x86 groupnorm instancenorm support bf16 storage with avx512bf16 dispatch

[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 98.44720% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 93.29%. Comparing base (cba96a6) to head (cbec927).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
src/layer/x86/groupnorm_bf16s.h	97.41%	4 Missing ⚠️
src/layer/x86/groupnorm_x86.cpp	97.77%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##           master    #6594    +/-   ##
========================================
  Coverage   93.29%   93.29%            
========================================
  Files         852      856     +4     
  Lines      271477   271866   +389     
========================================
+ Hits       253266   253636   +370     
- Misses      18211    18230    +19     

☔ 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

This PR adds bfloat16 (bf16) storage execution paths for the x86 InstanceNorm and GroupNorm layers, including AVX512BF16 runtime dispatch plumbing consistent with existing x86 bf16 patterns in the repo.

Changes:

• Enable support_bf16_storage for InstanceNorm_x86 and GroupNorm_x86 and route forward_inplace() to bf16 implementations when opt.use_bf16_storage is enabled.
• Add bf16 kernels for InstanceNorm/GroupNorm (*_bf16s.h) using SSE/AVX/AVX512 code paths with scalar fallbacks.
• Add AVX512BF16 runtime-dispatch entrypoints (*_x86_avx512bf16.cpp) to allow calling a BF16-flag-compiled implementation when supported at runtime.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
src/layer/x86/instancenorm_x86_avx512bf16.cpp	Adds AVX512BF16 runtime-dispatch target symbol for InstanceNorm bf16 kernel.
src/layer/x86/instancenorm_x86.h	Declares InstanceNorm_x86() and bf16 forward helper.
src/layer/x86/instancenorm_x86.cpp	Enables bf16 storage support + implements bf16 inplace forward using new helper kernels.
src/layer/x86/instancenorm_bf16s.h	Adds bf16 mean/var + affine transform kernel with AVX512BF16 runtime dispatch hook.
src/layer/x86/groupnorm_x86_avx512bf16.cpp	Adds AVX512BF16 runtime-dispatch target symbol for GroupNorm bf16 kernel.
src/layer/x86/groupnorm_x86.h	Declares bf16 forward helper for GroupNorm.
src/layer/x86/groupnorm_x86.cpp	Enables bf16 storage support + implements bf16 inplace forward (including packing handling).
src/layer/x86/groupnorm_bf16s.h	Adds bf16 GroupNorm kernel with AVX512BF16 runtime dispatch hook.

---

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