Bit-Brick Cluster K1 — 4-Slot SpacemiT K1 RISC-V Cluster Board with GbE Switch

Specifications

Overview

The Bit-Brick Cluster K1 is a cluster carrier board that accepts up to four SSOM-K1 system-on-modules, each built around an eight-core SpacemiT K1 RISC-V processor. It is aimed at developers, researchers and system integrators working on distributed edge computing, AI workloads and high-performance embedded experiments — the RISC-V analogue of the Raspberry Pi cluster boards that have been standard in lab and teaching environments.

Each SSOM-K1 node carries an 8-core X60 RISC-V CPU at 1.6 GHz with the RV64GCVB instruction set and the RVA22 RVV 1.0 vector extension, an Imagination IMG BXE-2-32 GPU at 819 MHz rated at 20 GFLOPS, and 2.0 TOPS of INT8 AI throughput via CPU core fusion. Nodes are individually configured with 4 GB or 8 GB of RAM and 16, 32 or 64 GB of eMMC, so a cluster can be assembled with uneven node capacity for experiments that need it.

What separates the Cluster K1 from a stack of single-board computers is the interconnect: an onboard Gigabit Ethernet switch chip links the four core boards, giving every node its own independent network port with automatic IP assignment and no external switch, cabling mess or manual addressing. Each module also exposes dual Gigabit PHY, HDMI 1.4 up to 1080p60 on the master slot, two 4-lane MIPI CSI camera inputs, and USB 3.0 plus USB 2.0. Per-node video handling covers 4K60 H.265/H.264/VP9/VP8 decode and 4K30 encode.

Key Benefits

Onboard Gigabit switch with a port per node and automatic IP assignment removes external switching and manual network setup from cluster builds. Up to four 8-core RISC-V nodes with RVV 1.0 vector extension give 32 RISC-V cores in one chassis. 2.0 TOPS INT8 per node supports distributed inference experiments, and per-node 4K60 decode plus dual MIPI CSI make it usable for multi-camera edge workloads. Modules are individually specified, so node capacity can be mixed.

Applications

Distributed edge computing and cluster research, RISC-V software development and CI farms, teaching and university lab clusters, multi-camera edge vision experiments, distributed AI inference proof-of-concept work.

Related Products