141,781 measured positions
MaleCNS v1.0 soma coordinates form the FLYGPU compute core. The source geometry is centered, scaled and preserved as a data-driven neural structure.
Explore the source data ↗Biology becomes
board architecture.
FLYGPU translates the spatial organization of a fruit-fly nervous system into a new visual language for compute: dense local structure, short signal paths and many small work units moving in parallel.
The measured neural cloud sits inside a modeled GPU package. CPU hash activity drives its cooling system, circuit traces and light flow in real time.
Double SHA-256 in-browser
POWER ON processes server-issued 80-byte challenges inside a duty-limited Web Worker. The live hash rate controls fan RPM, trace velocity and board illumination.
Compute you can inspect
A modeled PCB, neural package, dual cooling fans, power stages and signal traces turn each work cycle into a readable hardware scene.
From fly-brain geometry
to distributed compute.
Neural geometry × local hashing
Measured anatomy, controllable CPU work, verified challenges and a responsive 3D board establish the first FLYGPU core.
Partitioned work channels
Expand task routing, contributor discovery, result validation and fault-tolerant coordination across many lightweight workers.
A verifiable compute layer
Develop open scheduling and contribution measurement with performance claims grounded in reproducible benchmarks.
FLYGPU is an independent research concept. It is not affiliated with OKX, BNB Chain, MaleCNS collaborators or hardware manufacturers. Wallet connection does not start compute or grant asset permissions.