Project progress
Measured history across the public AIEN repositories
541108
1972
2506
Codebase growth
Physical lines, including blanks and commentsAugust is retained site history. September 19 is the runtime's first recorded release; its initial import already contains substantial code.
Revisions and review
Verification activity
The test footprint
Named test files, deduplicated by contents. Inline tests remain in the general source count.
The language shift
C includes headers and .inc files; other languages include shell, assembly, web, Mojo, Python and CUDA.
Evidence at different scales
Selected reported experiments · candidate-specific329 loop sites classified
0 unclassified · silicon reruns NOT_RUN
Omega #172Recent repository activity
Merged and open pull requestsFrom runtime to computing stack
Historical milestones · select for evidenceThe repositories behind the growth
Current pinned heads · source and test lines| Repository | Source lines | Revisions | PRs | CI runs |
|---|
Rows count each repository independently. The project total deduplicates exact matching file contents across repositories, so rows do not sum to the headline.
Scope, counting rules and pinned sources
Scope: the original 29 public AIEN repositories, including archived precursors, experimental projects, documentation and project websites; newly discovered public AIEN, Spark, Omega, Atlas, Physics and related named repositories are added automatically. Forks are excluded. Private repositories, unrelated private applications, unmerged branches, deleted history and work before the retained imports are excluded.
Lines: physical source lines, including blanks and comments, in Rust, C/headers/.inc, C++, Mojo, assembly, shell, Python, web, CUDA and Zig files. Evidence, receipts, data, fixture directories, generated/build/vendor trees and symlinks are excluded. Exact matching blobs are counted once per daily snapshot. Similar migrated copies can remain. Test classification follows file-path names; embedded tests are not separated. This is repository footprint, not original authored SLOC.
Revisions: distinct commit SHAs reachable from the pinned default branches, including merges. Snapshots use the latest first-parent state at each Chicago day end. Pull requests include all states; merged counts use the merge date. CI counts use available GitHub Actions records, deduplicated by run URL and limited to the measurement cutoff. Deleted CI records cannot be reconstructed; a run is a workflow execution, not an individual test case.
Charts connect daily observations. Hover values between observations are interpolated; the date scrubber and headline counts use exact daily snapshots. The latest day ends at the measurement cutoff. Collection runs hourly through GitHub Actions and on site updates; scheduled runs can be delayed. This page checks for published data every five minutes. An unsuccessful collection leaves the previous deployment intact.
There is no comparable project-wide count of individual executed tests: assertions, input sweeps, fuzz iterations and qualification gates measure different things. Reported evidence above is linked to its PR and applies to its candidate. These measurements inspect Git history and GitHub records; they do not rerun the reported runtime experiments or observe hardware. code growth and CI activity do not establish full-system qualification.