Installation¶

You can try Henad with its example models in the web app, with nothing to install. To run it on your machine, install the app and the command line from crates.io, or build them from a clone of the repository. To write your own models, start a project from the template instead.

Henad runs on any device with a CPU, optionally a GPU, and any operating system that can build wgpu.

Supported platforms

As per the wgpu documentation, Henad can be run on the following platforms as of wgpu v30:

API	Windows	Linux/Android	Web (wasm)
Vulkan
Metal
DX12
OpenGL	, or	(GL ES 3.0+)	(WebGL2)
WebGPU

= First Class Support
= Downlevel/Best Effort Support
= Requires the ANGLE translation layer (GL ES 3.0 only). On macOS/iOS, use the angle feature. On Windows, gles uses WGL by default. Build with cfg(windows_angle) to use ANGLE instead.
= Requires the MoltenVK translation layer

Installing the app and the command line¶

With a Rust toolchain from rustup, install both from crates.io:

cargo install --locked --config profile.release.opt-level=2 henad-app henad-cli

This puts henad-app and henad-cli on your path, with the ten example models. --locked builds with the dependency versions the release was tested with. The --config builds at opt-level 2, the level every Henad number is measured at. Without it cargo install builds at 3.

Building from a clone¶

git clone https://github.com/micfong-z/henad.git
cd henad

On the first build, rustup installs the toolchain that the repository pins. cargo run --release --bin henad-app and cargo run --release -p henad-cli then run the two binaries, as Running Henad shows.

The web build needs the dated nightly toolchain the repository pins, with the rust-src component for wasm threads. Run this from the repository root:

rustup toolchain install "$(cat templates/model-project/scripts/web-toolchain)" --profile minimal \
  --component rust-src,clippy --target wasm32-unknown-unknown

It also needs Trunk, at the release the repository pins:

cargo install --locked trunk --version "$(cat templates/model-project/scripts/trunk-version)"

Use scripts/build_web.sh to build the web app.

Starting a project¶

Your own models live in your own project, with its own app, command line, web build, tests and CI, on the published crates. Your own project fetches the template and sets it up, and the first-model tutorials add models to it.

A Rust program that runs the example models, sweeps them or opens the app needs no template, and Using Henad from code shows one.

Check the install¶

henad-cli --info

From a clone, run cargo run -p henad-cli -- --info instead. This prints the host details and the GPU adapter wgpu selected. If an adapter line is printed, the four GPU models are available. Otherwise, only the CPU models are available.

Packages on Linux¶

A desktop distribution ships everything Henad needs. A server or container image might lack a Vulkan loader and driver, and henad-cli --info then prints no adapter. On Debian and Ubuntu, install libvulkan1 with mesa-vulkan-drivers or your GPU vendor's driver. On Fedora, install vulkan-loader with mesa-vulkan-drivers. On X11 the app needs libxkbcommon-x11, packaged as libxkbcommon-x11-0 on Debian and Ubuntu. Its file dialogs need xdg-desktop-portal with a backend for your desktop, such as xdg-desktop-portal-gtk.