Skip to content
CAI
Software that uses CAICheck a score

sirupsen/napkin-math

66.1

Adequate · 14 November 2026

267

lines of production code

Ruby

with Python

5

measurements over time

CAI band scale
CAI trend line
CAI lens gauges

What this system is

Features

Added newsletter 14 and 20 benchmarking scripts

Added initial scripts for newsletter 14 (syncing) and newsletter 20 (compound vs combining indexes), including a Jupyter notebook, Ruby benchmarking tools, and data population scripts to measure database query performance.

newsletter · high confidence

Initial release of Napkin benchmarking tool

The application has been initialized with a comprehensive set of benchmarks for file I/O (sequential and random reads/writes), memory operations, hashing (SHA-256), and database interactions (MySQL, Redis). The tool now supports command-line argument parsing via Clap, utilizes the Jemalloc allocator for performance, and provides detailed throughput and latency metrics for various operations.

src · high confidence

Behavioural changes

2 commits (0 fixes) modifying examples

A change to existing behaviour in examples — 2 commits, 12 files: “Move the tools to examples/, and target-cpu=native to .cargo/config.toml”; “bench: 16 GiB latency chases, a loopback echo baseline for services, C lz4 and unchecked lz4\_flex”.

examples · medium confidence · unverified

8 commits (0 fixes) modifying script

A change to existing behaviour in script — 8 commits, 3 files: “bench: skip setup that a filtered run won't use; cache builds per CPU”; “check README tables in CI, and clippy every bench”; “bench: AVX-512 reads, whole-VM SSD numbers, self-deleting bench VMs”; and 5 more.

script · medium confidence · unverified

Replaced legacy memory benchmarks with comprehensive CPU and compression performance tests

The benchmarking suite has been restructured to provide more detailed performance insights. The previous simple sequential read/write benchmarks (base\_rates.rs) have been removed and replaced with new, more sophisticated benchmarks: compressed\_memory\_read.rs, which measures the performance of bitpacking compression and decompression using the BitPacker8x algorithm, and memory\_read.rs, which evaluates single-threaded and multi-threaded memory read throughput, including AVX2 vectorized and non-vectorized paths on x86\_64 architectures. This allows for a deeper analysis of memory bandwidth and CPU-specific optimizations. Also in this area: Benchmark suite reorganized and renamed for clarity; Added benchmark suites for encryption, handshakes, memory writes, spawning, and synchronization; Benchmark suite overhaul: new tests, Criterion migration, and allocator update. Also in this area: 1 commit (1 fix) fixing benches; 3 commits (1 fix) modifying benches; 5 commits (2 fixes) modifying benches; 1 commit (0 fixes) modifying benches.

benches, benches/benchmarks · low confidence

Dependencies

Initial dependency setup for Go, Ruby, and Rust projects

This change introduces the initial dependency manifests for the project's various language implementations. For the Go module, it defines the base module path and includes the \golang.org/x/text\ library. In the Ruby newsletter examples, it adds database driver dependencies (\mysql2\ for the syncing example, and \pg\, \mysql2\, and \fiber\_scheduler\ for the compound indexes example). For the main Rust application, it updates \Cargo.toml\ to rename the package to \napkin-math\, adds a comprehensive set of dependencies including \rand\, \sha2\, \redis\, \clap\, \jemallocator\, and \criterion\, and configures the release profile with debug symbols enabled. Also in this area: 1 commit updating dependencies (2 manifests); 2 commits updating dependencies (2 manifests); Renames benchmark targets and removes Go module; Updated Arrow/Parquet dependencies and expanded encoding benchmarks; Add postcard serialization and TLS/crypto dependencies; and 1 more.

(dependencies) · medium confidence · unverified

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

How this codebase got here

This is the PUBLIC form of this artifact. Findings are listed in full, but the details of SECURITY findings — which rule fired, in which file, on which line, and how to fix it — are deliberately withheld, and any secret-scanner results are excluded entirely. Where detail is absent here it was REMOVED FOR PUBLICATION; it is not missing from the analysis. The complete artifact is available from the repository owner.

Score

  • CAI 61 → 66 (+5.3)
  • Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.

Lenses

  • Code Health 73 → 70 (-3.1)
  • Architecture 69 (new)
  • Maturity 57 → 62 (+5.4)
  • Readiness 52 → 65 (+13.1)
  • Security 93 → 90 (-3.4)
  • Performance 90 → 90 (+0.0)

Resolved (5)

  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • Medium: security finding (details withheld)
  • Medium: security finding (details withheld)

New (6)

  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • Low: security finding (details withheld)
  • Medium advisory (unsound): RUSTSEC-2026-0442 (Cargo.lock)

Changes since last survey

  • 43 commits — 38 feature/other, 5 fixes

By area

  • (root) — 35 commits
  • benches/benchmarks — 7 commits
  • script/bench-vm — 1 commit

Notable commits

  • fix: bench: fix clippy warnings on x86 and aarch64
  • fix: bench: fixes from a measurement review
  • fix: readme: November 2026 audit fixes, three-CPU suite, disk latencies
  • fix: readme: fix contradictions, add independent random reads; chase on 1 GiB pages
  • fix: readme: rerun after the bench fixes; hash throughput from memory
  • change: Move the tools to examples/, and target-cpu=native to .cargo/config.toml
  • change: bench-vm: random suffix on run IDs
  • change: bench: 16 GiB latency chases, a loopback echo baseline for services, C lz4 and unchecked lz4_flex
  • change: bench: AVX-512 reads, whole-VM SSD numbers, self-deleting bench VMs
  • change: bench: one-command search index tables over 53 datasets
  • change: bench: services round trips, bitpacking, 1 MiB QD1 sequential disk tests
  • change: bench: share disk helpers and the Sattolo shuffle
  • change: bench: skip setup that a filtered run won't use; cache builds per CPU
  • change: bench: storage on io_uring from one thread; label single-thread rows
  • change: bench: two-stream NEON reads, huge-page sequential row; 1-thread read 30 GiB/s
  • change: check README tables in CI, and clippy every bench
  • change: readme: 1-year prices on every row, and an HDD shape
  • change: readme: 1-year prices only where there's a discount
  • change: readme: CPU Decode Bitpacked Integers, all cores
  • change: readme: Memory Random Read is one 64 B cache line
  • …and 23 more

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Survey your own repository

sirupsen/napkin-math was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.

About this page

  • The score is its most recent published measurement, taken on 14 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit 44e8f8ab477d91138d15edbac4bfe4614019cc2c — the exact code this score is about.
  • Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
  • Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.65+preprod-e1a3b69fc8ee.