Metal.jl

Metal programming in Julia

With Metal.jl it's possible to program GPUs on macOS using the Metal programming framework.

The package is a work-in-progress. There are bugs, functionality is missing, and performance hasn't been optimized. Expect to have to make changes to this package if you want to use it. PRs are very welcome!

Requirements

Mac device with M-series chip
Julia 1.8-1.10
macOS 13 (Ventura) or 14 (Sonoma)

These requirements are fairly strict, and are due to our limited development resources (manpower, hardware). Technically, they can be relaxed. If you are interested in contributing to this, see this issue for more details. In practice, Metal.jl will probably work on any macOS 10.15+, and other GPUs that are supported by Metal might also function (if only partially), but such combinations are unsupported for now.

Quick start

Metal.jl can be installed with the Julia package manager. From the Julia REPL, type ] to enter the Pkg REPL mode and run:

pkg> add Metal

Or, equivalently, via the Pkg API:

julia> import Pkg; Pkg.add("Metal")

For an overview of the toolchain in use, you can run the following command after importing the package:

julia> using Metal

julia> Metal.versioninfo()
macOS 13.5.0, Darwin 22.6.0

Toolchain:
- Julia: 1.9.3
- LLVM: 14.0.6

Julia packages:
- Metal.jl: 0.5.0
- Metal_LLVM_Tools_jll: 0.5.1+0

1 device:
- Apple M2 Max (64.000 KiB allocated)

Array abstraction

The easiest way to work with Metal.jl, is by using its array abstraction. The MtlArray type is both meant to be a convenient container for device memory, as well as provide a data-parallel abstraction for using the GPU without writing your own kernels:

julia> a = MtlArray([1])
1-element MtlArray{Int64, 1}:
 1

julia> a .+ 1
1-element MtlArray{Int64, 1}:
 2

Kernel programming

The above array abstractions are all implemented using Metal kernels written in Julia. These kernels follow a similar programming style to Julia's other GPU back-ends, and with that deviate from how kernels are implemented in Metal C (i.e., indexing intrinsics are functions not arguments, arbitrary aggregate arguments are supported, etc):

julia> function vadd(a, b, c)
           i = thread_position_in_grid_1d()
           c[i] = a[i] + b[i]
           return
       end
vadd (generic function with 1 method)

julia> a = MtlArray([1,1,1,1]); b = MtlArray([2,2,2,2]); c = similar(a);

julia> @metal threads=2 groups=2 vadd(a, b, c)

julia> Array(c)
4-element Vector{Int64}:
 3
 3
 3
 3

Metal API wrapper

Finally, all of the above functionality is made possible by interfacing with the Metal libraries through ObjectiveC.jl. We provide low-level objects and functions that map These low-level API wrappers, along with some slightly higher-level Julia wrappers, are available in the MTL submodule exported by Metal.jl:

julia> dev = MTLDevice(1)
<AGXG13XDevice: 0x14c17f200>
    name = Apple M1 Pro

julia> dev.name
NSString("Apple M1 Pro")

Hacking

Metal.jl relies on a custom LLVM with an AIR back-end, provided as a JLL. Normally, this JLLis built on Yggdrasil. If you need to make changes to the LLVM back-end, have a look at the build_llvm.jl in the deps/ folder. This scripts builds a local version of the LLVM back-end, and configures a local preference such that any environment depending on the corresponding JLLs will pick-up the modified version (i.e., do julia --project in a clone of Metal.jl).

Acknowledgements

This package builds upon the experience of several Julia contributors to CUDA.jl, AMDGPU.jl and oneAPI.jl.

amadeusine/Metal.jl