triton-ascend-case-elemwise-broadcast-3d
CommunityTwo-stage kernels for fast 3D elementwise broadcast
Authorxchang1121
Version1.0.0
Installs0
System Documentation
What problem does it solve?
This skill targets performance improvements for cross-axis 3D broadcast by employing a two-stage kernel strategy, unlocking better vectorization and parallelism.
Core Features & Use Cases
- Two-stage kernel approach: first broadcast expansion and reshape to 2D, then a second kernel operates on the flattened dimension for high throughput.
- Performance uplift: especially effective when the last dimension is very small, improving utilization of hardware cores.
- Use Case: accelerate 3D broadcast operations on tensors with shapes like (B, H, W) and (1, H, 1).
Quick Start
Run the two-stage kernel workflow on a 3D broadcast task to observe improved throughput.
Dependency Matrix
Required Modules
None requiredComponents
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: triton-ascend-case-elemwise-broadcast-3d Download link: https://github.com/xchang1121/AutoResearch-CC-hook/archive/main.zip#triton-ascend-case-elemwise-broadcast-3d Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.