rotaryembed/tanh/selu/mish/hardswish/hardsigmoid/gelu/erf/elu/eltwise/dropout/quantize/dequantize/bnll x86 support bf16 storage

[nihui]

No description provided.

[tencent-adm]

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[codecov-commenter]

Codecov Report

❌ Patch coverage is 97.79710% with 38 lines in your changes missing coverage. Please review.
✅ Project coverage is 93.17%. Comparing base (371bbad) to head (a6281e3).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
src/layer/x86/quantize_bf16s.h	87.26%	34 Missing ⚠️
src/layer/x86/dequantize_bf16s.h	99.29%	1 Missing ⚠️
src/layer/x86/dropout_x86.cpp	87.50%	1 Missing ⚠️
src/layer/x86/eltwise_x86.cpp	90.90%	1 Missing ⚠️
src/layer/x86/hardswish_bf16s.h	98.75%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6624      +/-   ##
==========================================
- Coverage   93.53%   93.17%   -0.36%     
==========================================
  Files         874      902      +28     
  Lines      281162   281968     +806     
==========================================
- Hits       262977   262720     -257     
- Misses      18185    19248    +1063     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

Adds x86-side bf16 storage support for the RotaryEmbed layer, including a bf16 implementation and an AVX512BF16 runtime-dispatched wrapper.

Changes:

• Enable support_bf16_storage for RotaryEmbed_x86 and route bf16 inputs to a dedicated forward_bf16s() path.
• Introduce rotaryembed_bf16s.h implementing the bf16 rotary-embedding kernel (with SIMD paths and optional AVX512BF16 runtime dispatch).
• Add rotaryembed_x86_avx512bf16.cpp wrapper entrypoint for the AVX512BF16-dispatched bf16 kernel.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
src/layer/x86/rotaryembed_x86.h	Declares bf16 forward helper behind NCNN_BF16.
src/layer/x86/rotaryembed_x86.cpp	Enables bf16 storage and dispatches to bf16 kernel when applicable.
src/layer/x86/rotaryembed_bf16s.h	Adds bf16 kernel implementation + runtime AVX512BF16 dispatch hook.
src/layer/x86/rotaryembed_x86_avx512bf16.cpp	Adds AVX512BF16 wrapper function for the runtime-dispatched bf16 path.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 56 out of 56 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.