Add SVE2 optimization for Fill32f

[ermig1979]

Summary

• Add Simd::Sve2::Fill32f using SVE2/SVE predicated stores for aligned body and tail handling.
• Route SimdFill32f through SVE2 before NEON when available.
• Extend Fill32fAutoTest for SVE2 and register the new source in VS2022 SVE2 project files.
• Document the SVE2 Fill32f optimization in the 7.2.163 release notes.

Testing

• cmake ./prj/cmake -B build -DCMAKE_BUILD_TYPE=Release -DCMAKE_CXX_COMPILER=g++ -DCMAKE_C_COMPILER=gcc -DSIMD_TOOLCHAIN="g++" -DSIMD_TARGET="" -DSIMD_AVX512VNNI=ON -DSIMD_AMXBF16=ON -DSIMD_TEST_FLAGS="-march=native" -DSIMD_SHARED=ON
• cmake --build build --parallel$(nproc)
• export LD_LIBRARY_PATH="/workspace/build:$LD_LIBRARY_PATH" && ./Test "-r=.." -fi=Fill32f -tt=1 -ts=1
• aarch64-linux-gnu-g++ -march=armv9-a+sve2 -msve-vector-bits=scalable -DSIMD_STATIC -I./src -c src/Simd/SimdSve2Fill.cpp -o /tmp/SimdSve2Fill.o
• aarch64-linux-gnu-g++ -march=armv9-a+sve2 -msve-vector-bits=scalable -DSIMD_STATIC -I./src -c src/Simd/SimdLib.cpp -o /tmp/SimdLib.o
• aarch64-linux-gnu-g++ -march=armv9-a+sve2 -msve-vector-bits=scalable -I./src -D_GLIBCXX_USE_NANOSLEEP -c src/Test/TestFill.cpp -o /tmp/TestFill.o