---
id: 20260805-T0-15
title: "Nova编译器：端到端MLIR方案，加速深度学习模型映射"
title_en: "Nova: An End-to-End MLIR Compiler for Deep Learning"
url: https://ai.daily.yangsir.net/daily/20260805-T0-15
issue_date: 2026-08-05
publish_date: 2026-08-04T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2608.00029
---

# Nova编译器：端到端MLIR方案，加速深度学习模型映射

arXiv新论文提出Nova，一款端到端MLIR编译器，用于深度学习模型的高效编译。论文指出，深度学习模型在大规模部署时，性能高度依赖高级数学运算与底层硬件之间的映射效率。Nova通过端到端的MLIR编译流程，优化了从张量框架到物理硬件的映射过程，减少了中间转换损耗。该编译器有望提升AI模型的训练和推理效率。

## English Version

**Nova: An End-to-End MLIR Compiler for Deep Learning**

A new arXiv paper presents Nova, an end-to-end MLIR compiler for deep learning. It argues that performance at scale depends on how well high-level math ops map to underlying hardware. Nova optimizes this mapping via a full MLIR compilation pipeline, cutting intermediate conversion overhead. The compiler could improve training and inference efficiency for AI models.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2608.00029)

**详情页**：https://ai.daily.yangsir.net/daily/20260805-T0-15

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*