编译器
结合编译原理、AI 编译器论文与项目实践，理解程序表示和优化。
- 寄存器分配
- HIDA: A Hierarchical Dataflow Compiler for High-Level Synthesis
- MLIR: Scaling Compiler Infrastructure for Domain Specifific Computation
- The Deep Learning Compiler: A Comprehensive Survey
- TIRAMISU: A Polyhedral Compiler for Expressing Fast and Portable Code
GitHub 历史论文笔记
- Rammer: Enabling Holistic Deep Learning Compiler Optimizations with rTasks
- Roller: Fast and Efficient Tensor Compilation for Deep Learning
- Welder: Scheduling Deep Learning Memory Access via Tile-graph
- Cocktailer: Analyzing and Optimizing Dynamic Control Flow in Deep Learning
- AStitch: Enabling a New Multi-dimensional Optimization Space for Memory-Intensive ML Training and Inference on Modern SIMT Architectures
- Graphene: An IR for Optimized Tensor Computations on GPUs
- BOLT: BRIDGING THE GAP BETWEEN AUTO-TUNERS AND HARDWARE-NATIVE PERFORMANCE
- Chimera: An Analytical Optimizing Framework for Effective Compute-intensive Operators Fusion
- AMOS: Enabling Automatic Mapping for Tensor Computations On Spatial Accelerators with Hardware Abstraction
循环变换理论：未完成摘记按现有清单归入计算基础。