benchmark

A command for running performance checks on a kernel project. It can run a quick check, selected workload strides, or a full benchmark, including compilation, correctness, latency, and numerical-error measurements.

In plain words
What is it for?
It helps compile and test kernels, measure latency against a reference, count passes and failures, and report maximum absolute and relative error.
Why use it?
Kernel changes need repeatable checks to show whether they still work and whether they became faster or less accurate. This command collects those results across small and large workloads.

Command for Claude Code

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.

agentmods
npx agentmods add commands/dogacel/auto-gpu-kernel/benchmark
Clone the repo
git clone --depth 1 https://github.com/Dogacel/auto-gpu-kernel

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 197 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00197
Opus 5 $0.00000 $0.00098
Sonnet 5 $0.00000 $0.00039
Haiku 4.5 $0.00000 $0.00020

Measured 2d ago against content hash 6b3d064e8743, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

benchmark scanned grade A with 0 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 2d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

dsa_sparse_attention_h16_ckv512_kpe64_topk2048_ps64/.claude/commands/benchmark.md · 14 lines

What it actually says

/benchmark

Arg Command When
quick (default) modal run scripts/run_modal.py --quick Compile + correctness (smallest + largest workload)
stride N modal run scripts/run_modal.py --stride N Iteration (default N=2 → ~10 workloads)
full modal run scripts/run_modal.py Lock in final numbers (128 workloads, 10-15 min)

Pipe output to a file (e.g. bench.log in repo root) — /log-experiment will attach it.

If Modal crash-loops (container fails to boot repeatedly, not just slow), cancel and diagnose; don't sit waiting.

Report back: pass/fail counts, absolute kernel latency (min / median / max, split small vs large when both are present), max abs/rel error, reference latency.

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.

  1. 2d ago First seen · 14 lines · 0 tokens per session scan A 6b3d064e8743

Subscribe to this mod's changes

benchmark is a command published in the GitHub repository Dogacel/auto-gpu-kernel (157 stars, last pushed 11d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 197 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.