[MLAS] Fix rotary avx2 kernel invalid access

[tianleiwu]

This fixes an issue that _mm256_maskload_ps intrinsic used in remainder-handling logic introduced in #23694.

The core of the problem is that _mm256_maskload_ps (and its store equivalent) can read beyond the masked elements.
Even if mask correctly specifies that you only want to load, for example, 3 floats, the intrinsic may still read the full 32 bytes (8 floats) from the provided memory address.

The invalid access occurs when one of buffers (input, sin_data, or cos_data) ends near the boundary of a memory page, and the part of the 32-byte read that you don't care about (i.e., the masked-off part) falls onto an unmapped page. This will cause a segmentation fault (invalid access).

The Solution: Use a Scalar Remainder Loop

The simplest, safest, and most robust solution is to replace the masked AVX remainder logic with a simple scalar loop. This is the exact strategy already used by your RopeKernel_Avx2_fp16_Impl functions, which are safe from this bug.

The performance impact of this change will be negligible, as this loop only processes the final 1-15 elements.

[Copilot]

Pull Request Overview

This PR fixes a critical memory access violation in the AVX2 rotary embedding kernel. The issue stems from the use of _mm256_maskload_ps and _mm256_maskstore_ps intrinsics, which can read/write beyond the masked elements, potentially causing segmentation faults when buffers are near page boundaries.

Key Changes:

• Removed masked AVX2 remainder handling logic that could cause invalid memory access
• Replaced with safe scalar loops for processing trailing elements (1-15 elements)

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}