cpu: rv64: gemm: add bf16 gemm SIMD optimization with RISC-V V Extension

[ryanle1017]

Description

This PR introduces a SIMD-optimized bfloat16 GEMM kernel for the RISC-V 64-bit architecture, leveraging the RISC-V Vector (V) Extension. This work extends the foundational f32 GEMM implementation from PR #3785, enabling high-performance mixed-precision computations.
The primary motivation is to accelerate inference on emerging RISC-V platforms. As bfloat16 becomes a critical data type for modern deep learning models, offering significant memory bandwidth savings with a dynamic range comparable to f32, this optimized kernel fills a crucial performance gap.
This implementation focuses on the bf16:bf16:f32 data type combination (bfloat16 inputs, float32 accumulation and output), which is a common and numerically robust approach for mixed-precision GEMM.

Key Changes

• RVV-Optimized BF16 Kernel: Added a new GEMM kernel (rvv_gemm_bf16bf16f32) specifically designed for bfloat16 inputs and float32 outputs/accumulation.
• GEMM Dispatch Integration: The main GEMM dispatch logic in src/cpu/gemm/gemm.cpp is updated to route bf16bf16f32 requests to the new RVV kernel when running on a compatible RISC-V platform.
• Platform Feature Detection: Extended platform::mayiuse_bf16() to correctly detect and enable bfloat16 support when RVV intrinsics are available for RISC-V.

Checklist

General

• [√] Do all unit and benchdnn tests (make test and make test_benchdnn_*) pass locally for each commit?
• [√] Have you formatted the code using clang-format?

Performance improvements

• [√] Have you submitted performance data that demonstrates performance improvements?

All performance data was measured on a
--TODO
The performance baseline is the default ref_gemm implementation in the oneDNN main branch. Below are some example performance benchmarks for different problem sizes.

Matmul Primitives Performance(--dt=bf16:bf16:f32)

New features

• Have you published an RFC for the new feature?
• Was the RFC approved?
• Have you added relevant tests?

Bug fixes

• Have you included information on how to reproduce the issue (either in a github issue or in this PR)?
• Have you added relevant regression tests?

RFC PR

• Does RFC document follow the template?
• Have you added a link to the rendered document?