windows: CUDA serving build — cuda-guard + cmake/llvm runtime paths (rescued from #2056) - #2249
Merged
Merged
Conversation
…ime_path Three real fixes that get continuum-core to build+detect CUDA on native windows-msvc (validated: features cuda,directml -> detect_cuda -> 27 GiB on a 5090, was a bogus 4GB under directml-only): - cargo-features.sh: only add `cuda` on Windows+Nvidia when cl.exe is actually reachable, else directml-only. candle affine.cu needs nvcc->cl.exe; without it the whole build hard-failed instead of degrading. - install-manifest.toml: cmake + llvm-libclang had no [module.runtime_path], so start-server.sh installed them to ~/.continuum/tools but never put them on PATH -> every 'cmake not found' / bindgen libclang failure. Added windows runtime_path, regenerated manifests. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Q4NU4VNiELPQfBpCacDZGc
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Third rescue from the stalled queue, same method as #2244 and #2246.
#2056 is a 56-commit superset branch.
git cherryreports 55 unique by patch-id, but patch identity is not semantic supersession — the branch contains #2053's commits verbatim (97011f4f8,0ea9538a1,175cd6d51,10fed2798), several of which target a demand API canary has already replaced. Rebasing the whole thing means re-litigating a superseded feature 20 commits deep.So it gets taken apart into coherent units instead. This is the Windows-CUDA build unit — 4 files, +21/−4, touching
install-manifest.toml,cargo-features.sh, and the generated Windows manifest. Cherry-picked onto canary, applied clean.Scoped deliberately to the build plumbing, which is self-contained and platform-critical: it is the difference between the Windows/5090 node building a CUDA serving binary and not.
Still to rescue from #2056 (each verified against canary first, then its own PR):
98a6a16c8→b7b526455):ModelResidencyView,ResidencyBeacon/ResidencyLedger,GridResidencyModule,route_grid_overflow, and the persona effector closureexpert_observeharness (315298940,e47661aaf)afef13294ChatModule per-request,9d4509bceairc room-route echo)#2056 itself closes once its unique content is landed in reviewable pieces — it is not going to rebase as one branch, and pretending otherwise is what nearly destroyed it earlier tonight (blind
rebase --skipdropped commits that CREATE files; restored from origin only because nothing had been force-pushed).