awesome-tensor-compilers

Why we included this project

Picking a compiler for deep learning workloads usually means digging through scattered project pages and papers. This index collects the main ones in one place, including TVM, MLIR, Triton, and Halide, and organizes the papers by the problems people actually run into: auto-tuning, cost models, quantization, sparse computation, and optimization for specific hardware like CPUs, GPUs, and NPUs. That structure lets you jump straight to the work that matches your bottleneck instead of assembling the list yourself. Teams comparing compiler stacks and researchers surveying the field will both find it useful. It is a reference index rather than software you deploy, but it saves real time when you need to know what compiler options exist and which papers explain them.