mm_cmpistrm: fix one test that wasn't being run

[mr-c]

The test data is indeed unit and not bit based.

The previous bad test wasn't being run as "SWORD_RANGES_UNIT" is specified on line 12481, and so the test using "SWORD_RANGES_BIT" wasn't triggered.

---

Summary by cubic

Fixes a misconfigured mm_cmpistrm SWORD range test so it runs in the correct mode. Replaces IMM_SWORD_RANGES_BIT with IMM_SWORD_RANGES_UNIT to match the unit-based data.

• Bug Fixes
  • Switched the test immediate to IMM_SWORD_RANGES_UNIT in tests/impl.cpp.
  • The suite runs SWORD_RANGES_UNIT, so the BIT-based test never executed; this now validates the intended case.

^{Written for commit 6cbcc1a. Summary will update on new commits.}

[cubic-dev-ai]

No issues found across 1 file

[jserv]

Thank @mr-c for contributing!

[mr-c]

Thank you all for the excellent hand written test cases! We are copying some of them into SIMDe 🙏🏻

[jserv]

Thank you all for the excellent hand written test cases! We are copying some of them into SIMDe 🙏🏻

I am pleased to hear that SSE2NEON has been helpful to the development of SIMDe.

With the release of SSE2NEON v1.9.x, I believe that all known translations from SSE intrinsics to their Armv7-A and Armv8-A counterparts have been fully implemented and validated. The current codebase reflects a comprehensive and carefully tested mapping across the supported intrinsic set.

It would be excellent to see selected improvements from SSE2NEON cherry-picked into SIMDe, further strengthening it as a unified and robust SIMD translation solution across architectures.