skills.rest
dongg622
dongg622COMMUNITY
Explore all skills by this author →
GitHub

mindspeed-llm-train-profiler

Configure and automate CPU, memory, and step-level profiling for MindSpeed-LLM training on Ascend NPU.

Updated Aug 27, 2026
Search Tags:#performance#optimization#profiling#training#ascend#npu
One-click install
npx skills add https://github.com/dongg622/china-ai-chip-skill --skill mindspeed-llm-train-profiler
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mindspeed-llm-train-profiler
Source: https://github.com/dongg622/china-ai-chip-skill/tree/main/Ascend/mindspeed-llm-train-profiler
Command: npx skills add https://github.com/dongg622/china-ai-chip-skill --skill mindspeed-llm-train-profiler

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill guides users through efficiently configuring and automating profiling data collection during MindSpeed-LLM training on Ascend NPU devices.

Core Features & Use Cases

  • Training Profiling Automation: Assists in setting up and executing profiling for model training, including CPU, memory, and performance metrics.
  • Custom Configuration: Supports adjusting profiling levels, steps, and resource to tailor data collection for performance analysis.
  • Use Case: A user wants to collect detailed performance data from a training script to identify bottlenecks, enabling optimization of training workflows.

Quick Start

Provide the training script path and profiling preferences, then the Skill will generate a profiling script to run and collect data accordingly.

Frequently Asked Questions about mindspeed-llm-train-profiler

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I profile LLM training performance on Ascend NPU?▼

To profile LLM training on Ascend NPU, you provide your training script path and preferences to generate a profiling script. This script automates CPU, memory, and step-level data collection to identify performance bottlenecks.

What metrics can I collect during MindSpeed-LLM training profiling?▼

During MindSpeed-LLM training profiling, you can collect CPU, memory, and step-level performance metrics. This data collection is fully customizable, allowing you to adjust profiling levels and steps to tailor the analysis.

Can I customize the profiling configuration for my Ascend NPU training script?▼

Yes, you can customize profiling configuration for Ascend NPU training scripts by adjusting profiling levels, steps, and resource allocation. This tailors data collection specifically for your performance analysis needs.

What are the prerequisites for automating performance profiling on Ascend NPU?▼

Prerequisites for automating performance profiling on Ascend NPU include proper environment setup and ensuring training script correctness. Meeting these conditions ensures generated profiling commands execute and collect data without errors.

How does step-level profiling help identify bottlenecks in model training?▼

Step-level profiling helps identify bottlenecks in model training by capturing detailed performance data at specific execution steps. This granular data collection isolates CPU and memory inefficiencies within the training workflow.

Related Skills

Explore More Software Engineering Skills→
Tianyi-Billy-MaTianyi-Billy-Ma

pytorch-lightning

Automate PyTorch training workflows with LightningModule, Trainer, and LightningDataModule.

Community
Advanced
jacexhjacexh

python-performance-optimization

Profile Python CPU and memory usage with cProfile and memory_profiler.

Community
Advanced
qoobaqooba
2

tweaktune-synthesizer

Generate tweaktune pipelines for synthesizing LLM training datasets via interactive Q&A flows.

Community
Advanced
mtsatryanmtsatryan
1

performance-engineer

Profile CPU, memory, I/O, and latency to identify performance bottlenecks.

Community
Advanced