open_jev: add native Rust/CUDA text worker

[hsliuustc0106]

Purpose

Closes #54

Serve Open-Jev-27B-v1.1 through the existing Rust frontend with a native Rust/CUDA text worker. It compiles and tokenizes independent candidate prompts, applies the trained scalar decision head and saved temperature, and returns typed choice, score and noul answers.

• Extract the Qwen prefill implementation shared by Cua-S1 and Open-Jev, preserving Cua-S1's public imports.
• Add a CPU checkpoint export recipe and real inference before worker readiness.
• Fuse attention gating, cache residual RMSNorm values and pack BF16 MLP SiLU loads/stores while preserving rounding and reduction order. Other SiLU layouts use the scalar fallback.
• Add reference contract, tokenization and kernel coverage, plus setup and L20X results.

The CUDA ABI is now 4; rebuild the shared library and both native workers together.

Test Plan

System1-Omni Version / Commit: base 1be7d41eb74d5b49fd25194042041319950e5dce; head 40062bc12ac4b0d6364f61f0db963a49a1fde4f9.

Open-Jev tests and fixtures live under the repository-level tests/open_jev/; shared Qwen JSON/configuration and CUDA tests live under tests/qwen3_5/. Cargo explicitly registers the integration tests, and private unit tests load their files from that same top-level tree. Frontend CPU mock-worker coverage is tracked separately in issue #46 and PR #58.

Check prompts, typed answers and token IDs against pinned Open-Jev fixtures. Run the workspace checks and reserved-GPU kernel/worker validation. The recorded performance comparison fixes cached RMSNorm in both native variants and measures scalar versus packed SiLU on 74 single-candidate JevBench requests: one excluded feasibility pass and two measured passes per configuration. Hardware, revisions, controls and acceptance gates are in the linked results.

Test Result

Passed locally after cleanup:

cargo fmt --all --check
cargo clippy --workspace --locked --all-targets -- -D warnings
cargo test --workspace --locked
cargo build --workspace --release --locked

• After moving the tests: all workspace Rust checks passed, including 29 CPU tests; the pinned-checkpoint tokenizer test and strict MkDocs build passed. All 38 registered test names are preserved, including ignored GPU/checkpoint cases. The relocated six-test GPU suite compiled in release mode without device execution.
• Six CUDA ABI 4 tests and live worker/frontend smoke checks previously passed on reserved L20X GPU 2 at 61b83b3. The CUDA source and GPU test contents are unchanged; their location and registration changed, and device checks were not repeated. SM89 compilation passed previously; SM89 device execution and full Cua-S1 checkpoint inference were not checked.
• Recorded packed-SiLU improvement: warm HTTP mean 49.527→48.297 ms (2.483%), with all 74 native probabilities unchanged. Long-request SiLU time 21.217→6.380 ms (69.93%), closing 94.19% of that measured kernel-family gap to OpenJev-Fast.

The comparison covers single-candidate, eager text inference. Fast's preparation variability prevents claiming a stable aggregate winner, and the measured outputs do not establish native/Fast accuracy parity. Multi-candidate prefix sharing and Open-Jev CUDA Graph replay remain unvalidated. Nsight Systems supplied timelines because Nsight Compute counters were denied.

Self-review

Contributor checklist from CONTRIBUTING.md, left for completion before requesting review:

• I have reviewed the full diff and addressed the issues I found.
• I have checked that the change follows the project's architecture and stays focused on the stated purpose.
• I have run the checks appropriate to this change and reported commands, results, and anything I could not verify above.
• I have checked that the PR description, documentation, and any accuracy or performance claims match the implementation and available evidence.