fix(promote-strided): include hip.npi_* ops in strided-input promotion

[kirankumarmatam-amd]

Problem

PromoteStridedHipOperands filtered ops using op->getDialect() == HipDialect.
NPI plugin ops (hip.npi_quantize, hip.npi_dequantize, etc.) are registered
in NpiDialect — a separate dialect object — even though their op names start
with hip. and they use the same contiguous-pointer kernel ABI. The dialect
check therefore excluded them, so strided memref.subview inputs (e.g. Q/K/V
slices produced by onnx.Split) were passed directly to NPI kernels that read
them as flat contiguous arrays, producing silently wrong values. On Qwen 7B VLM
(gfx1151) this caused cosine similarity to drop from ~1.0 to ~0.36.

Fix

Replace the dialect-object check with an op-name prefix check
(fullOpName.starts_with("hip.")), which covers both HipDialect ops and
hip.npi_* ops regardless of which dialect object owns them.

Testing

Verified on gfx1151 (Strix Halo, Windows D3D12 HMM): cosine similarity for
Qwen 7B VLM restored to ~1.0 after this fix.

[github-actions]

Thanks for opening a PR!

This project follows LLVM's incremental-development and AI-tool-use
guidance. See CONTRIBUTING.md
for the project workflow.

Before requesting review, please check that:

1. The change is focused. Substantial work links the relevant issue
   or design discussion.
2. The PR documents relevant test results and updates affected
   documentation.
3. If AI tools provided substantial assistance, the description
   explains what was assisted and how it was validated, and commit
   trailers identify the tool. The contributor has reviewed and
   understands the result.

Reviewers are assigned through
CODEOWNERS where ownership
is configured.

[github-actions]

L2 Accuracy Results (EP vs CPU)

Model	Combined L2	Total Elems	Skipped NaN/Inf
conv_test_hybrid	4.8668E-07	64	0
GroupQueryAttention_seq256	25.2366	2621440	0
MatMulNBits_o_seq128	259.906	368640	0
QMoE_seq128	34.957	368640	0

Threshold: 0.01 | Run: 3535 - Commit: 07eedbc

[github-actions]

MorphiZen EP Performance Results

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	7.52	6.20	361	3	1242
GroupQueryAttention_seq128	4406.44	1.69259	12	6	310
matmul_down_seq128	499.02	2.45	74	3	352

EPContext Export Performance

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	7.55	44.05	359	3	15590

EPContext Import Performance

Model	QPS	Session (s)	1st Infer (ms)	CPU%	Mem (MB)
full_model_seq128	7.53	8.66	362	3	15760

OGA Benchmark Results

Model	Warmup	Reps	Prompt Len	Gen Tokens	TTFT (ms)	TPS	Peak Mem (GB)	GPU Mem (GB)
gpt-oss-20b-webgpu-int4-rtn-block-32	1	5	128	128	169.2	79.1	1.33	13.57
Llama-3.1-8B-awq-g128-int4-asym-fp16-onnx-dml	1	5	128	128	345.0	40.0	1.22	6.43

OGA Wheel Smoke (Python benchmark_e2e.py)

Model	TTFT (ms)	TPS
Llama-3.1-8B-awq-g128-int4-asym-fp16-onnx-dml	194	39.3

Run: 3535 - Commit: 07eedbc