[huong8373tt-beep]

Summary

Fix unresolved softplus and softplus2 references in Inductor-generated Triton kernels. Replace platform-selected function aliases with named JIT forwarding functions so generated callers and helper definitions use matching names. The existing implementations, arithmetic, and platform selection are preserved. See also pytorch/pytorch#198050.

The final diff only changes fla/ops/utils/softplus.py.

Test plan

• Passed offline Inductor source-collection and Triton compilation checks on PyTorch 2.9.1 / Triton 3.5.1 for both helpers and both NVIDIA/generic implementations. All four baseline cases reproduce the missing-name error; the fixed cases compile.
• Passed pre-commit, header checks, and git diff --check.
• GPU runtime validation remains pending in a submitted H200 job. No test files are changed in this PR.

Benchmark / NCU

Offline PTX instruction comparisons match the baseline for all four helper/implementation combinations, excluding debug information. GPU latency measurements are pending; NCU was not run.

Breaking changes

None. Public signatures, numerical algorithms, and backend selection are unchanged.

Checklist

• I have read CONTRIBUTING.md and follow its conventions (code style, docstrings, commit prefixes).
• I have read AGENTS.md and, where my change matches its scope, the relevant skill under .agents/skills.
• Dependent tests pass locally or in CI, and new behavior is covered by tests where applicable (tick as N/A for changes with no testable code, e.g. docs-only).
• Kernel changes include same-hardware before/after benchmark numbers, dense + varlen where applicable (tick as N/A when no kernel code changed).
• This PR is minor/cosmetic-only (typo, formatting, style-only tweaks) — tick only if it is, and justify below.
• I understand this PR exceeds 500 changed lines and have explained its single purpose and why it should stay together below.

[yzhangcs]

Merged as 51e80c5. The final diff only changes fla/ops/utils/softplus.py; no test files or Dynamo-disable decorators are added.

Inductor can serialize softplus = softplus_nv as a definition named softplus_nv, while generated callers still reference softplus. Named JIT forwarding functions preserve the names of both helpers without changing arithmetic or platform selection. See also pytorch/pytorch#198050.

All four offline Inductor/Triton compilation checks pass and preserve the baseline PTX instructions. Static checks pass. GPU runtime validation and timings remain pending.