Skip to content

docs: cudapp.lock is not on omega-mv main; name the real pins #212

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
aien-spark merged 1 commit into from
Oct 31, 2026
Merged

docs: cudapp.lock is not on omega-mv main; name the real pins #212

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions AUDIT_SCOPE.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,13 +132,18 @@ make hello-gpu
## Dependencies and provenance

- No lockfile at the root; C with no external libraries beyond libc and libm for the library.
`native/learn53/omega653.lock` is a sub-component lock.
Sub-component pins: `om/omega-bridge.lock` (omega-bridge commit for the Omega bindings and CI),
`om/pipeline/dependencies.lock`, `native/learn53/omega653.lock`,
`native/learn53/moe-closure.lock` and `tools/c5/omc.pin`.
- Vendored: `third_party/nvidia-open-580.173.02/` (67 NVIDIA headers, MIT, unchanged,
`SHA256SUMS` in that folder). Mesa NVK/NAK is cited in comments for hardware facts, no code
copied. Notices in `THIRD_PARTY.md` and `NOTICE`. History and licence audit in `docs/AUDIT.md`,
`docs/LICENSE_PROVENANCE.md`, `docs/LICENSING_TRANSITION.md`.
- AGENTS.md: no NVIDIA binary, cubin, PTX, firmware or NVIDIA tool listing goes in the repository.
- Omega builds the same sources at a pinned commit (`cudapp.lock` in omega-mv).
- Omega does not build these sources at a pin yet. omega-mv main has no `cudapp.lock`; it still
builds its own diverged copy of the GPU code (#106) and takes the driver client from physics-mv
(#107). The lock, with its `check-cudapp` target and `tools/cudapp-stage.sh`, is proposed in
the open draft omega-mv#924.
- No network access at build or test time is documented. At the described revision the repository
had no GitHub Actions workflow files; the audit-readiness change adds `.github/workflows/ci.yml`
(item 10 above), whose only network use is fetching the pinned actions on the runner.
Expand Down
11 changes: 7 additions & 4 deletions docs/START_HERE.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,10 +62,13 @@ blob, with refusal controls. Originals and historical commits are unchanged.
On x86-64 the same target is outside `make test` and exits 2, because the
pinned compiler is not part of the portable suite.

Canonical consumer integration uses the existing single-commit `cudapp.lock`,
not a second artifact or qualification registry. The isolated Omega consumer
cutover builds an authenticated committed source archive and links its actual
callers to that canonical library:
Canonical consumer integration is meant to use one single-commit `cudapp.lock`
in Omega, not a second artifact or qualification registry. That lock does not
exist on omega-mv main yet: `cudapp.lock`, the `check-cudapp` make target and
`tools/cudapp-stage.sh` are proposed in the open draft
[omega-mv#924](https://github.com/aien-dev/omega-mv/pull/924). That isolated
Omega consumer cutover builds an authenticated committed source archive and
links its actual callers to that canonical library; with #924 checked out:

```sh
OMEGA_OMC=/absolute/path/to/pinned/omc \
Expand Down
Loading