NVIDIA/nccl holds a health index of 84 out of 100, placing it in the Excellent band. It scores highest on Vitality (96/100) and lowest on Security (21/100). It was last updated today. 10 contributors account for most of its recent work.
84
overall / 100
Excellent
Software health index
Metrics are grouped into weighted categories on one standardized 1–100 scale. Overall starts as their weighted mean, calibrated against the distribution of the public record so bands carry percentile meaning; when public evidence triggers the High-Risk Jurisdiction Policy, the rating is adjusted and receives an At Risk ceiling of 34.
84
Exceptional93-100The record's top tier (≈ top 5%); essentially all checked criteria met
Excellent80-92Strong across the board; minor gaps
Good65-79Healthy; gaps are limited and manageable
Moderate50-64Acceptable with notable gaps; review recommended
Weak35-49Material weaknesses across several areas
At Risk20-34Significant weaknesses; adoption warrants caution
Critical1-19Severe problems (abandoned, single-maintainer, no hygiene)
Score profile
Each axis is a category. The shape matters more than the average — a healthy subject fills the whole shape, while a spike-and-crater profile means strength in one dimension is masking risk in another.
The weighted overall 71 is calibrated to 84 on the published index scale (record calibration 2026-08-02).
Direct dependencies free of known advisories — no direct dependency carries a known advisory
0/25
Indirect dependencies free of known advisories — transitive set not separable from development and test dependencies in this scope
0/40
No advisories left outstanding — no advisory carries a publication date
Inputs used
source
osv
advisories
0
affected_packages
0
assessed_packages
89
unassessed_packages
15
affected_by_severity
none
direct_affected_packages
0
Excluded from scoring (no data or not applicable): Indirect dependencies free of known advisories, No advisories left outstanding. Remaining weights renormalized. Matched 89 resolved dependencies against OSV. 15 could not be assessed — no resolved version, an unsupported ecosystem, or beyond the reported package list. This repository publishes no package the index resolves, so the repository dependency graph was assessed instead. That graph mixes development and test pins with shipped dependencies, so only the declared runtime dependencies are scored; transitive findings are reported as context and excluded from the score. Reachability is not analyzed.
How well is the repo equipped to be developed and maintained with AI coding agents? Carries a deliberately small weight (4%): agent tooling is a real maintenance signal, but a repository with none can still reach 100/100.
API schema (OpenAPI/GraphQL/proto) — not applicable to this kind of software
0/20
MCP server — not applicable to this kind of software
40/40
Runnable examples — example, examples
Inputs used
example_dirs
example, examples
has_mcp_signal
no
api_schema_files
—
interfaces_expected_of
—
Excluded from scoring (no data or not applicable): API schema (OpenAPI/GraphQL/proto), MCP server. Remaining weights renormalized.
Key facts
4,988GitHub stars
100contributors
1,035commits, last 12 months
0days since last push
14releases
10bus factor
264open issues
—package ecosystems
Data collection warnings
Star history unavailable: GitHub GraphQL error: Resource not accessible by personal access token
First-time contributor figures cover 12 of 19 authors (cap 12)
Could not fetch pypi package 'pace_cpp' from its registry
Could not fetch pypi package 'ubx._C' from its registry
pypi package 'pace' points at a different repository (https://github.com/Dispersive-Hydrodynamics-Lab/PACE); excluded from ecosystem scoring
Could not fetch pypi package 'ubx' from its registry
OpenSSF Scorecard did not return a usable result (2026/08/12 23:13:18 Warning: PATs stored in env variables GITHUB_AUTH_TOKEN and GITHUB_TOKEN differ. Scorecard will use the former.); skipping Scorecard checks
More detail
Star and fork history 0 ★ / 1,375 ⇿
0Stars
1,375Forks
12Releases
When each star and fork was added, collected from GitHub and bucketed by day. Cumulative growth sits directly above the daily additions it is made of, so the two read against each other: steady organic accretion looks nothing like an abrupt, short-lived burst. Where that difference is measurable, it is reported as growth authenticity.
Only the most recent history is shown — this repository exceeds the collection window, so the earliest history is not captured.
Major 0Minor 0Patch 0
Each point covers 6 days.
Direct dependencies 5
Registry
Package
Version constraint
Manifest
PyPI
packaging
—
bindings/nccl4py/pyproject.toml
PyPI
numpy
—
bindings/nccl4py/pyproject.toml
PyPI
cuda-core
~=1.0
bindings/nccl4py/pyproject.toml
PyPI
cuda-pathfinder
>=1.5.4,<2.0.0
bindings/nccl4py/pyproject.toml
PyPI
torch
>=2.1
contrib/nccl_ubx/pyproject.toml
All dependencies 104
Full resolved dependency set from the GitHub dependency graph: 8 direct and 96 indirect (transitive) packages. The transitive closure is complete when the repository commits a lockfile.
Registry
Package
Version
Relation
PyPI
cuda-core
1.1.1
direct
PyPI
cuda-pathfinder
1.6.0
direct
PyPI
numpy
—
direct
PyPI
numpy
2.2.6
direct
PyPI
numpy
2.4.6
direct
PyPI
numpy
2.5.1
direct
PyPI
packaging
26.3
direct
PyPI
torch
—
direct
crates.io
aho-corasick
1.1.4
indirect
crates.io
anyhow
1.0.103
indirect
crates.io
bindgen
0.71.1
indirect
crates.io
bitflags
2.13.0
indirect
crates.io
cexpr
0.6.0
indirect
crates.io
cfg-if
1.0.4
indirect
crates.io
clang-sys
1.8.1
indirect
crates.io
crc32fast
1.5.0
indirect
crates.io
crunchy
0.2.4
indirect
crates.io
cuda-bindings
0.1.0
indirect
crates.io
cuda-core
0.1.0
indirect
crates.io
cuda-device
0.1.0
indirect
crates.io
cuda-host
0.1.0
indirect
crates.io
cuda-macros
0.1.0
indirect
crates.io
either
1.16.0
indirect
crates.io
equivalent
1.0.2
indirect
crates.io
foldhash
0.1.5
indirect
crates.io
glob
0.3.3
indirect
crates.io
half
2.7.1
indirect
crates.io
hashbrown
0.15.5
indirect
crates.io
hashbrown
0.17.1
indirect
crates.io
indexmap
2.14.0
indirect
crates.io
itertools
0.13.0
indirect
crates.io
libc
0.2.186
indirect
crates.io
libloading
0.8.9
indirect
crates.io
libnvvm-sys
0.1.0
indirect
crates.io
log
0.4.33
indirect
crates.io
memchr
2.8.2
indirect
crates.io
minimal-lexical
0.2.1
indirect
crates.io
nom
7.1.3
indirect
crates.io
nvjitlink-sys
0.1.0
indirect
crates.io
object
0.36.7
indirect
crates.io
oxide-artifacts
0.1.0
indirect
crates.io
prettyplease
0.2.37
indirect
crates.io
proc-macro2
1.0.106
indirect
crates.io
quote
1.0.46
indirect
crates.io
regex
1.12.4
indirect
crates.io
regex-automata
0.4.14
indirect
crates.io
regex-syntax
0.8.11
indirect
crates.io
reserved-oxide-symbols
0.1.0
indirect
crates.io
rustc-hash
2.1.2
indirect
crates.io
shlex
1.3.0
indirect
crates.io
syn
2.0.118
indirect
crates.io
thiserror
2.0.18
indirect
crates.io
thiserror-impl
2.0.18
indirect
crates.io
unicode-ident
1.0.24
indirect
crates.io
windows-link
0.2.1
indirect
crates.io
zerocopy
0.8.52
indirect
crates.io
zerocopy-derive
0.8.52
indirect
PyPI
alabaster
0.7.13
indirect
PyPI
backports-strenum
1.3.1
indirect
PyPI
colorama
0.4.6
indirect
PyPI
coverage
7.15.3
indirect
PyPI
cuda-bindings
12.9.7
indirect
PyPI
cuda-bindings
13.2.0
indirect
PyPI
cuda-python
12.9.7
indirect
PyPI
cuda-python
13.2.0
indirect
PyPI
cupy-cuda12x
—
indirect
PyPI
cupy-cuda13x
—
indirect
PyPI
docutils
0.20.1
indirect
PyPI
duckdb
—
indirect
PyPI
exceptiongroup
1.3.1
indirect
PyPI
iniconfig
2.3.0
indirect
PyPI
jinja2
3.1.6
indirect
PyPI
markupsafe
2.1.5
indirect
PyPI
matplotlib
—
indirect
PyPI
ml-dtypes
0.5.4
indirect
PyPI
mpi4py
—
indirect
PyPI
mpi4py
4.1.2
indirect
PyPI
nccl4py
—
indirect
PyPI
nvidia-cuda-nvdisasm
13.3.73
indirect
PyPI
nvidia-cutlass-dsl
4.7.0
indirect
PyPI
nvidia-cutlass-dsl-libs-base
4.7.0
indirect
PyPI
nvidia-cutlass-dsl-libs-core
4.7.0
indirect
PyPI
nvidia-cutlass-dsl-libs-cu12
4.7.0
indirect
PyPI
nvidia-cutlass-dsl-libs-cu13
4.7.0
indirect
PyPI
nvidia-nccl-cu12
2.30.7
indirect
PyPI
nvidia-nccl-cu13
2.30.7
indirect
PyPI
pandas
—
indirect
PyPI
pluggy
1.6.0
indirect
PyPI
protobuf
6.33.6
indirect
PyPI
pyarrow
—
indirect
PyPI
pygments
2.20.0
indirect
PyPI
pytest
—
indirect
PyPI
pytest
9.1.1
indirect
PyPI
pytest-cov
7.1.0
indirect
PyPI
pytest-mpi
—
indirect
PyPI
pytest-mpi
0.6
indirect
PyPI
pytest-run-parallel
0.10.0
indirect
PyPI
ruff
—
indirect
PyPI
setuptools
—
indirect
PyPI
sphinx
7.1.2
indirect
PyPI
sphinx-rtd-theme
2.0.0
indirect
PyPI
tomli
2.4.1
indirect
PyPI
tqdm
—
indirect
PyPI
typing-extensions
4.16.0
indirect
Dependency advisories 0
This repository publishes no package the index resolves, so its own dependency graph was assessed — 89 packages, which also include development and test pins that never ship: 0 carry known advisories, of which 0 are direct. 15 could not be assessed — no resolved version, an unsupported ecosystem, or beyond the reported package list.
No known advisories affect the assessed dependencies.
An advisory means the version recorded in the dependency graph falls inside an advisory’s affected range. Reachability is not analysed, and the graph includes development and test pins — a finding may concern tooling rather than shipped software.
Raw JSON report machine-readable
Feedback
Spotted something off in this report, or have thoughts to share? Wrong measurements, missed tooling, ideas, questions — anything is welcome. Every message is read and gets a response.
Related records
Inspected repositories sharing catalogue tags or ecosystems with NVIDIA/nccl. Read them side by side →
Scores are signals, not warranties. They reflect publicly visible practices on GitHub — not a code audit, and not a security guarantee.
Missing data is excluded and weights renormalized, never scored as zero. Methodology is versioned and open: metrics v2.10.0, schema v0.31.0 — full methodology · metrics wiki.
