Official agent skill

Compileiq Booster Pack

by NVIDIA in NVIDIA/CompileIQ

Use BEFORE running a full CompileIQ search. An agent skill from NVIDIA/CompileIQ.

OfficialApache-2.0Auto-check: notesDevOps & Cloud

Install Compileiq Booster Pack

skills CLI
$ npx skills add NVIDIA/CompileIQ --skill compileiq-booster-pack -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install NVIDIA/CompileIQ compileiq-booster-pack --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/NVIDIA/CompileIQ.git skills-src && mkdir -p .claude/skills && cp -r skills-src/agent-skills/compileiq-booster-pack .claude/skills/compileiq-booster-pack && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
compileiq-booster-pack
GitHub stars
137
Token cost
~2.1k tokens
SKILL.md length
686 words
Files
2 (incl. scripts)
Skills in repo
7
Repo updated
First seen
Licence
Apache-2.0

At a glance

Use BEFORE running a full CompileIQ search. An agent skill from NVIDIA/CompileIQ.

  • Works in 5 steps: Pre-flight: the O0/O3 ACF-injection… → Download → Apply one ACF at a time → …
  • Speed up without searching
  • SKILL.md covers When, Available packs (today), Steps and Self-test, plus 2 more sections
  • Runs Shell scripts from its folder; calls python, gh and bash

What it does

Compileiq Booster Pack is an agent skill from NVIDIA/CompileIQ, published by the product's own GitHub organization. Use BEFORE running a full CompileIQ search. Walks through downloading a Booster Pack from NVIDIA/CompileIQ GitHub Releases, applying ACF candidates one at a time to the user's compiler (raw PTXAS, NVCC, Triton, Helion, FlashInfer), and keeping only candidates that compile, pass correctness, and beat the no-ACF baseline. Includes the mandatory Debug-pack O0/O3 ACF-injection canary that proves the ACF is reaching PTXAS. Triggers on "booster pack", "ACF", "apply-controls", "speed up without searching", "helion fp8"…

Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files, including scripts (for example `scripts/apply_one_acf.sh`).

It sits in DevOps & Cloud, covering Deployment. It works with NVIDIA AI Platform and GitHub. The repository describes itself as: An Optimizer for Nvidia Compilers. The licence is Apache-2.0.

When your agent uses it

  • Speed up without searching
  • Flashinfer batch decode

Example prompts

  • “booster pack”
  • “apply-controls”
  • “speed up without searching”
  • “/compileiq-booster-pack”

Requirements

  • Python 3
  • A Bash shell
  • Pre-approved tools (allowed-tools): Bash, Read

Workflow steps

5 steps, taken from the step headings in SKILL.md.

  1. Pre-flight: the O0/O3 ACF-injection canary (mandatory first step)
  2. Download
  3. Apply one ACF at a time
  4. Validate every candidate
  5. Reproducibility log

What it can do on your machine

Read from SKILL.md and the folder at commit 743aca4. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves these tools, so the agent can use them without asking each time:

    • Bash
    • Read

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Shell), which the agent can run.

    Shell commands in SKILL.md call:

    • python
    • gh
    • bash

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use gh, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Compileiq Booster Pack loads about 2.1k tokens when it runs. Until then it costs about 146 tokens; SKILL.md has 686 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~146
When it runs · the whole SKILL.md, loaded when a task matches
~2.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NotePre-approves every shell command (allowed-tools: Bash)SKILL.md
    allowed-tools: Bash, Read

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from NVIDIA/CompileIQ at commit 743aca4, republished under its Apache-2.0 licence (© NVIDIA). 686 words, ~2,079 tokens.

Download SKILL.mdSave it as .claude/skills/compileiq-booster-pack/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
compileiq-booster-pack
description
Use BEFORE running a full CompileIQ search. Walks through downloading a Booster Pack from NVIDIA/CompileIQ GitHub Releases, applying ACF candidates one at a time to the user's compiler (raw PTXAS, NVCC, Triton, Helion, FlashInfer), and keeping only candidates that compile, pass correctness, and beat the no-ACF baseline. Includes the mandatory Debug-pack O0/O3 ACF-injection canary that proves the ACF is reaching PTXAS. Triggers on "booster pack", "ACF", "apply-controls", "speed up without searching", "helion fp8", "flashinfer batch decode", "debug pack".
allowed-tools
Bash, Read
when_to_use
- Workload is close to a published pack (Helion FP8 quant / causal depthwise conv / Gated DeltaNet fwd; FlashInfer BatchDecode is a known related workload)…
license
Apache-2.0
metadata.version
1.0.0
metadata.author
NVIDIA CompileIQ
metadata.domain
compiler-optimization
paths
**/*.cu, **/*.cuh, **/*.acf, **/*.py, **/*.sh

compileiq-booster-pack

Try curated .acf candidates before running a full CompileIQ search. A Booster Pack is a zip of ACFs that NVIDIA validated against a specific workload family. They are not guaranteed speedups; treat every candidate as workload-specific and validate it on your own benchmark.

Authoritative narrative: docs/booster_packs.md, docs/flashinfer_booster.md.

When

If this is trueUse this path
Workload is close to a Booster Pack's intended workload, compiler, GPU, and validation context.Try the Booster Pack first.
Workload differs materially or no pack candidate helps.Run a full CompileIQ search (compileiq-run-search).
Baseline, correctness check, compiler path, or benchmark setup are not in place.Wait. Fix those before applying any ACF.

Available packs (today)

PackWorkloads it was validated againstNotes
booster-pack-helion.zipHelion FP8 Quantization, Causal Depthwise Convolution, Gated DeltaNet ForwardHas shown benefit on FlashInfer BatchDecodeWithPagedKVCacheWrapper; related attention workloads worth testing.
booster-pack-debug.zipDiagnostic ACFs (O0, O3, others that disable or alter selected optimizations)Not for speed; for debugging. Use the O0/O3 canary below before trusting any other pack.

The public release shape is documented in docs/booster_packs.md. Read each candidate's compiler_stages from its pack manifest and use every listed stage. There is no runtime download API today. Don't invent one.

Steps

0. Pre-flight: the O0/O3 ACF-injection canary (mandatory first step)

The most common silent failure when applying ACFs is a framework cache (Triton, Helion, FlashInfer's flashinfer_cubin/flashinfer_jit_cache, NVCC build cache) serving a stale binary that ignored the ACF. The Debug pack has two ACFs with predictable, opposite-direction signatures:

  • ptxas_opt0.acf: forces unoptimized PTXAS compilation. Applied → expect a measurable regression (often 2-10x slower) vs. baseline.
  • ptxas_opt3.acf: forces the default PTXAS optimization level. Applied → expect to match baseline (the no-ACF default is already -O3).
bash
# Baseline
T_BASE_MS=$(./run-benchmark.sh)

# O0 must regress
PTXAS_OPTIONS="--apply-controls=booster-pack-debug/ptxas_opt0.acf" T_O0_MS=$(./run-benchmark.sh)

# O3 must match baseline
PTXAS_OPTIONS="--apply-controls=booster-pack-debug/ptxas_opt3.acf" T_O3_MS=$(./run-benchmark.sh)

python -c "
import sys
base, o0, o3 = $T_BASE_MS, $T_O0_MS, $T_O3_MS
if o0 < base * 1.05:
    print('FAIL: O0 did not regress; ACF is NOT reaching PTXAS. Fix the cache-bust.')
    sys.exit(1)
if abs(o3 - base) / base > 0.05:
    print(f'WARN: O3 differs from baseline by >5%; baseline may not be -O3 or framework caching differs.')
print('PASS: ACF injection is wired up correctly.')
"

If this fails, stop. Fix the cache-bust before trying any real pack candidate:

  • Triton: export TRITON_ALWAYS_COMPILE=1, unique TRITON_CACHE_DIR per eval.
  • Helion: export HELION_SKIP_CACHE=1.
  • FlashInfer: confirm flashinfer_cubin and flashinfer_jit_cache packages are absent (docs/flashinfer_booster.md:56-64).
  • Raw nvcc: clean the build dir between candidates.
1. Download

Browse https://github.com/NVIDIA/CompileIQ/releases, find the latest tag matching booster-packs-*, and download the relevant pack zip plus the top-level booster-pack-catalog.json.

bash
BOOSTER_TAG="$(gh release list -R NVIDIA/CompileIQ --limit 100 --json tagName,isDraft \
  --jq '.[] | select(.isDraft == false) | select(.tagName | startswith("booster-packs-")) | .tagName' \
  | head -n 1)"
echo "Using $BOOSTER_TAG"
gh release download "$BOOSTER_TAG" -R NVIDIA/CompileIQ -p 'booster-pack-helion.zip' -p 'booster-pack-catalog.json' -D ./packs
unzip ./packs/booster-pack-helion.zip -d ./packs
cat ./packs/booster-pack-helion/booster-pack-manifest.json

Always read the per-pack manifest before applying: it lists the intended workload, compiler version, GPU target, validation context, and known caveats. For a reproducible rerun, set BOOSTER_TAG to the exact tag printed above.

2. Apply one ACF at a time
TargetInjection
Raw PTXASptxas -v -arch=sm_100 --apply-controls candidate.acf kernel.ptx
NVCC (CUDA source)nvcc -Xptxas --apply-controls=candidate.acf -arch=sm_100 kernel.cu -o exe
TritonPTXAS_OPTIONS="--apply-controls=candidate.acf" TRITON_ALWAYS_COMPILE=1 python bench.py
HelionHelion's official ACF API + HELION_SKIP_CACHE=1 (see helionlang.com/examples/acfs/softmax_acf.html).
FlashInferFLASHINFER_EXTRA_CUDAFLAGS="--ptxas-options=--apply-controls=$ACF_FILE" python bench.py (see docs/flashinfer_booster.md:107).

Apply exactly one ACF per run. If it fails to compile, hangs, crashes, returns wrong answers, or regresses, reject that candidate and move to the next.

Show full SKILL.md (244 more words)Show less
3. Validate every candidate
  • Compare against a known-good reference (correctness, not just speed).
  • Test multiple input shapes when shape matters.
  • Use compile and runtime timeouts to bound runaway candidates.
  • Run multiple performance trials if the benchmark is noisy.
  • Record the reproducibility checklist below (one row per candidate).
4. Reproducibility log

For every candidate you accept or reject, append a row to booster-pack-log.csv with:

  • ACF filename (and sha256)
  • Manifest / release version
  • Benchmark command
  • GPU model + driver version
  • CTK version
  • nvcc and ptxas paths + versions
  • Framework version or commit
  • Input shape
  • Baseline result (mean ± std)
  • Candidate result (mean ± std)
  • Correctness status
  • Decision: KEPT or REJECTED:<reason>

This is the same checklist docs/flashinfer_booster.md:135-148 recommends. The scripts/apply_one_acf.sh helper does most of this automatically.

Self-test

bash
bash scripts/apply_one_acf.sh --self-test

Dry-runs a "baseline vs baseline" comparison (no ACF applied to either side) and confirms the helper correctly reports "NOT a real improvement". Catches misconfigured script invocations before they pollute the reproducibility log.

Gotchas

  • Pack name is not a hard boundary. Helion Pack helps some FlashInfer cases (docs/booster_packs.md:34); test before assuming.
  • Booster Packs are not search-space inputs. Don't try to feed an ACF through PtxasSearchSpace(...); packs are already-generated .acf candidate bundles, not inputs to PtxasSearchSpace or NvccSearchSpace.
  • Force recompilation. If you can't prove a recompile happened between candidates, don't trust the measurement. See the cache-bust hints under the pre-flight canary.

Next

  • If no pack candidate helps your workload, go to compileiq-run-search for a full CompileIQ search over PtxasSearchSpace().
  • For attention workloads specifically, also see compileiq-search-space (variant="att").

© NVIDIA, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (scripts) in agent-skills/compileiq-booster-pack of NVIDIA/CompileIQ.

  • SKILL.md
  • scripts/apply_one_acf.sh

Open the folder on GitHubat commit 743aca4

Compare with similar skills

Compileiq Booster Pack next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Compileiq Booster Pack compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Compileiq Booster Pack this skillNVIDIA/CompileIQ137—~2.1kAutomated safety check: NotesApache-2.0
AI News RadarLearnPrompt/ai-news-radar1.8k—~2.5kAutomated safety check: NotesMIT
Reflexo ReleaseMyriad-Dreamin/typst.ts1.2k—~1.5kAutomated safety check: PassApache-2.0
Use Vercel Actionamondnet/vercel-action764—~2.7kAutomated safety check: PassMIT
1panel App Builderarch3rPro/1Panel-Appstore213—~1.3kAutomated safety check: PassMIT
Prepare Cloudflare Production DeploymentLubomirGeorgiev/cloudflare-workers-nextjs-saas-template786—~5.9kAutomated safety check: NotesMIT

Similar skills

  • AI News Radar

    LearnPrompt/ai-news-radar

    A skill your agent uses when working on AI News Radar, 24 小时 AI 更新雷达, AI 更新雷达, 伯乐Skill, or Scout Skill: finding high-signal AI/tech sources, adding RSS/OPML/GitHub feeds, checking source health…

    1.8k GitHub stars~2.5k tokensUpdated today
    DevOps & CloudAuto-check: notes
  • Reflexo Release

    Myriad-Dreamin/typst.ts

    Guide Reflexo/typst.ts release preparation and operator handoffs.

    1.2k GitHub stars~1.5k tokensUpdated yesterday
    DevOps & CloudAuto-check passed
  • Use Vercel Action

    amondnet/vercel-action

    Wire amondnet/vercel-action into a GitHub Actions workflow to deploy Vercel projects from CI.

    764 GitHub stars~2.7k tokensUpdated 2 mo ago
    DevOps & CloudAuto-check passed
  • 1panel App Builder

    arch3rPro/1Panel-Appstore

    A skill your agent uses when packaging Docker deployments as 1Panel local app store apps, including GitHub projects, docker-compose.yml files, docker run commands, app metadata, version directories…

    213 GitHub stars~1.3k tokensUpdated today
    DevOps & CloudAuto-check passed
  • Prepare Cloudflare Production Deployment

    LubomirGeorgiev/cloudflare-workers-nextjs-saas-template

    Source-of-truth runbook for preparing this Vinext Cloudflare Workers SaaS template for production deployment.

    786 GitHub stars~5.9k tokensUpdated yesterday
    DevOps & CloudAuto-check: notes
  • Broccoli Oss GCP Deploy

    besimple-oss/broccoli

    Deploy this repository to a new Google Cloud project using the repo's existing Cloud Run, Cloud Run Jobs, Cloud SQL, Secret Manager, and Artifact Registry scripts.

    286 GitHub stars~4.1k tokensUpdated 5 mo ago
    DevOps & CloudAuto-check passed

More from NVIDIA/CompileIQ

  • Compileiq Bootstrap

    NVIDIA/CompileIQ

    Official

    A skill your agent uses when starting a fresh CompileIQ project, hitting a socket timeout, or before running any other compileiq- skill.

    137 GitHub stars~1.3k tokensUpdated 15 days ago
    Auto-check: notes
  • Compileiq Debug

    NVIDIA/CompileIQ

    Official

    A skill your agent uses when something is wrong: Search() hangs, all evaluations return INVALIDSCORE, scores aren't improving, every config returns the same number, ptxas errors fill the log, CV% is…

    137 GitHub stars~2.9k tokensUpdated 15 days ago
    Auto-check: notes
  • Compileiq Run Search

    NVIDIA/CompileIQ

    Official

    A skill your agent uses when composing the Search(...) call and calling .start().

    137 GitHub stars~2.2k tokensUpdated 15 days ago
    Auto-check: notes
  • Official

    Use AFTER a Search has completed and BEFORE claiming any speedup or shipping an ACF.

    137 GitHub stars~2.2k tokensUpdated 15 days ago
    Auto-check: notes
  • Official

    A skill your agent uses when writing the objectivefunction= passed to Search().

    137 GitHub stars~2.3k tokensUpdated 15 days ago
    Auto-check: notes
  • Compileiq Search Space

    NVIDIA/CompileIQ

    Official

    A skill your agent uses when picking the searchspace= argument for Search().

    137 GitHub stars~1.9k tokensUpdated 15 days ago
    Auto-check: notes

Categories

Questions about Compileiq Booster Pack

What does Compileiq Booster Pack do?

Use BEFORE running a full CompileIQ search. An agent skill from NVIDIA/CompileIQ. Compileiq Booster Pack is an agent skill from NVIDIA/CompileIQ, published by the product's own GitHub organization. Use BEFORE running a full CompileIQ search.

When should I use Compileiq Booster Pack?

Compileiq Booster Pack fits situations like: speed up without searching; flashinfer batch decode.

How do I install Compileiq Booster Pack in Claude Code?

Run `npx skills add NVIDIA/CompileIQ --skill compileiq-booster-pack -a claude-code`. Or copy the skill folder (agent-skills/compileiq-booster-pack in NVIDIA/CompileIQ) into .claude/skills/compileiq-booster-pack in your project. Claude Code loads it when a task matches its description.

How do I install Compileiq Booster Pack in Codex?

Run `npx skills add NVIDIA/CompileIQ --skill compileiq-booster-pack -a codex`. Or copy the skill folder (agent-skills/compileiq-booster-pack in NVIDIA/CompileIQ) into .agents/skills/compileiq-booster-pack in your project. Codex loads it when a task matches its description.

Can I use Compileiq Booster Pack in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add NVIDIA/CompileIQ --skill compileiq-booster-pack -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/compileiq-booster-pack, .gemini/skills/compileiq-booster-pack, .github/skills/compileiq-booster-pack and .opencode/skills/compileiq-booster-pack in your project.

What does Compileiq Booster Pack need to run?

Going by SKILL.md and its folder, Compileiq Booster Pack needs a shell for the scripts in its folder and the command-line tools its instructions call (python, gh and bash). Our summary lists: Python 3; A Bash shell. Its frontmatter pre-approves these tools: Bash, Read.

Does Compileiq Booster Pack access the network?

SKILL.md contains no URLs. Its commands use gh, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Compileiq Booster Pack safe to install?

Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Compileiq Booster Pack use?

Compileiq Booster Pack is published under the Apache-2.0 licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Compileiq Booster Pack use?

About 2.1k tokens (SKILL.md is roughly 8.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Compileiq Booster Pack?

Skills that share tags, products or a category with Compileiq Booster Pack: AI News Radar (LearnPrompt/ai-news-radar, 1.8k stars), Reflexo Release (Myriad-Dreamin/typst.ts, 1.2k stars), Use Vercel Action (amondnet/vercel-action, 764 stars) and 1panel App Builder (arch3rPro/1Panel-Appstore, 213 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Compileiq Booster Pack?

NVIDIA (a GitHub organization, an official publisher) maintains it in NVIDIA/CompileIQ, which has 137 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on September 23, 2026.

Source: NVIDIA/CompileIQ on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.