Software:Cranelift
| Developer(s) | Bytecode Alliance |
|---|---|
| Initial release | 2016 |
| Stable release | 0.93.0
|
| Type | Compiler |
| License | Apache License with LLVM exceptions[1] |
| Website | https://github.com/bytecodealliance/wasmtime/blob/main/cranelift/codegen/ |
Cranelift (formerly known as Cretonne) is an optimizing compiler backend written in Rust, which converts a target-independent intermediate representation into executable machine code. The project started in 2016 and is currently developed by Bytecode Alliance.[2][3] Unlike compiler backends such as LLVM which focuses on ahead-of-time compilation, the emphasis of Cranelift is on just-in-time compilation with short compile time being an explicit goal of the project.[4]
As of 2023, Cranelift supports instruction set architectures such as x86-64, AArch64, RISC-V, and IBM z/Architecture.
History
Prior to the backend framework rewrite in 2020, the project made use of only one intermediate representation (IR) across all compilation stages prior to machine code emission, namely CLIF (Cranelift IR Format). The old design made use of instruction legalizations, which involved transforming the high-level IR continuously until each CLIF instruction corresponded to an instruction of the target machine, at which point the executable code could be emitted.
In 2020, the backend was rewritten to use a separate, machine-specific IR called VCode for later compilation stages instead. After this rewrite, a high-level input CLIF would be optimized by a middle layer before being lowered into VCode, which would be further transformed by remaining passes in the backend.[4] Peepmatic, a tool used to generate a peephole optimizer from a domain-specific language (DSL), was also added.[5]
In 2022, a new register allocator, which was adapted from the register allocator from IonMonkey, was added. In addition, the backend framework was reworked to utilize ISLE, another DSL in the project that was made to ease the development of the instruction selection stage in backends.[6]
The mid-end saw addition of optimization passes that make use of e-graphs in 2022.[7] The optimization passes are enabled by default starting in 2023.
Applications
The project is being used by multiple WebAssembly runtimes such as Wasmtime and Wasmer, with performance comparable to LLVM-based runtimes.[8]
An alternative code generator for the Rust compiler that is powered by Cranelift also exists.[9]
References
- ↑ "LICENSE.md". https://github.com/bytecodealliance/wasmtime/blob/main/cranelift/codegen/LICENSE.
- ↑ "GitHub - bytecodealliance/cranelift at 399be07fcf00d0974d2a83d1af55f0bee9861d54" (in en). https://github.com/bytecodealliance/cranelift/tree/399be07fcf00.
- ↑ "wasmtime/cranelift at main · bytecodealliance/wasmtime" (in en). https://github.com/bytecodealliance/wasmtime/tree/main/cranelift.
- ↑ 4.0 4.1 Fallin, Chris. "A New Backend for Cranelift, Part 1: Instruction Selection". https://cfallin.org/blog/2020/09/18/cranelift-isel-1/.
- ↑ "Introduce peepmatic: a peephole optimizations DSL and peephole optimizer compiler by fitzgen · Pull Request #1647 · bytecodealliance/wasmtime" (in en). https://github.com/bytecodealliance/wasmtime/pull/1647.
- ↑ Fallin, Chris. "Cranelift's Instruction Selector DSL, ISLE: Term-Rewriting Made Practical". https://cfallin.org/blog/2023/01/20/cranelift-isle/.
- ↑ "egraph-based midend: draw the rest of the owl (productionized). by cfallin · Pull Request #4953 · bytecodealliance/wasmtime" (in en). https://github.com/bytecodealliance/wasmtime/pull/4953.
- ↑ Denis, Frank. "Performance of WebAssembly runtimes in 2023 | Frank DENIS random thoughts.". https://00f.net/2023/01/04/webassembly-benchmark-2023/.
- ↑ "Cranelift codegen backend for rust". 26 January 2023. https://github.com/bjorn3/rustc_codegen_cranelift.
 |  |
