OpenArkCompiler 发行版

## Maple 1.0.0 Release Notes

Maple compiler infrastructure provides a C compiler, stable aarch64 static code generators, as well as a large suite of  optimizations.

### Front End

Hir2mpl supports .ast, .dex, .class and .jar as inputs. Currently, .dex, .class and .jar are not enabled. It enables the corresponding compilation process based on the inputs.

Hir2mpl takes .ast file generated by clang as input and converts it to a mpl file. For more details about mpl file and MapleIR, see [MapleIRDesign](https://gitee.com/openarkcompiler/OpenArkCompiler/blob/master/doc/en/MapleIRDesign.md).

### Middle End

Function inlining

Constant folding

Auto vectorization

Loop unrolling

CFG simplification

Value range propagation

Dead store elimination

Copy propagation

Partial Redundancy Elimination

Induction variable optimizations

Code sinking

Optimized basic block layout

### Back End

Copy propagation

Target specific propagation

Peephole optimization

Dead code elimination

If conversion optimization

Extended block optimization

Instruction scheduling

CFG optimization

### Other Information

At this time, Maple can work properly with SPEC CPU 2017, gcc.c-torture. We use Clang as our parser&lexer and aarch64-linux-gnu-gcc as our assembler&linker.

Our own parser&lexer is under development.

GVP方舟编译器 / OpenArkCompiler

搜索帮助

GVP 方舟编译器 / OpenArkCompiler