Skip to main content

triton-ascend-matmul

适用于矩阵乘法(matmul)类算子的优化指南。当算子的核心计算涉及二维或更高维的矩阵乘法时应选择此指南,典型算子包括:matmul, mm, bmm, linear, gemm, outer_product, einsum(含矩阵乘), conv(转为矩阵乘实现)等。涵盖 Cube Core 使用、分块(tiling)策略、Swizzle 优化、大 K 维处理等关键技巧。不适用于纯逐元素运算或纯归约运算。对于 attention 机制中的 QK^T 和 score*V 矩阵乘,若算子整体是注意力计算,应优先选择 attention 指南。

Jump to install

Source facts

Repository
mindspore-ai/akg
Last source activity
April 7, 2026 at 02:26
Detected SKILL.md language
Chinese
Stars
259
Forks
48

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.