ascendc

Official

Guide AscendC transformer op development.

Authorascend-ai-coding
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Developing AscendC transformer operators is complex and error-prone; this Skill provides a structured, pattern-based guide to build host/kernel interfaces, tiling, shape inference, and CANN examples for AscendC operators.

Core Features & Use Cases

  • Templates and best practices for AscendC operator development across FFN, GMM, and MoE variants.
  • Clear guidance on aligning op_host definitions, op_kernel implementations, tiling/infershape logic, and CANN examples with existing references.
  • Scaffolding support for the end-to-end workflow from frontmatter-driven discovery to production-ready operator code.

Quick Start

Use the genop tool in ops-transformer to scaffold a new AscendC operator and implement its host and kernel logic following the provided templates.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: ascendc
Download link: https://github.com/ascend-ai-coding/awesome-ascend-skills/archive/main.zip#ascendc

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.