🌐 Web UI First · v0.2.1 on PyPI

nsys-ai

Open Nsight Systems profiles directly in a browser-first timeline UI, then drill into terminal tools when you need them.

$ pip install nsys-ai click to copy

nsys-ai profile.nsys-rep

GPU 0 | 39.0s – 42.0s 39.127s 39.532s 39.937s N0 [───────── Iteration 142 ──────────────────────────────────────────────────] N1 [── forward ──────────] [── backward ─────────────────────────────────] N2 [Attn] [MLP] [Norm] [Attn▸bwd] [MLP▸bwd] [AllReduce] ─────────────────────────────────────────────────────── S7 flash_fwd 26ms ████████░░████████████░██░░░████████████████░░██████████████████████ S12 ██░░████░░░░░░░░██░░░░░░████████░░░░░░░░░░░░░░░██████ S15 █████████████████████████░░░░░░░░░███████████████████████████████ ─────────────────────────────────────────────────────── 39.128s→39.154s │ 26.2ms [S7] │ flash_fwd_splitkv_kernel └─ 📁 Iteration 142 └─ 📁 forward └─ ▸ Attention ▶ flash_fwd_splitkv_kernel 26.2ms

🚀 Try It Now

Get started in 60 seconds with a real GPU training profile

Quick Start — Example 20: Megatron-LM DistCA

# 1. Install $ pip install nsys-ai # 2. Clone & download a real profile (~700 MB) $ git clone https://github.com/GindaChen/nsys-ai && cd nsys-ai $ cd examples/example-20-megatron-distca $ python download_data.py # 3. Open the web viewer (default command) $ nsys-ai output/megatron_distca.nsys-rep # 4. Optional: profile overview in terminal $ nsys-ai info output/megatron_distca.sqlite # 5. Optional: terminal timeline TUI $ nsys-ai timeline output/megatron_distca.sqlite --gpu 4 --trim 39 42

See all examples → examples/

🌐

Web Timeline

Default experience: multi-GPU browser viewer with progressive rendering, NVTX hierarchy bars, pinch-to-zoom, trackpad pan, and AI chat. No --trim required.

🖥️

Timeline TUI

Perfetto-style horizontal timeline with per-stream kernel visualization, NVTX hierarchy bars, and time-cursor navigation. Zoom, pan, and bookmark positions.

🌲

Tree TUI

Interactive NVTX hierarchy browser. Expand/collapse nodes, see kernel durations, filter by name or threshold, and navigate with vim-style keybindings.

🌐

HTML Exports

Generate interactive HTML viewers, Perfetto JSON traces for ui.perfetto.dev, and flat CSV/JSON exports for scripting and sharing.

🔍

Smart Search

Search kernels and NVTX annotations by name, with hierarchical search support. Filter by GPU, time window, and parent NVTX scope.

📊

Overlap Analysis

Compute vs NCCL overlap, collective breakdowns by type, and automatic training iteration detection for distributed workloads.

🤖

AI Analysis

Optional AI module for auto-commentary on kernel distributions, NVTX annotation suggestions, and bottleneck detection.

nsys-ai

🚀 Try It Now

Web Timeline

Timeline TUI

Tree TUI

HTML Exports

Smart Search

Overlap Analysis

AI Analysis

Keyboard-Driven Navigation

Navigation

Analysis

Bookmarks

📚 Documentation

Interactive Tools

🔍 SQLite Schema Explorer

📊 NVTX Stack Visualizations