin mining with a digital fly brain
FutureBit tests Bitcoin mining with a digital fly brain

​Mining hardware manufacturer FutureBit has tested an unusual method of performing SHA-256 computations using a digital model of a fruit fly brain. The HashFly experiment uses 2,914 simulated neural connections, but currently runs on conventional computer hardware and cannot compete with ASIC miners.

The company says a theoretical system based on real biological neurons could achieve energy efficiency of around 1 watt per TH/s. That would be roughly 10 times more efficient than modern ASIC chips built on a 3nm process, although such a system remains purely hypothetical.

HashFly is based on MaleCNS v1.0, a digital map of the brain and central nervous system of an adult male fruit fly. FutureBit plans to expand the experiment to all neurons in the dataset and publish its research findings.

Digital neurons perform SHA-256 computations

To mine Bitcoin, machines repeatedly perform SHA-256 computations in an attempt to find a value that meets the network's required difficulty. Mining hardware performance is measured by hashrate, or the number of such attempts it can make per second.

In HashFly, part of the computational process is performed using 2,914 simulated neural connections. Users can adjust the target difficulty, but the experiment operates at a level that is not comparable with the real Bitcoin network.

For comparison, FutureBit's compact Apollo III ASIC miner can deliver around 18 TH/s. HashFly runs directly in a browser on conventional hardware and is not designed for practical Bitcoin mining.

The most intriguing part of the experiment is FutureBit's theoretical estimate for a biological version of the system.

The company calculated that if real neurons could be made to continuously perform mining-related computations and their output were compared with the energy consumption of a fruit fly, efficiency could reach around 1 watt per TH/s. No real-world system with such characteristics currently exists.

Fly brains are also being used to control mining

A separate project called FlyMiner is being developed using a digital map containing 139,255 fruit fly neurons and 16.8 million connections between them. Unlike HashFly, the system does not directly perform all mining computations. Instead, it determines when a conventional algorithm should begin searching for a solution.

When activity among simulated neurons associated with the fly's movement reaches a predefined threshold, the program begins testing solutions to a task received from the CKPool solo mining pool. The process reaches around 700,000 attempts per second.

Digital models of fruit fly brains have also been experimentally used not only for mining but for trading. Such projects test whether an organism's neural activity, without any understanding of financial markets, can be used as a source of signals for trading decisions.

One potential difference between this approach and human behavior is the absence of emotional reactions to market movements. However, such experiments have yet to demonstrate that biological or simulated neural systems outperform professional trading algorithms.

Earlier, we reported that HOLY Mining launched an intelligent cloud computing platform to improve mining efficiency. The company is focusing on automating the management of computing power and infrastructure.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.