AI Compilers Are Redefining Execution Boundaries

青稞AI · wechat · 2026-07-18

This article systematically reviews the evolution of AI compilers and inference/training infrastructure. The core argument is: the traditional operator-bound compilation paradigm is no longer sufficient. Current work is pushing execution units down to the tile/task level, deepening semantic fusion, and incorporating communication into the IR. Key sections include:

1) Compilation boundaries shifting from operators to tiles/tasks

The author argues that workloads like LLMs, long context, KV cache, MoE, dynamic batching, and distributed training make the "one operator per boundary" abstraction too coarse.

Representative directions mentioned include:

2) Moving from plain fusion to "semantic fusion"

The article notes that traditional operator fusion falls short for patterns with reduction dependencies, like attention. FlashAttention's key breakthrough was using online softmax to bypass the limitation of materializing full scores.

It highlights Neptune:

Also mentioned:

3) Superoptimizers, megakernels, and dynamic execution

The article attributes the recent resurgence of superoptimizers to the sheer volume of model variants, shapes, dtypes, and attention masks, making it impossible for experts to handwrite all optimal kernels.

Representative works include:

The author emphasizes that truly effective optimization isn't just about "searching more," but organizing the search space correctly.

4) Distributed compilers start incorporating communication into the IR

The bottleneck on the training side is often communication, not pure compute. The article summarizes approaches to making communication explicit:

5) Conclusion

The author's final assessment is:

Original post →

More from Infra

Infra channel →