Coremesh
분산형 GPU 컴퓨팅 마켓플레이스
공연
RD 인기지표
RD 성장지표
비교
Coremesh 소개
Coremesh is a middleware layer that connects idle GPU capacity to a network: a lightweight client on each device takes jobs during available hours, covering AI inference, zero-knowledge proof generation and validator hosting, while a scheduler matches jobs by memory, bandwidth and latency and settles in USDC or the platform token. Jobs ship as container images with runtime and dependencies pinned inside the image.
It addresses the long tail on both sides, where small server rooms and personal cards produce idle capacity between tasks and inference or proving jobs have uneven per-machine requirements. The protocol isolates jobs in containers, constrains node behavior with staking and ties payout to delivery quality, shifting compute from per-machine rental to per-task pricing. Nodes submit periodic liveness proofs and lose job priority after repeated absence, and cluster mode groups machines into one scheduling unit.