mojo build hello.mojo
./hello
- Only amd64 and arm64 - NVIDIA GPU support Q1 2024.
- JIT compilation with LLVM ORC
- Multi-Level Intermediate Representation MLIR - single static assignment for all the things.
- Mojo simd - n-body demo
- Mojo autotune - memset demo
- const/enum linting
- GPU pin of hot cache - PowerInfer: Fast Large Language Model Serving with a Consumer-grade GPU - PowerInfer Github
- tensor compression
- More MILR for IO/cache levels
- Equality saturation
2023 LLVM Dev Mtg - Mojo 🔥: A system programming language for heterogenous computing