simd-optimization

A guide to SIMD optimization on x86-64 and ARM processors. SIMD is a way to apply one operation to many data values at once, and x86-64 and ARM are different processor families.

In plain words
What is it for?
It helps implement or debug SIMD code, compare vectorized changes on named ARM and x86-64 chips, and report benchmarks without treating one machine’s results as universal.
Why use it?
It warns that an optimization’s speed can differ between processor types, especially when memory access limits performance.

Skill for Claude CodeCodex

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 skills/rust-works/succinctly/simd-optimization
Any agent
npx skills add rust-works/succinctly --skill simd-optimization
Clone the repo
git clone --depth 1 https://github.com/rust-works/succinctly

Made for: Claude Code, Codex.

Per session 65 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,691 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 $0.00065 $0.02691
Opus 5 $0.00032 $0.01345
Sonnet 5 $0.00013 $0.00538
Haiku 4.5 $0.00006 $0.00269

Measured 2d ago against content hash 221e1e02c683, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

simd-optimization 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.

.claude/skills/simd-optimization/SKILL.md · 241 lines

How it starts

The opening of the file, as written. The whole thing — 241 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SIMD Optimization Skill

Patterns and learnings from SIMD optimization in this codebase.

Comprehensive documentation: See docs/optimizations/simd.md for full details on SIMD techniques.

Key Insight: Memory-Bound Effects Do Not Port Across Platforms

For anything cache- or bandwidth-bound, the effect size differs by architecture, not just the noise — so a single-platform measurement can mischaracterise a change, not merely blur it.

O6 (#106) removed a repeated interest-bitmap rescan. The same commit measured 6.1x on Apple M4 Pro and 16.4x on Ryzen 9 7950X. Post-fix times were comparable (124 ms vs 137 ms); the pre-fix times differed 3x (758 ms vs 2240 ms), because Apple's memory subsystem absorbed the thrashing far better than Zen 4. Measuring only Apple Silicon understated the fix by 2.7x. The same asymmetry is why project_benchmark_feature_flags-style AVX-512 findings must never be presented as universal.

Rule: any claim about a memory-bound path needs both an ARM and an x86_64 number, and the tables must name the chip. See docs/guides/benchmarking.md § A/B Benchmarking Method.

Key Insight: Wider SIMD != Automatically Faster

Two AVX-512 optimizations implemented with dramatically different results:

AVX512-VPOPCNTDQ: ~5–9× vs a Baseline count_ones(), ≈1× Native (Compute-Bound)

Implementation: src/bits/popcount.rs

  • Processes 8 u64 words (512 bits) in parallel
  • Hardware _mm512_popcnt_epi64 instruction
  • Result: ~5–9× faster than a baseline-build count_ones() (which lowers to scalar broadword) — e.g. 96.8 GiB/s vs 18.5 GiB/s.

Why it wins — but only vs a baseline build: Pure compute-bound and embarrassingly parallel, so explicit VPOPCNTDQ crushes scalar broadword. But compile with -C target-cpu=native and count_ones() auto-vectorizes to VPOPCNTDQ itself, reaching ≈1× parity — the explicit path's remaining value is portable binaries that still reach VPOPCNTDQ via runtime is_x86_feature_detected! dispatch. Measured data: Popcount Strategies (#45).

Read the full file on GitHub · 241 lines

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 · 241 lines · 65 tokens per session scan A 221e1e02c683

Subscribe to this mod's changes

simd-optimization is a skill published in the GitHub repository rust-works/succinctly (51 stars, last pushed 2d ago), licensed MIT. It adds 65 tokens to every session and 2,691 once invoked, about $0.0003 per session on Opus 5. 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.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens