From d81741514bf2fd9fee26bc43ba5833d887494ca8 Mon Sep 17 00:00:00 2001
From: Ludovic Raess <ludovic.rass@gmail.com>
Date: Tue, 1 Oct 2024 18:55:40 +0200
Subject: [PATCH] Update doc

---
 docs/src/concepts/architectures.md | 8 +++++++-
 docs/src/concepts/grids.md         | 3 +++
 docs/src/getting_started.md        | 6 +++++-
 docs/src/index.md                  | 1 +
 4 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/docs/src/concepts/architectures.md b/docs/src/concepts/architectures.md
index 3a673841..c9a8a2f5 100644
--- a/docs/src/concepts/architectures.md
+++ b/docs/src/concepts/architectures.md
@@ -2,7 +2,7 @@
 
 ## Backend Selection & Architecture Initialization
 
-Chmy.jl supports CPUs, as well as CUDA and ROC backends for Nvidia and AMD GPUs through a thin wrapper around the [`KernelAbstractions.jl`](https://github.com/JuliaGPU/KernelAbstractions.jl) for users to select desirable backends.
+Chmy.jl supports CPUs, as well as CUDA, ROC and Metal backends for Nvidia, AMD and Apple M-serie GPUs through a thin wrapper around the [`KernelAbstractions.jl`](https://github.com/JuliaGPU/KernelAbstractions.jl) for users to select desirable backends.
 
 ```julia
 # Default with CPU
@@ -21,6 +21,12 @@ using AMDGPU
 arch = Arch(ROCBackend())
 ```
 
+```julia
+using Metal
+
+arch = Arch(MetalBackend())
+```
+
 At the beginning of program, one may specify the backend and initialize the architecture they desire to use. The initialized `arch` variable will be required explicitly at creation of some objects such as grids and kernel launchers.
 
 ## Specifying the device ID and stream priority
diff --git a/docs/src/concepts/grids.md b/docs/src/concepts/grids.md
index 3cf07b29..ec7c50f7 100644
--- a/docs/src/concepts/grids.md
+++ b/docs/src/concepts/grids.md
@@ -48,6 +48,9 @@ grid   = UniformGrid(arch;
                     dims=(nx, ny, nz))
 ```
 
+!!! warning "Metal backend"
+    If using the Metal backend, ensure to initialise the grid using `Float32` (`f0`) values in the `origin` and `extent` tuples.
+
 !!! info "Interactive Grid Visualization"
     - [grids_2d.jl](https://github.com/PTsolvers/Chmy.jl/blob/main/examples/grids_2d.jl):  Visualization of a 2D `StructuredGrid`
     -  [grids_3d.jl](https://github.com/PTsolvers/Chmy.jl/blob/main/examples/grids_3d.jl):  Visualization of a 3D `StructuredGrid`
diff --git a/docs/src/getting_started.md b/docs/src/getting_started.md
index 4dc281df..b087995c 100644
--- a/docs/src/getting_started.md
+++ b/docs/src/getting_started.md
@@ -47,6 +47,7 @@ using KernelAbstractions # for backend-agnostic kernels
 using Printf, CairoMakie # for I/O and plotting
 # using CUDA
 # using AMDGPU
+# using Metal
 ```
 
 In this introductory tutorial, we will use the CPU backend for simplicity:
@@ -56,7 +57,10 @@ backend = CPU()
 arch = Arch(backend)
 ```
 
-If a different backend is desired, one needs to load the relevant package accordingly. For example, if Nvidia or AMD GPUs are available, one can comment out `using CUDA` or `using AMDGPU` and make sure to use `arch = Arch(CUDABackend())` or `arch = Arch(ROCBackend())`, respectively, when selecting the architecture. For further information about executing on a single-device or multi-device architecture, see the documentation section for [Architectures](./concepts/architectures.md)
+If a different backend is desired, one needs to load the relevant package accordingly. For example, if Nvidia or AMD GPUs are available, one can comment out `using CUDA`, `using AMDGPU` or `using Metal` and make sure to use `arch = Arch(CUDABackend())`, `arch = Arch(ROCBackend())` or `arch = Arch(MetalBackend())`, respectively, when selecting the architecture. For further information about executing on a single-device or multi-device architecture, see the documentation section for [Architectures](./concepts/architectures.md)
+
+!!! warning "Metal backend"
+    Metal backend restricts floating point arithmetic precision of computations to `Float32` or lower. In Chmy, this can be achieved by initialising the grid object using `Float32` (`f0`) elements in the `origin` and `extent` tuples.
 
 ## Writing & Launch Compute Kernels
 
diff --git a/docs/src/index.md b/docs/src/index.md
index 087b885d..b5ee7a90 100644
--- a/docs/src/index.md
+++ b/docs/src/index.md
@@ -31,6 +31,7 @@ Chmy.jl provides a comprehensive framework for handling complex computational ta
 
 A general list of the features is:
 
+- Backend-agnostic implementation relying on [KernelAbstractions.jl](https://github.com/JuliaGPU/KernelAbstractions.jl)
 - Distributed computing support with [MPI.jl](https://github.com/JuliaParallel/MPI.jl)
 - Multi-dimensional, parameterizable discrete and continuous fields on structured grids
 - High-level interface for specifying boundary conditions with automatic batching for performance