A Shader for Nano-FPGA Fabric
May 1, 2022

A programmable fabric becomes easier to invent when its structure becomes visible. Brian Greenforest’s GLSL shader renders a nano-FPGA surface so designers can inspect tiling, cell roles, local connections, and the consequences of granularity.

The browser-based rendering also published the switching-fabric concept in a searchable medium and opened it to teaching, discussion, and continued development.

Render the Architecture as a Living Surface

Each pixel can represent local state or a cell function, while the shader evaluates many cells in parallel. That correspondence makes WebGL a natural medium for exploring arrays whose physical implementation also relies on massive local parallelism.

The early design used a single reconfigurable MUX as its finest switching element. Later work increased node capability so each cell could compute and manage adjacent configuration without wasting control circuitry.

Use Visualization to Accelerate Hardware Design

A visible fabric helps engineers compare granularity, routing, state, and communication before committing to a board or chip. It also gives students an immediate way to connect shaders with reconfigurable logic.

Explore the rendered fabric, inspect the newer Cartilage implementations, or collaborate on the next visualization and hardware mapping.

Execute A Shader for Nano-FPGA Fabric Through Cartilage Runtime Instantiation

Cartilage Runtime-Instantiation Architecture carries this system into nested modules that enter and connect inside a live spatial machine.

Cartilage Runtime-Instantiation Architecture · Cartilage Core · public source repository · https://youtube.com/watch?v=irna23iEQBA&feature=share

Originally posted on LinkedIn

Brian Greenforest · (2022-05-01 17:28:42 UTC)

Open the original LinkedIn post · LinkedIn activity 6926583190195253248

LinkedIn status when archived: Visible to Brian’s connections.

7 months ago I wrote a GLSL shader for nano-FPGA fabric. I think prior art law holds, because ShaderToy.com is a findable on Google Search web site. Don't know where patent attorney would look: people invent so much, that "who cares"? #opensource #openhardware PM for details and to schedule free learning lessons. Next version is also coming and in works right now.

Original LinkedIn media

Media source note: Current thumbnail from the video URL recorded in Brian Greenforest’s LinkedIn export.