Rust is supported on Linux/arm64 systems as a tier1 platform along side x86.
All available Graviton processors (excluding the first and not recommended Graviton1) have support for the Armv8.2 instruction set. Armv8.2 specification includes the large-system extensions (LSE) introduced in Armv8.1. LSE provides low-cost atomic operations:
LSE improves system throughput for CPU-to-CPU communication, locks, and mutexes.
The improvement can be up to an order of magnitude when using LSE instead of
load/store exclusives. LSE can be enabled in Rust and we've seen cases on
larger machines where performance is improved by over 3x by setting the RUSTFLAGS
environment variable and rebuilding your project.
export RUSTFLAGS="-Ctarget-feature=+lse"
cargo build --release
If you're running only on Graviton2 or newer hardware you can also enable other instructions by setting the cpu target such as the example below:
export RUSTFLAGS="-Ctarget-cpu=neoverse-n1"
cargo build --release
When Rust is configured to use LLVM 12 or newer, target feature
+outline-atomics is available. Outline-atomics produces a binary containing
two versions of the atomic operation following the hardware capabilities. When
the code executes on a newer hardware such as Graviton2, the processor will
execute LSE instructions; when the code executes on older hardware without LSE
instructions, the processor will execute Armv8.0 atomics instructions.
Rust 1.57 (release on December 2, 2021) enables by default outline-atomics target feature when compiling for arm64-linux with LLVM 12 or newer. When using older Rust releases, outline-atomics target feature can be enabled with
export RUSTFLAGS="-Ctarget-feature=+outline-atomics"