Replace QH16 bf16 kernel with a new one that does not use ptr_RP

[JohnNikolay84]

Motivation

#2729 has introduced a new QH64 kernel that is not writing directly to ptr_RP and instead is writing split data into ptr_R/logits.

As this #2983 states other kernels like MLA_A16W16_1TG_4W_32mx1_16nx1_Coex0_Msk1_QH16.co do not follow the same logic and write into a null pointer instead.

Technical Details

This change is introducing a new kernel for nhead=32 bf16 that is using same convention as QH64 kernel. However I have not been able to find a kernel with mfma layouts 32x32x16, instead I am using the one with 16x16x32.

Test Plan

Run a new test in aiter and make sure it pass torch reference.
Run DeepSeek in TP4 and make sure it is not crashing.

Test Result

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

[github-actions]

🏷️ CI Guide

Runs automatically on every PR:

• ✅ Pre-checks (submodule verification, code formatting)
• ✅ Aiter op tests (gfx942 + gfx950)
• ✅ Triton tests on MI35X (only when aiter/ops/triton/** or related paths are changed)

Extended tests (opt-in via labels):

Label	Tests
ci:triton-300x	Run an additional Triton test job on MI300X in PRs; main branch always runs both MI35X and MI300X
ci:sglang	SGLang integration tests
ci:atom	ATOM benchmark (DeepSeek-R1 + GPT-OSS)
ci:vllm	vLLM benchmark
ci:all	All of the above

Add labels via the sidebar or gh pr edit 2999 --add-label <label>