What is nsys-profile-analysis?

A guide for examining Nsight Systems trace files, which record when CPU and GPU work ran during an application. It uses the VeloQ command-line tool to query timelines, idle gaps, launches, correlations, concurrency, and captured metrics.

Use it to investigate GPU idle time, find why kernels or launches happened, connect CPU and GPU activity, inspect NVTX markers, examine CUDA graph replays, and check overlap or hardware metrics.

It turns large performance traces into structured results and helps show how CPU work led to GPU work. It also separates Nsight Systems traces from Nsight Compute reports and PyTorch profiling traces.

nsys-profile-analysis — Skill for Claude Code by lucifer1004/VeloQ

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add lucifer1004/VeloQ --skill nsys-profile-analysis

Clone the repo

git clone --depth 1 https://github.com/lucifer1004/VeloQ

Made for: Claude Code.

Or install VeloQ, the plugin that ships this one along with the rest of its 3 skills.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/lucifer1004/veloq/nsys-profile-analysis/github.svg)](https://agentmods.dev/skills/lucifer1004/veloq/nsys-profile-analysis)

Your own site

<a href="https://agentmods.dev/skills/lucifer1004/veloq/nsys-profile-analysis"><img src="https://agentmods.dev/badge/skills/lucifer1004/veloq/nsys-profile-analysis/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/lucifer1004/veloq/nsys-profile-analysis"><img src="https://agentmods.dev/badge/skills/lucifer1004/veloq/nsys-profile-analysis.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 60 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,408 The whole file, excluding the scripts and references it only reads on demand.

Security scan C 2 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00060	$0.01408
Opus 5.5	$0.00024	$0.00563
Sonnet 5.5	$0.00012	$0.00282
Haiku 4.5	$0.00006	$0.00141

Measured 2mo ago against content hash 2e6b3544bd87, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade C, and why

nsys-profile-analysis scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Downloads and executes remote codehighSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

curl -fsSL https://raw.githubusercontent.com/lucifer1004/veloq/main/scripts/install.sh | bash

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -fsSL https://raw.githubusercontent.com/lucifer1004/veloq/main/scripts/install.sh | bash

plugins/veloq/skills/nsys-profile-analysis/SKILL.md · 114 lines

How it starts

The opening of the file, as written. The whole thing — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Nsight Systems Profile Analysis

Use veloq as the evidence extractor for NSys timeline traces: what ran, when, how CPU work caused GPU work, where idle gaps are, and whether captured metric streams are trustworthy. Requires veloq on PATH.

For Nsight Compute .ncu-rep kernel reports use ncu-profile-analysis; for PyTorch/Kineto Chrome traces use pytorch-profile-analysis.

Quickstart

veloq info trace.nsys-rep   # detect source, trace map, applicable recipes
veloq summary trace.nsys-rep | jq '.data.auxiliary.capabilities'

Then query with stats / search / inspect / correlate / slices / gaps / timeline / concurrency / graph-replays / metrics / hardware. Per-command flags and the response schema live in veloq <cmd> --help and veloq schema <cmd>; canonical multi-step workflows live in veloq recipes / veloq recipes <id>. Those are the SSOT — do not duplicate them here.

stdout is always one JSON envelope: data.rows[] on success (every row carries a stable key), error on failure. Parse .data/.error, never stderr.

What veloq gives you over raw tables

Querying the exported sqlite/parquet directly is possible but you re-implement — and can silently get wrong — things veloq already does:

Correlation decode: runtime↔kernel/memcpy/memset joins through process-local correlationId bridged by TARGET_INFO_CUDA_CONTEXT_INFO and the globalTid PID mask (correlate/inspect do this for you).
NVTX attribution: forward (range → GPU work) and reverse (kernel → enclosing ranges) trees, incl. nesting depth and --nvtx scoping.
A correlation index + sidecar caches (<trace>.veloq/) reused across queries; first .nsys-rep access runs nsys export for you.
Stable row keys (kernel:1234, gap|pid:..|@..) — diff two captures with INDEX(.data.rows; .key) in jq.

If you do read raw tables anyway (user asked, or veloq lacks the query), you own the invariants in the pitfalls list below — every one of them has produced a plausible-but-wrong answer in practice.

Read the full file on GitHub · 114 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2mo ago First seen · 114 lines · 60 tokens per session scan C 2e6b3544bd87

Subscribe to this mod's changes

nsys-profile-analysis is a skill published in the GitHub repository lucifer1004/VeloQ (131 stars, last pushed 25d ago), licensed MIT. It adds 60 tokens to every session and 1,408 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Analytics and metrics GPU Scientific cli cuda ncu nsys profiling

modal-serverless-gpu

Run approved CPU or GPU work through OpenScience computejob on the user's configured Modal account. Use for isolated scientific scripts, dependency provisioning, durable outputs, logs, status, cancellation, and recovery. Never invoke the Modal SDK or CLI directly.

synthetic-sciences/openscience · 54 tokens

cuopt-developer

Modify, build, test, debug, and contribute to NVIDIA cuOpt (C++/CUDA, Python, server, CI). Use for solver internals, PRs, DCO, and code conventions.

NVIDIA/cuopt · 47 tokens

cuopt-user-rules

Base rules for end users calling NVIDIA cuOpt (routing/LP/MILP/QP/install/server). Not for cuOpt internals — use cuopt-developer for those.

NVIDIA/cuopt · 42 tokens

iket-profiling

Profile inside a CuTe DSL (cutlass.cute) kernel with IKET (In-Kernel Event Tracing) and the run-iket profiler. Use when developing or tuning a CuTe DSL kernel and you need per-warp timelines of kernel phases (setup, mainloop, epilogue), producer/consumer (TMA/MMA) activity, or time spent in pipeline and mbarrier…

humanfia/iket-profiling-skill · 134 tokens

cuda-kernel-optimizer

Use when optimizing or diagnosing GPU workload performance with CUDA, CUTLASS, Triton, PyTorch, vLLM or TensorRT-LLM; interpreting NCU, Nsys or PyTorch Profiler evidence; or preparing the workload and correctness checks for such optimization. General coding, skill maintenance and conceptual GPU questions do not…

troycheng/cuda-kernel-optimizer · 77 tokens

cuopt-numerical-optimization-formulation

LP, MILP, QP — concepts, problem-text parsing, and formulation patterns (parameters, constraints, decisions, objective). Concepts only; no API.